mirror of
https://git.triforceproject.com/triforce/source
synced 2026-11-22 17:19:58 -04:00
Compare commits
15 Commits
| Author | SHA1 | Date | |
|---|---|---|---|
|
|
e6f2f248f2 | ||
|
|
70c210cee1 | ||
|
|
043db52c62 | ||
|
|
6c7117635c | ||
|
|
b0e46adb48 | ||
|
|
fb26495475 | ||
|
|
a750a21025 | ||
|
|
72793f2231 | ||
|
|
6f338c8f56 | ||
|
|
93eeb714c2 | ||
|
|
183b05d8c4 | ||
|
|
b2ca6a6bb6 | ||
|
|
23a743a977 | ||
|
|
40aa5f5595 | ||
|
|
aff80bd56f |
9
.github/workflows/linux-appimage.yml
vendored
9
.github/workflows/linux-appimage.yml
vendored
@@ -9,6 +9,13 @@ on:
|
||||
tags: ['*']
|
||||
pull_request:
|
||||
workflow_dispatch:
|
||||
inputs:
|
||||
# (a build that replaces a release's Linux file names that release: a game built from it compares its version with
|
||||
# the latest release's tag by name (host/hos_update.c), and git describe's v0.3.6-1-g... would be told of v0.3.6)
|
||||
version:
|
||||
description: "The version the AppImage names itself (empty: git describe's)"
|
||||
required: false
|
||||
default: ''
|
||||
permissions:
|
||||
contents: read
|
||||
concurrency:
|
||||
@@ -95,6 +102,8 @@ jobs:
|
||||
ls -la work work/image
|
||||
|
||||
- name: AppImage
|
||||
env:
|
||||
HOOKSHOT_VERSION: ${{ inputs.version }} # (empty but for a started run that names one: appimage.sh asks git then)
|
||||
run: scripts/linux/appimage.sh
|
||||
|
||||
# (AMD FSR 4's Linux side, a build option the AppImage leaves off (gpu/vk/vk_fsr4.cpp: dlopen, the older backend
|
||||
|
||||
2
.gitignore
vendored
2
.gitignore
vendored
@@ -18,6 +18,8 @@ gamefile/
|
||||
|
||||
# ---- build output and caches ----
|
||||
build/
|
||||
# (but launcher/Logic/Build/ is source: git on Windows ignores case, so build/ hid it and v0.5.2 was committed without it)
|
||||
!launcher/Logic/Build/
|
||||
__pycache__/
|
||||
*.pyc
|
||||
*.o
|
||||
|
||||
@@ -37,8 +37,15 @@ Open an [issue](https://github.com/ruckus7382sfgsgdf/CBeta0.2/issues) and includ
|
||||
- **a game problem:** `triforce_log.txt` from your game folder (the folder with `Triforce.exe`). If you started the
|
||||
game again after the problem, send `triforce_log.1.txt` too: each start keeps the four runs before it as
|
||||
`triforce_log.1.txt` (the last) to `triforce_log.4.txt`.
|
||||
- **a Setup or build problem:** `setup.log` from `%LOCALAPPDATA%\Programs\Triforce`, and for a failed build the text
|
||||
from **Show details** on the build page.
|
||||
- **a freeze:** `triforce_hang.txt` from the game folder as well, when it is there. After 20 seconds without a frame
|
||||
the game writes down where each of its threads was (addresses in the program, no game data).
|
||||
- **something seen in play:** press **F8** when you see it, and send the folder it names (`captures\<date>-<time>`
|
||||
in the game folder): a picture of the frame, the end of the log, your settings and your save, with no game file
|
||||
among them. The System tab's **Export save for a bug report** makes one zip of the save and the logs instead.
|
||||
- **a Setup or build problem:** `setup.log` from `%LOCALAPPDATA%\Programs\Triforce`. For a build that stopped, press
|
||||
**Copy report** on the build page and paste it: the Setup and tools versions, what the build was started with,
|
||||
the last 200 lines of **Show details** and of `setup.log`, with your user folder written as `%USERPROFILE%` so
|
||||
your name is not in it.
|
||||
|
||||
Before you report:
|
||||
|
||||
@@ -48,7 +55,7 @@ Before you report:
|
||||
injectors: they change what Triforce does.
|
||||
- If graphics look wrong after an update, clearing the shader caches often helps: close the game and delete
|
||||
`triforce_programs.bin`, `triforce_programs.idx`, `triforce_pipelines.bin`, `triforce_pipeline_states.bin`,
|
||||
`triforce_computes.bin`, `mkw_vk_pipelines.bin` and `triforce_prepared.bin` from the game folder (never
|
||||
`triforce_computes.bin`, `mkw_vk_pipelines.bin`, `mkw_vk_pipelines_play.bin` and `triforce_prepared.bin` from the game folder (never
|
||||
`triforce_shaders.bin`). The next start prepares them again.
|
||||
|
||||
One topic per issue or pull request, in English, please.
|
||||
@@ -90,7 +97,70 @@ To build from a clone instead:
|
||||
The launcher lives in `setup/`: a .NET Framework 4.8 WPF app, built with `powershell -File setup\build.ps1` (needs the
|
||||
C# compiler from Visual Studio Build Tools). `setup\pack_tools.ps1` packs the tools zip a release carries. Never run a
|
||||
development build's install, update or uninstall without `TRIFORCE_TEST_ROOT` set to a scratch folder: it would act
|
||||
on your real install.
|
||||
on your real install. With it the install, the settings (a text file in place of the registry), the shortcuts, the
|
||||
game folders and the build's work folder (`<root>\build`) all stay inside that folder. The network is not part of
|
||||
that: such a run still asks the release site and downloads, unless `TRIFORCE_RELEASE_API` names another address.
|
||||
|
||||
What keeps a run of Setup from damaging an install is in `setup/Guard.cs`, written without the window so that
|
||||
`tests/test_setup_*.py` compile it alone (with the same C# compiler) and run it, no Setup started:
|
||||
|
||||
- **One Setup at a time for an install.** The first one holds a lock named after the folder it works in (and, in a
|
||||
test run, after `TRIFORCE_TEST_ROOT`, so a test never meets the real install's Setup). A second start brings the
|
||||
first one's window to the front and ends; an uninstall is refused while Setup is open.
|
||||
- **Room on the disk, asked first.** Before the tools are downloaded Setup asks the disk for room for the zip and for
|
||||
what it unpacks to (five times the zip's size, until the zip is there and says it exactly before its first file is
|
||||
unpacked), and stops with both numbers, what it takes and what is free, when there is none. With less than 60 GB on
|
||||
the game folder's disk the Build button says so in the same way and still lets the player go on. An update of a game
|
||||
folder is not asked about: what it takes depends on what changed.
|
||||
- **Downloads that go on, and are checked.** A file is fetched into `<file>.part`. One that was cut off (the
|
||||
connection, Cancel, Setup closed) is gone on from at the next try with a Range request, but only as a part of the
|
||||
same file: `<file>.part.of` names the address and size it was begun for and the server's ETag (else its date) for
|
||||
it, which goes back as If-Range; anything else is fetched anew. A server that sends nothing for 60 seconds is given
|
||||
up. The file gets its name only once it has the size the release lists, which the server must not contradict. A
|
||||
tools zip that is there whole is not fetched again; one that does not unpack is.
|
||||
- **Copy report**, beside Show details after a build that stopped (not one the player cancelled): what Setup knows of
|
||||
the build and the last 200 lines of its details and of `setup.log` go to the clipboard, with the account's folder
|
||||
(in either slash, doubled backslashes, any case, and its short name) replaced by `%USERPROFILE%`. A name in a path
|
||||
outside that folder stays.
|
||||
- **Its own update, undone when the new Setup does not start.** The new exe takes the running one's name, the running
|
||||
one goes on as `TriforceSetup.old.exe` and starts the new one with `--relaunched`. No second program watches: the
|
||||
old Setup itself stays, out of sight, until the new one's window is up (a named event), 30 seconds at most. A new
|
||||
Setup that cannot be started, or that ends with an error's exit code before that, is replaced by the old exe again in
|
||||
one rename; the old window comes back and says so, and that release's Setup is not fetched again on this PC
|
||||
(`SetupSkip` in `HKCU\Software\Triforce`; the tools still are). One that runs on without the sign, or ends without
|
||||
an error, is left alone.
|
||||
|
||||
Three more parts are written the same way, each a file without the window and a test that runs it on folders of
|
||||
its own:
|
||||
|
||||
- **The Mods page** (`setup/Mods.cs`, `tests/test_setup_mods.py`; the page is `Shell.Mods.cs`). It keeps no list:
|
||||
the mods folder is read as `host/mods.c` reads it, and every change is one a player could make by hand. On and off
|
||||
is the `-` in front of a folder's name; the order is a number in front (`01-`, two or three digits and a hyphen),
|
||||
kept apart from the mod's own name, and moving a mod numbers the folders of the mods folder anew, all of them or
|
||||
none; a removed mod moves into `mods\.removed`; an archive is unpacked into `mods\.adding-<id>` (a name with a
|
||||
dot in front is not looked at by the game), checked as it lies there (no path out of the folder, no link, no
|
||||
device's name, a bound on size and count), and only its mod folder moves into place. `.zip` is read with .NET,
|
||||
`.7z` and `.rar` with the `tar.exe` in System32 and nothing else: do not add an unpacker to the repository. Every
|
||||
write goes through one question, "is this inside the mods folder?", and no change is made while the game of that
|
||||
folder runs. Code patches (`exefs`) are built in: Setup keeps a mark of them at each build (`ModsBuilt` in
|
||||
`HKCU\Software\Triforce`) and says "Build again to apply" while the mods' mark is another one. Cheats are not in
|
||||
that mark: the game reads them when it starts.
|
||||
- **Add to Steam** (`setup/Steam.cs`, `tests/test_setup_steam.py`; the pictures are drawn in `SteamArt.cs`, the
|
||||
link is `Shell.Steam.cs`). Steam's `shortcuts.vdf` is another program's file: it is read whole, other shortcuts are
|
||||
copied through as bytes and never rebuilt, a file that does not read to its end is refused, the new file takes the
|
||||
old one's name in one rename, and the old one is kept (`.triforce-backup`, `.before-triforce`). Nothing is
|
||||
written while Steam runs, and only from the question the player answers. **No test and no test run may touch a
|
||||
real Steam folder:** Steam's folder is always handed in, tests author their own in a temporary folder, and with
|
||||
`TRIFORCE_TEST_ROOT` Setup looks only at `<root>\Steam` and never asks the registry.
|
||||
- **Sounds** (`setup/Sound.cs`, `tests/test_setup_sound.py`; where they play is `Shell.Sounds.cs`). Setup plays the
|
||||
game's own decide and cancel sounds, `Sound/Resource/Stream/SE_Sys_Do00.bwav` and `SE_Sys_Cancel00.bwav`, read from
|
||||
the player's game folder (or from the mod the game itself would read them from) and decoded in memory. **No sound
|
||||
is in the repository, in a release or in Setup, and none may be added:** before a game is built, and for a file that
|
||||
is missing or does not decode, Setup is silent. Nothing decoded is written to the disk. The tests author their own
|
||||
BWAV files and never read a game's. Sounds play only for what the player does (a click, a row chosen, a choice
|
||||
changed; not a page the programme turns), not while the game runs, and never in a run without a player
|
||||
(`--snapshot`, `TRIFORCE_UI_SCRIPT`, `TRIFORCE_TEST_ROOT`): such a run gets, with `TRIFORCE_SOUND_LOG=<file>`, a
|
||||
line in that file for each sound that would have played, and still makes none.
|
||||
|
||||
## Building
|
||||
|
||||
@@ -117,6 +187,10 @@ Each replacement is named by the CRC-32 of the module the renderer's own emitter
|
||||
coverage line in the log to say so: dump the modules again (`RX_VK_DUMP_ALL_SPV=<folder>`) and match them by what they
|
||||
read. Never share a game folder built with it.
|
||||
|
||||
What RenoDX's mod does not have (RenoDRT as a tone mapper, the gamma correction) is in `gpu/vk/shaders/renodx_glue`:
|
||||
files of Triforce's own that are compiled together with the fetched shaders and call RenoDX's library there. Keep it
|
||||
that way: nothing of RenoDX's source goes into them or anywhere else in the repository.
|
||||
|
||||
## Testing
|
||||
|
||||
The game-free checks need no game files:
|
||||
@@ -158,7 +232,7 @@ it builds on. By submitting a change you agree to license it under those terms.
|
||||
an incompatible license, and never paste decompiled game code or anything generated from the game files.
|
||||
|
||||
The third-party parts (Eden, suyu, Sirit, the Vulkan Memory Allocator, SDL, Opus, Dear ImGui, zstd, the Libertinus
|
||||
fonts, zbic and the a64-kit tools) keep their own notices. Preserve them.
|
||||
fonts, SMAA, zbic and the a64-kit tools) keep their own notices. Preserve them.
|
||||
|
||||
Thanks for testing, reporting and sending fixes. Every report from a different GPU and every area played further
|
||||
moves the port forward.
|
||||
|
||||
6
assets/triforce_assets/CREDITS.txt
Normal file
6
assets/triforce_assets/CREDITS.txt
Normal file
@@ -0,0 +1,6 @@
|
||||
Majora's Mask (majora_mask.bin, majora_alb.png, majora_emc.png)
|
||||
3D model "Majora's Mask" by PigeonMage, https://sketchfab.com/3d-models/majoras-mask-e9cfcdf9875846f6b6137f0156cd4eb2
|
||||
Licence: Creative Commons Attribution 4.0 (CC BY 4.0), https://creativecommons.org/licenses/by/4.0/
|
||||
Changes: the mesh placed on the Happy Mask Salesman's head bone, scaled, and its textures resized to 1024 x 1024.
|
||||
|
||||
The Happy Mask Salesman himself is not in these files: Triforce reads his model and animations from your own game files.
|
||||
BIN
assets/triforce_assets/majora_alb.png
Normal file
BIN
assets/triforce_assets/majora_alb.png
Normal file
Binary file not shown.
|
After Width: | Height: | Size: 1.4 MiB |
BIN
assets/triforce_assets/majora_emc.png
Normal file
BIN
assets/triforce_assets/majora_emc.png
Normal file
Binary file not shown.
|
After Width: | Height: | Size: 129 KiB |
BIN
assets/triforce_assets/majora_mask.bin
Normal file
BIN
assets/triforce_assets/majora_mask.bin
Normal file
Binary file not shown.
@@ -10,7 +10,21 @@ public partial class App : Application
|
||||
// Why it stopped is one "error <text>" line on stdout (the stream a launcher reads) and again on stderr
|
||||
protected override void OnStartup(StartupEventArgs e)
|
||||
{
|
||||
if (e.Args.Length > 0 && e.Args[0] == "--cli")
|
||||
var cli = e.Args.Length > 0 && e.Args[0] == "--cli";
|
||||
// ((Triforce) one builder at a time for its tools, Instance.cs: asked before any build or window)
|
||||
if (Tools.Find() is { } tools && !Instance.Take(tools.Root))
|
||||
{
|
||||
if (cli)
|
||||
{
|
||||
var said = "error " + Instance.Refusal(tools.Root).Replace("\n\n", " ");
|
||||
Console.Out.WriteLine(said);
|
||||
Console.Error.WriteLine(said);
|
||||
}
|
||||
else MessageBox.Show(Instance.Refusal(tools.Root), "Hookshot Builder", MessageBoxButton.OK, MessageBoxImage.Information);
|
||||
Environment.Exit(1);
|
||||
return;
|
||||
}
|
||||
if (cli)
|
||||
{
|
||||
Environment.Exit(Cli(e.Args.Skip(1).ToList()));
|
||||
return;
|
||||
|
||||
53
builder/HookshotBuilder/Instance.cs
Normal file
53
builder/HookshotBuilder/Instance.cs
Normal file
@@ -0,0 +1,53 @@
|
||||
using System;
|
||||
using System.IO;
|
||||
using System.Threading;
|
||||
|
||||
namespace HookshotBuilder;
|
||||
|
||||
/// (Triforce, handoff §5) One Hookshot Builder at a time for its build tools. Two of them on one tools folder ran make_game.py on
|
||||
/// the same work folder, and make_game.py stops a build it finds running there: the first one's build ended without a reason.
|
||||
/// The first builder holds a named mutex for as long as it runs (Windows lets go of it when the process ends, however it ends);
|
||||
/// a second one for the same tools says so and ends. The name comes from the tools folder (Tools.Root), so builders of other
|
||||
/// folders, Triforce Setup (its lock is named after its own folder) and the game (after its game folder) never meet it.
|
||||
/// HOOKSHOT_SINGLE_INSTANCE=0: no lock (as before). Plain logic: tests/test_builder_instance.py compiles this file alone
|
||||
static class Instance
|
||||
{
|
||||
static Mutex? held;
|
||||
|
||||
/// The lock's name for a tools folder: one for each spelling of it (case, '/' or '\', a separator at its end), a 64-bit FNV-1a
|
||||
/// of that spelling. Local\: this user's session
|
||||
public static string Name(string root)
|
||||
{
|
||||
try { root = Path.GetFullPath(root); } catch (Exception) { }
|
||||
var text = root.Replace('/', '\\').TrimEnd('\\').ToUpperInvariant();
|
||||
var h = 14695981039346656037UL; // (a name, not a secret)
|
||||
foreach (var c in text) { h ^= c; h *= 1099511628211UL; }
|
||||
return @"Local\HookshotBuilder-" + h.ToString("x16");
|
||||
}
|
||||
|
||||
/// True when this process now holds the tools folder (or did already; it waited up to waitMs for a builder that was ending),
|
||||
/// false when another builder does. Where no lock can be made at all nobody is refused
|
||||
public static bool Take(string root, int waitMs = 5000)
|
||||
{
|
||||
if (held != null) return true;
|
||||
if (Environment.GetEnvironmentVariable("HOOKSHOT_SINGLE_INSTANCE") == "0") return true;
|
||||
Mutex m;
|
||||
try { m = new Mutex(false, Name(root)); }
|
||||
catch (UnauthorizedAccessException) { return false; } // (it is there, and belongs to a builder run as administrator)
|
||||
catch (Exception) { return true; }
|
||||
try
|
||||
{
|
||||
if (!m.WaitOne(waitMs)) { m.Dispose(); return false; }
|
||||
}
|
||||
catch (AbandonedMutexException) { } // (its holder ended without letting go: Windows hands it on, and it is ours)
|
||||
catch (Exception) { m.Dispose(); return true; }
|
||||
held = m;
|
||||
return true;
|
||||
}
|
||||
|
||||
/// What a second builder says
|
||||
public static string Refusal(string root) =>
|
||||
"Hookshot Builder is already running for these build tools:\n\n" + root +
|
||||
"\n\nUse the window that is open, or close it and start the builder again. (Two builders on one tools folder would build " +
|
||||
"into the same work folder and stop each other's build.)";
|
||||
}
|
||||
@@ -952,6 +952,15 @@ public partial class MainWindow : Window
|
||||
clock.Stop();
|
||||
elapsed.Stop();
|
||||
runner = null;
|
||||
// ((Triforce, 11-18) a Cancel first, whatever the run said: Runner.Cancel asks the game's shader preparation to stop
|
||||
// and it ends well, so make_game.py goes on to its @@done. Taken as done, a Cancel in "Preparing shaders" said "Your
|
||||
// game is ready", made the shortcuts and freed the build's files, where it went back to the files, as before)
|
||||
if (failure == "cancelled")
|
||||
{
|
||||
SetStep(0);
|
||||
ShowPage(FilesPage);
|
||||
return;
|
||||
}
|
||||
if (finished.Length > 0)
|
||||
{
|
||||
foreach (var (id, r) in rows)
|
||||
@@ -964,12 +973,6 @@ public partial class MainWindow : Window
|
||||
ShowReady();
|
||||
return;
|
||||
}
|
||||
if (failure == "cancelled")
|
||||
{
|
||||
SetStep(0);
|
||||
ShowPage(FilesPage);
|
||||
return;
|
||||
}
|
||||
if (current != null && rows.TryGetValue(current, out var row)) row.Stopped();
|
||||
ShowProblem(failure ?? "The build tools stopped without saying why.", tools: failure?.Contains("antivirus", StringComparison.OrdinalIgnoreCase) == true);
|
||||
}
|
||||
@@ -1231,7 +1234,7 @@ public partial class MainWindow : Window
|
||||
void CloseWindow(object sender, RoutedEventArgs e)
|
||||
{
|
||||
if (runner != null && MessageBox.Show(this, T(S.CloseQuestion), "Hookshot", MessageBoxButton.YesNo) != MessageBoxResult.Yes) return;
|
||||
runner?.Cancel();
|
||||
runner?.Cancel().Wait(); // (the app ends with the window: a shader preparation asked to stop is waited for, or killed)
|
||||
Close();
|
||||
}
|
||||
}
|
||||
|
||||
123
builder/HookshotBuilder/Reuse.cs
Normal file
123
builder/HookshotBuilder/Reuse.cs
Normal file
@@ -0,0 +1,123 @@
|
||||
using System;
|
||||
using System.IO;
|
||||
using System.Linq;
|
||||
using System.Threading;
|
||||
|
||||
namespace HookshotBuilder;
|
||||
|
||||
/// (Triforce) What the setup (builder/setup/setup.c) lends from the installed version to the one it puts together, and the lock
|
||||
/// every setup holds while it does. An update moves the folders of parts that did not change (Python, the compiler...) from
|
||||
/// versions\<installed> into versions\<tag>.part and writes each down first in its reused.txt (the installed version's tag,
|
||||
/// then a folder a line); a cut-off setup leaves them there. Free gives them back (UndoEveryReuse, setup.c undo_every_reuse,
|
||||
/// the same rules) and never deletes a <tag>.part that still holds them (Lent): deleted, the installed version lost them for
|
||||
/// good. Plain logic: tests/test_builder_setup_lock.py compiles this file alone
|
||||
static class Reuse
|
||||
{
|
||||
public const string Note = "reused.txt";
|
||||
|
||||
/// True when folder is an unfinished version (a versions\*.part) that still holds folders of another version
|
||||
public static bool Lent(string folder) =>
|
||||
Path.GetFileName(folder.TrimEnd('\\', '/')).EndsWith(".part", StringComparison.OrdinalIgnoreCase) &&
|
||||
File.Exists(Path.Combine(folder, Note));
|
||||
|
||||
/// The setup's lock for a setup folder (setup.c setup_lock_name and setup_parts.lock_name, the same): a 64-bit FNV-1a of the
|
||||
/// folder's full path with '\' between its parts, none at its end, a to z in upper case. Local\: this user's session
|
||||
public static string LockName(string setupBase)
|
||||
{
|
||||
try { setupBase = Path.GetFullPath(setupBase); } catch (Exception) { }
|
||||
var text = setupBase.Replace('/', '\\').TrimEnd('\\');
|
||||
var h = 14695981039346656037UL; // (a name, not a secret)
|
||||
foreach (var c in text) { h ^= c >= 'a' && c <= 'z' ? (uint)(c - 32) : c; h *= 1099511628211UL; }
|
||||
return @"Local\HookshotBuilder-Setup-" + h.ToString("x16");
|
||||
}
|
||||
|
||||
/// The setup folder held by this thread until Dispose (one that holds nothing where no lock can be made at all), or null
|
||||
/// when a setup held it for waitMs: one is installing now
|
||||
public static Held? TakeLock(string setupBase, int waitMs = 0)
|
||||
{
|
||||
Mutex m;
|
||||
try { m = new Mutex(false, LockName(setupBase)); }
|
||||
catch (UnauthorizedAccessException) { return null; } // (it is there, and belongs to a setup run as administrator)
|
||||
catch (Exception) { return new Held(null); }
|
||||
try
|
||||
{
|
||||
if (!m.WaitOne(waitMs)) { m.Dispose(); return null; }
|
||||
}
|
||||
catch (AbandonedMutexException) { } // (its holder ended without letting go: Windows hands it on, and it is ours)
|
||||
catch (Exception) { m.Dispose(); return new Held(null); }
|
||||
return new Held(m);
|
||||
}
|
||||
|
||||
public sealed class Held : IDisposable
|
||||
{
|
||||
Mutex? m;
|
||||
internal Held(Mutex? m) => this.m = m;
|
||||
|
||||
public void Dispose()
|
||||
{
|
||||
if (m == null) return;
|
||||
try { m.ReleaseMutex(); } catch (Exception) { }
|
||||
m.Dispose();
|
||||
m = null;
|
||||
}
|
||||
}
|
||||
|
||||
/// The folders a setup moved from the installed version into install (a <tag>.part), moved back: each only where that version
|
||||
/// has none; the note goes once all went back. True: nothing is left to move back
|
||||
public static bool Undo(string versions, string install)
|
||||
{
|
||||
var note = Path.Combine(install, Note);
|
||||
string[] lines;
|
||||
try { lines = File.ReadAllLines(note).Where(l => l.Trim().Length > 0).ToArray(); }
|
||||
catch (Exception e) when (e is IOException or UnauthorizedAccessException) { return !File.Exists(note); }
|
||||
if (lines.Length == 0 || !NameOk(lines[0], false)) // (no version named: nothing was moved yet)
|
||||
{
|
||||
try { File.Delete(note); } catch (Exception e) when (e is IOException or UnauthorizedAccessException) { }
|
||||
return true;
|
||||
}
|
||||
var ok = true;
|
||||
foreach (var rel in lines.Skip(1))
|
||||
{
|
||||
if (!NameOk(rel, true)) continue;
|
||||
var from = Path.Combine(install, rel.Replace('/', '\\'));
|
||||
var back = Path.Combine(versions, lines[0], rel.Replace('/', '\\'));
|
||||
if (!(Directory.Exists(from) || File.Exists(from)) || Directory.Exists(back) || File.Exists(back)) continue;
|
||||
try
|
||||
{
|
||||
Directory.CreateDirectory(Path.GetDirectoryName(back)!);
|
||||
if (Directory.Exists(from)) Directory.Move(from, back);
|
||||
else File.Move(from, back);
|
||||
}
|
||||
catch (Exception e) when (e is IOException or UnauthorizedAccessException) { ok = false; }
|
||||
}
|
||||
if (ok)
|
||||
try { File.Delete(note); } catch (Exception e) when (e is IOException or UnauthorizedAccessException) { }
|
||||
return ok;
|
||||
}
|
||||
|
||||
/// Undo for every versions\*.part with a note; the first whose folders could not all go back (each is tried), null when none
|
||||
public static string? UndoEveryReuse(string versions)
|
||||
{
|
||||
string? left = null;
|
||||
string[] dirs;
|
||||
try { dirs = Directory.GetDirectories(versions); }
|
||||
catch (Exception e) when (e is IOException or UnauthorizedAccessException) { return null; }
|
||||
foreach (var d in dirs.OrderBy(d => d, StringComparer.Ordinal))
|
||||
if (Lent(d) && !Undo(versions, d)) left ??= d;
|
||||
return left;
|
||||
}
|
||||
|
||||
/// A name from the note (setup.c list_name_ok, the same): letters, digits and . _ + - only (with path, '/' or '\' between
|
||||
/// its parts), no "." or ".." part, no drive, no root
|
||||
static bool NameOk(string s, bool path)
|
||||
{
|
||||
if (s.Length == 0 || s[0] == '/' || s[0] == '\\') return false;
|
||||
foreach (var c in s)
|
||||
if (!(c is >= 'a' and <= 'z' or >= 'A' and <= 'Z' or >= '0' and <= '9' or '.' or '_' or '+' or '-' ||
|
||||
(path && (c == '/' || c == '\\'))))
|
||||
return false;
|
||||
foreach (var part in path ? s.Split('/', '\\') : new[] { s })
|
||||
if (part.Length == 0 || part == "." || part == "..") return false;
|
||||
return true;
|
||||
}
|
||||
}
|
||||
@@ -81,7 +81,12 @@ sealed class Runner
|
||||
this.ngxSdk = ngxSdk;
|
||||
}
|
||||
|
||||
static ProcessStartInfo Info(Tools tools, IEnumerable<string> args, int? workers = null, string? ngxSdk = null)
|
||||
/// (Triforce, 11-18) The event this run's game opens while it prepares its shaders (gpu/vk/vk_main.cpp PrepareStopWatch),
|
||||
/// named for this run alone: a Cancel here never stops another builder's preparation
|
||||
readonly string stopEvent = @"Local\TriforcePrepareStop." + Guid.NewGuid().ToString("N");
|
||||
|
||||
static ProcessStartInfo Info(Tools tools, IEnumerable<string> args, int? workers = null, string? ngxSdk = null,
|
||||
string? stopEvent = null)
|
||||
{
|
||||
var si = new ProcessStartInfo(tools.Python)
|
||||
{
|
||||
@@ -98,6 +103,9 @@ sealed class Runner
|
||||
// (NVIDIA DLSS: only when ticked. Without it, one set on the PC is not passed on either: a build is with DLSS only when asked)
|
||||
if (ngxSdk != null) si.Environment["HOOKSHOT_NGX_SDK"] = ngxSdk;
|
||||
else si.Environment.Remove("HOOKSHOT_NGX_SDK");
|
||||
// (not a HOOKSHOT_ name: make_game.py's game_env keeps it for the game)
|
||||
if (stopEvent != null) si.Environment["TRIFORCE_PREPARE_STOP"] = stopEvent;
|
||||
else si.Environment.Remove("TRIFORCE_PREPARE_STOP");
|
||||
// (the folder's own compiler, headers and libraries, when it has them: nothing of the PC's but Windows' own)
|
||||
var sets = new[] { ("EMSDK", "emsdk"), ("RX_CPM", "cpm"), ("WGPU_NATIVE", "wgpu-native") };
|
||||
foreach (var (name, folder) in sets)
|
||||
@@ -118,11 +126,16 @@ sealed class Runner
|
||||
{
|
||||
try
|
||||
{
|
||||
using var p = Process.Start(Info(tools, args, workers, ngxSdk))!;
|
||||
using var p = Process.Start(Info(tools, args, workers, ngxSdk, stopEvent))!;
|
||||
process = p;
|
||||
p.ErrorDataReceived += (_, e) => { if (e.Data != null) Line?.Invoke(e.Data); };
|
||||
p.BeginErrorReadLine();
|
||||
try { p.PriorityClass = ProcessPriorityClass.BelowNormal; } catch (Exception) { }
|
||||
// (Triforce, 11-22, the performance audit's OS-57, after Source A's) The build tools run at normal priority;
|
||||
// below normal, as every build was before, only with HOOKSHOT_LOW_PRIORITY=1. Source A's reason, a report
|
||||
// it did not measure: below normal, Windows kept the steps that are one process (the read, the lift) on a
|
||||
// hybrid Intel CPU's efficiency cores while the performance cores idled.
|
||||
if (Environment.GetEnvironmentVariable("HOOKSHOT_LOW_PRIORITY") == "1")
|
||||
try { p.PriorityClass = ProcessPriorityClass.BelowNormal; } catch (Exception) { }
|
||||
while (p.StandardOutput.ReadLine() is { } line)
|
||||
{
|
||||
if (line.StartsWith("@@error ")) error = line[8..];
|
||||
@@ -138,10 +151,26 @@ sealed class Runner
|
||||
});
|
||||
}
|
||||
|
||||
public void Cancel()
|
||||
/// (Triforce, 11-18) The game's --prepare-shaders run, when that is the step under way, is asked to stop first (the event
|
||||
/// gpu/vk/vk_main.cpp PrepareStopWatch waits on): it ends its driver's part cleanly and keeps what it made, where a kill
|
||||
/// used one of that part's three tries and left shaders out for good. Killed as before when it has not ended in 30 s,
|
||||
/// and at once when no preparation is under way. The wait is not on the caller's thread
|
||||
public Task Cancel()
|
||||
{
|
||||
cancelled = true;
|
||||
try { process?.Kill(entireProcessTree: true); } catch (Exception) { }
|
||||
var p = process;
|
||||
var asked = false;
|
||||
try
|
||||
{
|
||||
if (EventWaitHandle.TryOpenExisting(stopEvent, out var stop))
|
||||
using (stop) asked = stop.Set();
|
||||
}
|
||||
catch (Exception) { }
|
||||
return Task.Run(() =>
|
||||
{
|
||||
try { if (asked && p != null && p.WaitForExit(30000)) return; } catch (Exception) { }
|
||||
try { p?.Kill(entireProcessTree: true); } catch (Exception) { }
|
||||
});
|
||||
}
|
||||
|
||||
/// Is this the supported game and version? The answer in plain words.
|
||||
|
||||
@@ -75,17 +75,20 @@ static class Storage
|
||||
|
||||
/// What Free lets go. afterBuild: what goes once the game is ready, the compiled objects of the build (objcache, which make_game.py
|
||||
/// prunes to them) and their system headers' stamp kept, so that the next update compiles only what changed (issue 1287: every
|
||||
/// update compiled all ~600 files again, an hour and more). The Free up space button lets those go too
|
||||
static IEnumerable<string> Freeable(bool afterBuild = false)
|
||||
/// update compiled all ~600 files again, an hour and more). The Free up space button lets those go too. (Triforce) Never an
|
||||
/// unfinished version that still holds folders of the installed one (Reuse.Lent): deleted, the installed version lost its
|
||||
/// Python or its compiler for good. installing: a setup holds the folder now, and what it is putting together (downloads\,
|
||||
/// versions\) stays. Nothing is changed here: Measure counts it
|
||||
static IEnumerable<string> Freeable(bool afterBuild = false, bool installing = false)
|
||||
{
|
||||
foreach (var d in afterBuild ? new[] { "sa", "exefs", "probe" } : new[] { "objcache", "sa", "exefs", "probe" })
|
||||
yield return Path.Combine(Work, d);
|
||||
foreach (var d in new[] { "downloads", "update" })
|
||||
foreach (var d in installing ? new[] { "update" } : new[] { "downloads", "update" })
|
||||
yield return Path.Combine(Base, d);
|
||||
var versions = Path.Combine(Base, "versions");
|
||||
if (Directory.Exists(versions))
|
||||
if (Directory.Exists(versions) && !installing)
|
||||
foreach (var v in Directory.EnumerateDirectories(versions))
|
||||
if (!Kept(v)) yield return v;
|
||||
if (!Kept(v) && !Reuse.Lent(v)) yield return v;
|
||||
}
|
||||
|
||||
/// Bytes in the builder's folder and the build files folder, and in what Free lets go (an object hard-linked in both sa\build and
|
||||
@@ -103,12 +106,18 @@ static class Storage
|
||||
/// Lets go of what Measure counts as freeable; returns the bytes it could not delete (a file in use)
|
||||
public static long Free(bool afterBuild = false)
|
||||
{
|
||||
foreach (var d in Freeable(afterBuild)) Delete(d);
|
||||
// ((Triforce) what a cut-off setup moved from the installed version goes back first (Reuse.UndoEveryReuse), and then its
|
||||
// <tag>.part is freeable; one whose folders could not all go back stays. With the setup's folder held: a setup running now
|
||||
// is not undone under it, and its folders are left alone)
|
||||
using var setup = Reuse.TakeLock(Base);
|
||||
var installing = setup == null;
|
||||
if (!installing) Reuse.UndoEveryReuse(Path.Combine(Base, "versions"));
|
||||
foreach (var d in Freeable(afterBuild, installing)) Delete(d);
|
||||
if (!afterBuild) Delete(Path.Combine(Work, "system_headers.txt")); // (the stamp of the compiled objects' system headers goes with them)
|
||||
if (!afterBuild) return Measure().freeable;
|
||||
var seen = new HashSet<(uint, ulong)>(); // (what of the after-build set is still there: a file in use)
|
||||
long left = 0;
|
||||
foreach (var d in Freeable(afterBuild: true)) left += Size(d, seen);
|
||||
foreach (var d in Freeable(afterBuild: true, installing)) left += Size(d, seen);
|
||||
return left;
|
||||
}
|
||||
|
||||
|
||||
@@ -302,27 +302,85 @@ static class Updater
|
||||
using var response = await http.SendAsync(request, HttpCompletionOption.ResponseHeadersRead, token);
|
||||
if (!response.IsSuccessStatusCode) throw HttpFailure(response);
|
||||
|
||||
var total = response.Content.Headers.ContentLength ?? size;
|
||||
await using var from = await response.Content.ReadAsStreamAsync(token);
|
||||
await using var to = File.Create(path);
|
||||
var buffer = new byte[1 << 20];
|
||||
long have = 0;
|
||||
var shown = DateTime.MinValue;
|
||||
int n;
|
||||
while (true)
|
||||
// ((Triforce) the size the release lists, which the server must not contradict; the server's only where none is listed (the
|
||||
// server's went first, and a short or other file passed as whole). Room for it asked before the first byte, and the bytes
|
||||
// written to <file>.part, which becomes the file only once complete: a failure removes it)
|
||||
var told = response.Content.Headers.ContentLength;
|
||||
if (size > 0 && told is { } other && other != size)
|
||||
throw new IOException($"The server offers {other} bytes as {Path.GetFileName(path)}, and the release lists {size}. Nothing was downloaded.");
|
||||
var total = size > 0 ? size : told ?? 0;
|
||||
if (total > 0 && NoRoom(path, total) is { } none) throw new IOException(none);
|
||||
var part = path + ".part";
|
||||
try
|
||||
{
|
||||
try { n = await from.ReadAsync(buffer, token); }
|
||||
catch (IOException e) { throw new HttpRequestException("The download connection ended unexpectedly.", e); }
|
||||
if (n == 0) break;
|
||||
await to.WriteAsync(buffer.AsMemory(0, n), token);
|
||||
have += n;
|
||||
if (progress != null && total > 0 && DateTime.UtcNow - shown > TimeSpan.FromMilliseconds(200))
|
||||
await using (var from = await response.Content.ReadAsStreamAsync(token))
|
||||
await using (var to = File.Create(part))
|
||||
{
|
||||
shown = DateTime.UtcNow;
|
||||
progress(Math.Min(1.0, (double)have / total));
|
||||
var buffer = new byte[1 << 20];
|
||||
long have = 0;
|
||||
var shown = DateTime.MinValue;
|
||||
int n;
|
||||
while (true)
|
||||
{
|
||||
try { n = await from.ReadAsync(buffer, token); }
|
||||
catch (IOException e) { throw new HttpRequestException("The download connection ended unexpectedly.", e); }
|
||||
if (n == 0) break;
|
||||
await to.WriteAsync(buffer.AsMemory(0, n), token);
|
||||
have += n;
|
||||
if (total > 0 && have > total) throw new IOException($"The server sent more than the {total} bytes of {Path.GetFileName(path)}.");
|
||||
if (progress != null && total > 0 && DateTime.UtcNow - shown > TimeSpan.FromMilliseconds(200))
|
||||
{
|
||||
shown = DateTime.UtcNow;
|
||||
progress(Math.Min(1.0, (double)have / total));
|
||||
}
|
||||
}
|
||||
if (total > 0 && have != total) throw new HttpRequestException($"The download ended early ({have >> 20} of {total >> 20} MB).");
|
||||
}
|
||||
File.Move(part, path, true);
|
||||
}
|
||||
if (total > 0 && have != total) throw new HttpRequestException($"The download ended early ({have >> 20} of {total >> 20} MB).");
|
||||
catch (Exception)
|
||||
{
|
||||
try { File.Delete(part); } catch (Exception) { }
|
||||
throw;
|
||||
}
|
||||
}
|
||||
|
||||
/// ((Triforce) setup/Guard.cs Space, the same rule) Null when the disk of path has room for need bytes and 32 MB more (or
|
||||
/// cannot be asked); else the sentence that says how much it takes and how much is free
|
||||
static string? NoRoom(string path, long need)
|
||||
{
|
||||
long free;
|
||||
try { free = new DriveInfo(Path.GetPathRoot(Path.GetFullPath(path))!).AvailableFreeSpace; }
|
||||
catch (Exception) { return null; }
|
||||
need += 32L << 20;
|
||||
return free >= need ? null
|
||||
: $"There is no room for {Path.GetFileName(path)} on {Path.GetPathRoot(Path.GetFullPath(path))}: it takes {(need + (1 << 20) - 1) >> 20} MB there, and {free >> 20} MB are free. Make room and try again.";
|
||||
}
|
||||
|
||||
static readonly HashSet<string> Devices = new(StringComparer.OrdinalIgnoreCase) { "CON", "PRN", "AUX", "NUL", "CONIN$", "CONOUT$" };
|
||||
|
||||
/// ((Triforce) setup/Mods.cs Unsafe, the same rules) Why a zip's member must not be unpacked under its name, null when it may:
|
||||
/// a path from the disk's root or through "..", a colon (a drive's letter, a hidden stream), a name Windows keeps for a device
|
||||
/// (with or without an extension), a name that ends in a dot or a space (Windows drops them), a character Windows refuses
|
||||
internal static string? UnsafeName(string name)
|
||||
{
|
||||
var n = name.Replace('\\', '/');
|
||||
if (n.Trim().Length == 0) return "a file without a name";
|
||||
if (n[0] == '/') return "a path that starts at the disk's root";
|
||||
if (n.IndexOf(':') >= 0) return "a name with a colon or a drive's letter";
|
||||
foreach (var p in n.Split('/'))
|
||||
{
|
||||
if (p.Length == 0 || p == ".") continue;
|
||||
if (p == "..") return "a path that leaves its own folder (..)";
|
||||
if (p[p.Length - 1] == '.' || p[p.Length - 1] == ' ') return "a name that ends in a dot or a space, which Windows drops";
|
||||
var dot = p.IndexOf('.');
|
||||
var stem = (dot >= 0 ? p.Substring(0, dot) : p).TrimEnd(' ');
|
||||
if (Devices.Contains(stem) || stem.Length == 4 && (stem.StartsWith("COM", StringComparison.OrdinalIgnoreCase) || stem.StartsWith("LPT", StringComparison.OrdinalIgnoreCase)) &&
|
||||
(stem[3] is >= '1' and <= '9' || stem[3] == (char)0xB9 || stem[3] == (char)0xB2 || stem[3] == (char)0xB3))
|
||||
return "a name Windows keeps for a device (" + p + ")";
|
||||
if (p.Any(c => c < 32 || "<>\"|?*".IndexOf(c) >= 0)) return "a name with a character Windows does not allow";
|
||||
}
|
||||
return null;
|
||||
}
|
||||
|
||||
static HttpRequestException HttpFailure(HttpResponseMessage response) =>
|
||||
@@ -335,14 +393,41 @@ static class Updater
|
||||
error is HttpRequestException { StatusCode: { } status } ? $" (HTTP {(int)status})" : "";
|
||||
|
||||
/// The zip's HookshotBuilder folder, as target (an older one there replaced)
|
||||
/// ((Triforce) every name checked first (UnsafeName: such a zip is refused whole), a failure removes the part, and an older
|
||||
/// target is moved aside and back when the new one cannot take its place: before, it was deleted first, and a move that
|
||||
/// failed then left neither)
|
||||
static async Task<string> Unpack(string zip, string target)
|
||||
{
|
||||
var part = target + ".part";
|
||||
if (Directory.Exists(part)) Directory.Delete(part, true);
|
||||
await Task.Run(() => ZipFile.ExtractToDirectory(zip, part));
|
||||
try
|
||||
{
|
||||
await Task.Run(() =>
|
||||
{
|
||||
using (var z = ZipFile.OpenRead(zip))
|
||||
foreach (var e in z.Entries)
|
||||
if (UnsafeName(e.FullName) is { } why)
|
||||
throw new InvalidDataException($"The download holds {why} (\"{e.FullName}\"): it was not unpacked.");
|
||||
ZipFile.ExtractToDirectory(zip, part);
|
||||
});
|
||||
}
|
||||
catch (Exception)
|
||||
{
|
||||
try { if (Directory.Exists(part)) Directory.Delete(part, true); } catch (Exception) { }
|
||||
throw;
|
||||
}
|
||||
var inner = Path.Combine(part, "HookshotBuilder");
|
||||
if (Directory.Exists(target)) Directory.Delete(target, true);
|
||||
Directory.Move(Directory.Exists(inner) ? inner : part, target);
|
||||
var old = target + ".old";
|
||||
if (Directory.Exists(old)) Directory.Delete(old, true);
|
||||
if (Directory.Exists(target)) Directory.Move(target, old);
|
||||
try { Directory.Move(Directory.Exists(inner) ? inner : part, target); }
|
||||
catch (Exception)
|
||||
{
|
||||
try { if (!Directory.Exists(target) && Directory.Exists(old)) Directory.Move(old, target); } catch (Exception) { }
|
||||
try { if (Directory.Exists(part)) Directory.Delete(part, true); } catch (Exception) { }
|
||||
throw;
|
||||
}
|
||||
try { if (Directory.Exists(old)) Directory.Delete(old, true); } catch (Exception) { }
|
||||
if (Directory.Exists(part)) Directory.Delete(part, true);
|
||||
return target;
|
||||
}
|
||||
|
||||
@@ -11,6 +11,7 @@
|
||||
<application xmlns="urn:schemas-microsoft-com:asm.v3">
|
||||
<windowsSettings>
|
||||
<dpiAwareness xmlns="http://schemas.microsoft.com/SMI/2016/WindowsSettings">PerMonitorV2</dpiAwareness>
|
||||
<longPathAware xmlns="http://schemas.microsoft.com/SMI/2016/WindowsSettings">true</longPathAware>
|
||||
</windowsSettings>
|
||||
</application>
|
||||
</assembly>
|
||||
|
||||
@@ -1,7 +1,8 @@
|
||||
#!/bin/sh
|
||||
# The Hookshot AppImage: the builder window (builder/linux/hookshot_builder.py) with everything it builds the game with -
|
||||
# Python and the tools, clang and lld, a glibc 2.35 sysroot, FFmpeg, SDL2 and the Vulkan back end, built already.
|
||||
# Hookshot.AppImage the builder window (Play is in the window, once a game is built)
|
||||
# Hookshot.AppImage the launcher window (Play is in the window, once a game is built)
|
||||
# Hookshot.AppImage --classic the classic builder window instead
|
||||
# Hookshot.AppImage --play start the game the builder saved, without opening the window
|
||||
# Hookshot.AppImage --play <folder> start the game in that folder
|
||||
# Hookshot.AppImage <folder> the same
|
||||
@@ -35,4 +36,20 @@ fi
|
||||
if [ -n "${1:-}" ] && [ -d "$1" ]; then
|
||||
exec "$HOOKSHOT_TOOLS/builder/linux/hookshot-launch" "$1"
|
||||
fi
|
||||
# The window: the Triforce launcher (usr/bin/triforce-launcher, a WebKitGTK window). It exits 4 when this system has no
|
||||
# WebKitGTK 4.1 (e.g. some Steam Deck images), and the classic builder window opens instead; --classic or
|
||||
# TRIFORCE_CLASSIC=1 always opens the classic one.
|
||||
LAUNCHER="$HERE/usr/bin/triforce-launcher"
|
||||
if [ "${1:-}" = "--classic" ]; then
|
||||
shift
|
||||
TRIFORCE_CLASSIC=1
|
||||
fi
|
||||
if [ "${TRIFORCE_CLASSIC:-0}" != "1" ] && [ -x "$LAUNCHER" ]; then
|
||||
# (WebKitGTK's GPU path breaks on NVIDIA's driver and under gamescope; its plain path is fine for a launcher)
|
||||
export WEBKIT_DISABLE_DMABUF_RENDERER="${WEBKIT_DISABLE_DMABUF_RENDERER:-1}"
|
||||
export TRIFORCE_APPRUN="$HERE/AppRun"
|
||||
"$LAUNCHER" "$@"
|
||||
code=$?
|
||||
[ "$code" -ne 4 ] && exit "$code"
|
||||
fi
|
||||
exec "$PY" -I "$HOOKSHOT_TOOLS/builder/linux/hookshot_builder.py" "$@"
|
||||
|
||||
@@ -17,6 +17,15 @@ CREATE_NO_WINDOW = 0x08000000
|
||||
BELOW_NORMAL_PRIORITY_CLASS = 0x00004000
|
||||
|
||||
|
||||
def build_priority(env=None):
|
||||
"""(Triforce, 11-22, the performance audit's OS-57, after Source A's) The priority class the build tools are started
|
||||
with, as a creation flag: none (normal priority), and below normal, as every build was before, only with
|
||||
HOOKSHOT_LOW_PRIORITY=1. Source A's reason, a report it did not measure: below normal, Windows kept the steps that are
|
||||
one process (the read, the lift) on a hybrid Intel CPU's efficiency cores while the performance cores idled."""
|
||||
env = os.environ if env is None else env
|
||||
return BELOW_NORMAL_PRIORITY_CLASS if env.get("HOOKSHOT_LOW_PRIORITY") == "1" else 0
|
||||
|
||||
|
||||
# ---------------- the build tools (Runner.cs: Tools) ----------------
|
||||
|
||||
class Tools:
|
||||
@@ -96,7 +105,7 @@ class Runner:
|
||||
p = subprocess.Popen(build_command(self.tools, self.args), cwd=self.tools.root,
|
||||
env=build_env(self.tools, self.workers, self.ngx_sdk), stdout=subprocess.PIPE,
|
||||
stderr=subprocess.PIPE, stdin=subprocess.DEVNULL,
|
||||
creationflags=CREATE_NO_WINDOW | BELOW_NORMAL_PRIORITY_CLASS)
|
||||
creationflags=CREATE_NO_WINDOW | build_priority())
|
||||
except OSError as e:
|
||||
self.on_finished("The build tools could not start: %s" % e)
|
||||
return
|
||||
@@ -232,16 +241,21 @@ def _kept_version(d):
|
||||
return bool(cur) and _same(cur, d)
|
||||
|
||||
|
||||
def freeable(after_build=False):
|
||||
def freeable(after_build=False, installing=False):
|
||||
"""What free lets go (measure counts it; nothing is changed here). (Triforce) Never an unfinished version that still holds
|
||||
folders of the installed one (setup_parts.lent: its reused.txt): deleted, the installed version lost its Python or its
|
||||
compiler for good. installing: another install holds the setup's folder now, and what it is putting together
|
||||
(downloads\\, versions\\) stays"""
|
||||
from . import setup_parts
|
||||
for d in (("sa", "exefs", "probe") if after_build else ("objcache", "sa", "exefs", "probe")):
|
||||
yield os.path.join(work(), d)
|
||||
for d in ("downloads", "update"):
|
||||
for d in ("update",) if installing else ("downloads", "update"):
|
||||
yield os.path.join(paths.base(), d)
|
||||
v = paths.versions()
|
||||
if os.path.isdir(v):
|
||||
if os.path.isdir(v) and not installing:
|
||||
for name in os.listdir(v):
|
||||
p = os.path.join(v, name)
|
||||
if os.path.isdir(p) and not _kept_version(p):
|
||||
if os.path.isdir(p) and not _kept_version(p) and not setup_parts.lent(p):
|
||||
yield p
|
||||
|
||||
|
||||
@@ -275,13 +289,24 @@ def measure():
|
||||
|
||||
def free(after_build=False):
|
||||
"""Lets go of what measure counts as freeable; returns the bytes it could not delete (a file in use)"""
|
||||
for d in list(freeable(after_build)):
|
||||
_delete(d)
|
||||
if not after_build:
|
||||
_delete(os.path.join(work(), "system_headers.txt"))
|
||||
return measure()[1]
|
||||
seen = set()
|
||||
return sum(_size(d, seen) for d in freeable(True))
|
||||
from . import setup_parts
|
||||
# ((Triforce) what a cut-off install moved from the installed version goes back first (setup_parts.undo_every_reuse), and
|
||||
# then its <tag>.part is freeable; one whose folders could not all go back stays. With the setup's folder held: an install
|
||||
# running now is not undone under it, and its folders are left alone)
|
||||
lock = setup_parts.take_lock(0)
|
||||
try:
|
||||
if lock:
|
||||
setup_parts.undo_every_reuse()
|
||||
for d in list(freeable(after_build, installing=not lock)):
|
||||
_delete(d)
|
||||
if not after_build:
|
||||
_delete(os.path.join(work(), "system_headers.txt"))
|
||||
return measure()[1]
|
||||
seen = set()
|
||||
return sum(_size(d, seen) for d in freeable(True, installing=not lock))
|
||||
finally:
|
||||
if lock:
|
||||
lock.release()
|
||||
|
||||
|
||||
def gb(n):
|
||||
|
||||
@@ -22,12 +22,80 @@ from . import REPO, paths
|
||||
UA = "HookshotBuilder-Setup/2.0"
|
||||
WAITS = (3, 6, 12, 20, 30, 45, 60) # (8 tries over ~3 minutes: GitHub's release downloads answered 504 for hours on 2026-11-08)
|
||||
OWN = ("python", "emsdk", "cpm", "third_party", "wgpu-native")
|
||||
MARGIN = 32 << 20 # ((Triforce) what a disk keeps free beyond what is about to be written to it: setup/Guard.cs Space.Margin)
|
||||
REUSED = "reused.txt" # ((Triforce) in versions\<tag>.part: the folders moved over from the installed version, reuse_part)
|
||||
DEVICES = {"CON", "PRN", "AUX", "NUL", "CONIN$", "CONOUT$"}
|
||||
LOCK_WAIT = 20 # ((Triforce) seconds an install waits for another one into the same folder before it says so: setup.c SETUP_LOCK_WAIT_MS)
|
||||
|
||||
|
||||
class SetupError(Exception):
|
||||
pass
|
||||
|
||||
|
||||
def unsafe_member(name):
|
||||
"""(Triforce) Why a zip's member must not be unpacked under its name, None when it may (setup/Mods.cs Unsafe, the same
|
||||
rules): a path from the disk's root, a colon (a drive's letter, a hidden stream of another file), a path through "..", a
|
||||
name Windows keeps for a device (CON, NUL, COM1, lpt2.txt: with or without an extension), a name that ends in a dot or a
|
||||
space (Windows drops them, so the file would land under another name), a character Windows does not allow. zipfile's
|
||||
extractall drops some of these quietly and writes the file under another name; a device's name it writes to the device"""
|
||||
n = name.replace("\\", "/")
|
||||
if not n.strip():
|
||||
return "a member without a name"
|
||||
if n.startswith("/"):
|
||||
return "a path that starts at the disk's root"
|
||||
if ":" in n:
|
||||
return "a name with a colon or a drive's letter"
|
||||
for p in n.split("/"):
|
||||
if p in ("", "."):
|
||||
continue
|
||||
if p == "..":
|
||||
return "a path that leaves its own folder (..)"
|
||||
if p[-1] in ". ":
|
||||
return "a name that ends in a dot or a space, which Windows drops"
|
||||
stem = p.split(".", 1)[0].rstrip(" ").upper()
|
||||
# (COM1 to COM9 and LPT1 to LPT9; Windows takes the raised digits 1, 2 and 3 for digits there too)
|
||||
if stem in DEVICES or (len(stem) == 4 and stem[:3] in ("COM", "LPT") and stem[3] in "123456789\xb9\xb2\xb3"):
|
||||
return "a name Windows keeps for a device (%s)" % p
|
||||
if any(ord(c) < 32 or c in '<>"|?*' for c in p):
|
||||
return "a name with a character Windows does not allow"
|
||||
return None
|
||||
|
||||
|
||||
def check_zip(z, dest, what):
|
||||
"""(Triforce) Every member's name checked (unsafe_member) and the room for all of it asked before the first file is
|
||||
written: such a zip is refused whole"""
|
||||
need = 0
|
||||
for info in z.infolist():
|
||||
why = unsafe_member(info.filename)
|
||||
if why:
|
||||
raise SetupError("%s holds %s (\"%s\"): it was not unpacked." % (what, why, info.filename[:120]))
|
||||
need += info.file_size + 4096 # (a file takes whole clusters)
|
||||
room(dest, need, what)
|
||||
|
||||
|
||||
def free_bytes(folder):
|
||||
"""The free bytes of the disk that holds folder (asked of the nearest folder above it that is there); None when unknown"""
|
||||
d = os.path.abspath(folder)
|
||||
while not os.path.isdir(d):
|
||||
up = os.path.dirname(d)
|
||||
if up == d:
|
||||
return None
|
||||
d = up
|
||||
try:
|
||||
return shutil.disk_usage(d).free
|
||||
except OSError:
|
||||
return None
|
||||
|
||||
|
||||
def room(folder, need, what):
|
||||
"""(Triforce) SetupError before anything is written when the disk of folder has no room for need bytes and the margin"""
|
||||
free = free_bytes(folder)
|
||||
if free is not None and free < need + MARGIN:
|
||||
raise SetupError("There is no room for %s on %s: it takes %d MB there, and %d MB are free. Make room and try again."
|
||||
% (what, os.path.splitdrive(os.path.abspath(folder))[0] or folder, (need + MARGIN + (1 << 20) - 1) >> 20,
|
||||
free >> 20))
|
||||
|
||||
|
||||
class Part:
|
||||
def __init__(self, sha, size, dest, frm, to, url):
|
||||
self.sha, self.size, self.dest, self.frm, self.to, self.url = sha, int(size), dest, frm, to, url
|
||||
@@ -182,6 +250,17 @@ def reuse_part(tag, p, install):
|
||||
if not os.path.exists(src):
|
||||
return False
|
||||
os.makedirs(os.path.dirname(dst), exist_ok=True)
|
||||
# ((Triforce) written down before it moves: the installed version is the rollback point, and a cancelled or failed install
|
||||
# moves it back (undo_reuse). Before, a cancel left the installed version without the folders it lent and nothing ran)
|
||||
try:
|
||||
journal = os.path.join(install, REUSED)
|
||||
new = not os.path.isfile(journal)
|
||||
with open(journal, "a", encoding="utf-8") as f:
|
||||
f.write(("%s\n" % was if new else "") + rel + "\n")
|
||||
f.flush()
|
||||
os.fsync(f.fileno())
|
||||
except OSError:
|
||||
return False
|
||||
try:
|
||||
os.replace(src, dst)
|
||||
return True
|
||||
@@ -189,6 +268,139 @@ def reuse_part(tag, p, install):
|
||||
return False
|
||||
|
||||
|
||||
def undo_reuse(work):
|
||||
"""(Triforce) The folders reuse_part moved from the installed version into the unfinished one (work), moved back: what a
|
||||
cancelled, failed or cut-off install leaves is the version that was installed, whole. Each goes back only where the
|
||||
installed version has none; the list is removed once all went back"""
|
||||
journal = os.path.join(work, REUSED)
|
||||
try:
|
||||
with open(journal, encoding="utf-8") as f:
|
||||
lines = [l.rstrip("\n") for l in f if l.strip()]
|
||||
except OSError:
|
||||
return True
|
||||
if not lines:
|
||||
return True
|
||||
was, ok = lines[0], True
|
||||
for rel in lines[1:]:
|
||||
if os.path.isabs(rel) or ".." in rel.replace("\\", "/").split("/"):
|
||||
continue
|
||||
src, back = os.path.join(work, rel), os.path.join(paths.versions(), was, rel)
|
||||
if not os.path.exists(src) or os.path.exists(back):
|
||||
continue
|
||||
try:
|
||||
os.makedirs(os.path.dirname(back), exist_ok=True)
|
||||
os.replace(src, back)
|
||||
except OSError:
|
||||
ok = False
|
||||
if ok:
|
||||
try:
|
||||
os.remove(journal)
|
||||
except OSError:
|
||||
pass
|
||||
return ok
|
||||
|
||||
|
||||
def undo_every_reuse():
|
||||
"""(Triforce) undo_reuse for every unfinished version (versions\\*.part) that holds a reused.txt. Before, only the <tag>.part
|
||||
of the release being installed now was undone: an install of another release cut off before left the installed version's
|
||||
folders in its own <tag>.part, and the installed version ran without them. Returns the first such folder whose folders
|
||||
could not all go back (each is tried), None when nothing is left to move back"""
|
||||
v, left = paths.versions(), None
|
||||
try:
|
||||
names = sorted(os.listdir(v))
|
||||
except OSError:
|
||||
return None
|
||||
for name in names:
|
||||
work = os.path.join(v, name)
|
||||
if name.lower().endswith(".part") and os.path.isfile(os.path.join(work, REUSED)) and not undo_reuse(work):
|
||||
left = left or work
|
||||
return left
|
||||
|
||||
|
||||
def lent(folder):
|
||||
"""(Triforce) True when folder is an unfinished version (a versions\\*.part) that still holds folders of another version
|
||||
(its reused.txt): only undo_reuse may let it go, never a delete (core.freeable)"""
|
||||
return os.path.basename(folder.rstrip("\\/")).lower().endswith(".part") and os.path.isfile(os.path.join(folder, REUSED))
|
||||
|
||||
|
||||
def lock_name(base=None):
|
||||
"""(Triforce) The name of the lock every install into base holds (setup.c setup_lock_name and Reuse.cs LockName, the same):
|
||||
Local\\HookshotBuilder-Setup-<a 64-bit FNV-1a of the folder's full path with '\\' between its parts, none at its end, a to z
|
||||
in upper case, taken a UTF-16 unit at a time>. Local\\: this user's session"""
|
||||
full = os.path.abspath(base or paths.base()).replace("/", "\\").rstrip("\\")
|
||||
data, h = full.encode("utf-16-le", "surrogatepass"), 14695981039346656037 # (a name, not a secret)
|
||||
for i in range(0, len(data), 2):
|
||||
c = data[i] | data[i + 1] << 8
|
||||
h = ((h ^ (c - 32 if 97 <= c <= 122 else c)) * 1099511628211) & 0xFFFFFFFFFFFFFFFF
|
||||
return "Local\\HookshotBuilder-Setup-%016x" % h
|
||||
|
||||
|
||||
class Lock:
|
||||
"""(Triforce) The lock of the setup's folder, held (release() lets go of it; one that holds nothing where no lock could be
|
||||
made at all). Windows' named mutex belongs to the thread that took it: released on that thread"""
|
||||
|
||||
def __init__(self, handle=None, file=None):
|
||||
self.handle, self.file = handle, file
|
||||
|
||||
def release(self):
|
||||
if self.handle:
|
||||
k = ctypes_kernel32()
|
||||
k.ReleaseMutex(self.handle)
|
||||
k.CloseHandle(self.handle)
|
||||
self.handle = None
|
||||
if self.file:
|
||||
self.file.close() # (the flock goes with the file)
|
||||
self.file = None
|
||||
|
||||
|
||||
def ctypes_kernel32():
|
||||
import ctypes
|
||||
from ctypes import wintypes
|
||||
k = ctypes.WinDLL("kernel32", use_last_error=True)
|
||||
k.CreateMutexW.argtypes = [ctypes.c_void_p, wintypes.BOOL, wintypes.LPCWSTR]
|
||||
k.CreateMutexW.restype = wintypes.HANDLE
|
||||
k.WaitForSingleObject.argtypes = [wintypes.HANDLE, wintypes.DWORD]
|
||||
k.WaitForSingleObject.restype = wintypes.DWORD
|
||||
k.ReleaseMutex.argtypes = [wintypes.HANDLE]
|
||||
k.CloseHandle.argtypes = [wintypes.HANDLE]
|
||||
return k
|
||||
|
||||
|
||||
def take_lock(wait=LOCK_WAIT, base=None):
|
||||
"""(Triforce) The setup's folder for this install alone: Lock, or None when another install (this app's, or setup.c's)
|
||||
held it for wait seconds. Before, two installs into one folder ran at once, and one's sweep (undo_every_reuse) gave back
|
||||
the folders the other had just moved into its <tag>.part: that one then finished without its Python. Windows lets go of
|
||||
the mutex when its holder ends, however it ends; elsewhere a lock file (setup.lock in the folder) is flock'ed. Where no
|
||||
lock can be made at all nobody is refused"""
|
||||
base = base or paths.base()
|
||||
if os.name == "nt":
|
||||
k = ctypes_kernel32()
|
||||
h = k.CreateMutexW(None, False, lock_name(base))
|
||||
if not h: # (5: it is there, and belongs to a setup run as administrator)
|
||||
import ctypes
|
||||
return None if ctypes.get_last_error() == 5 else Lock()
|
||||
if k.WaitForSingleObject(h, int(max(0, wait) * 1000)) in (0, 0x80): # (0x80: its holder ended without letting go: ours)
|
||||
return Lock(handle=h)
|
||||
k.CloseHandle(h)
|
||||
return None
|
||||
try:
|
||||
import fcntl
|
||||
os.makedirs(base, exist_ok=True)
|
||||
f = open(os.path.join(base, "setup.lock"), "a")
|
||||
except (ImportError, OSError):
|
||||
return Lock()
|
||||
end = time.monotonic() + max(0, wait)
|
||||
while True:
|
||||
try:
|
||||
fcntl.flock(f, fcntl.LOCK_EX | fcntl.LOCK_NB)
|
||||
return Lock(file=f)
|
||||
except OSError:
|
||||
if time.monotonic() >= end:
|
||||
f.close()
|
||||
return None
|
||||
time.sleep(0.2)
|
||||
|
||||
|
||||
def rmtree(p):
|
||||
shutil.rmtree(p, ignore_errors=True)
|
||||
|
||||
@@ -207,6 +419,29 @@ def current():
|
||||
|
||||
def install(status, progress=lambda f: None, tag=None, repair=False, cancel=None):
|
||||
"""Installs the release (the latest, or tag) unless it is there already; returns (tag, its folder)"""
|
||||
# ((Triforce) one install at a time into the setup's folder, its sweep included (take_lock): another one waits for it, a while)
|
||||
lock = take_lock(0)
|
||||
if lock is None:
|
||||
status("Another Hookshot Builder setup is installing. Waiting for it to finish...")
|
||||
lock = take_lock(LOCK_WAIT)
|
||||
if lock is None:
|
||||
raise SetupError("Another Hookshot Builder setup is installing into %s right now. Wait for it to finish (or close it), "
|
||||
"then try again." % paths.base())
|
||||
try:
|
||||
return install_locked(status, progress, tag, repair, cancel)
|
||||
finally:
|
||||
lock.release()
|
||||
|
||||
|
||||
def install_locked(status, progress, tag, repair, cancel):
|
||||
"""install's work, with the setup's folder held"""
|
||||
# ((Triforce) an install cut off before (the app closed), of this release or another: what it moved over from the installed
|
||||
# version goes back first, before the installed version is looked at, and a part folder whose folders could not all go back
|
||||
# is kept, not deleted with them)
|
||||
left = undo_every_reuse()
|
||||
if left:
|
||||
raise SetupError("The last install could not be undone: %s still holds folders of the installed version. Close every "
|
||||
"Hookshot window and try again." % left)
|
||||
tag, parts = read_manifest(status, tag, cancel)
|
||||
versions, downloads = paths.versions(), os.path.join(paths.base(), "downloads")
|
||||
final = os.path.join(versions, tag)
|
||||
@@ -216,39 +451,23 @@ def install(status, progress=lambda f: None, tag=None, repair=False, cancel=None
|
||||
rmtree(final)
|
||||
os.makedirs(work, exist_ok=True)
|
||||
os.makedirs(downloads, exist_ok=True)
|
||||
n = len(parts)
|
||||
for i, p in enumerate(parts, 1):
|
||||
if cancel and cancel():
|
||||
raise SetupError("cancelled")
|
||||
if reuse_part(tag, p, work):
|
||||
status("Part %d of %d unchanged: kept from the installed version" % (i, n))
|
||||
continue
|
||||
zpath = os.path.join(downloads, p.sha + ".zip")
|
||||
if not (os.path.isfile(zpath) and sha256_file(zpath) == p.sha):
|
||||
download_part(status, progress, p, i, n, zpath, cancel)
|
||||
status("Unpacking part %d of %d..." % (i, n))
|
||||
progress(-1)
|
||||
dest = work if p.dest == "." else os.path.join(work, p.dest.replace("/", os.sep))
|
||||
os.makedirs(dest, exist_ok=True)
|
||||
try:
|
||||
with zipfile.ZipFile(zpath) as z:
|
||||
z.extractall(dest)
|
||||
except (OSError, zipfile.BadZipFile) as e:
|
||||
raise SetupError("Unpacking part %d of %d failed: %s" % (i, n, e)) from None
|
||||
if p.frm != "-" and p.to != "-": # (llvm-mingw-<version>-ucrt-x86_64 -> llvm-mingw)
|
||||
src, dst = os.path.join(dest, p.frm), os.path.join(dest, p.to)
|
||||
if not os.path.exists(dst):
|
||||
try:
|
||||
os.replace(src, dst)
|
||||
except OSError:
|
||||
raise SetupError("Could not set up part %d of %d (renaming %s)." % (i, n, p.frm)) from None
|
||||
if not os.path.isfile(os.path.join(work, "HookshotBuilder", "tools", "nx2", "make_game.py")):
|
||||
raise SetupError("The build tools are missing after unpacking. Your antivirus may have removed them: allow the folder %s "
|
||||
"and try again." % paths.base())
|
||||
try:
|
||||
unpack_parts(status, progress, tag, parts, work, downloads, cancel)
|
||||
except BaseException:
|
||||
# ((Triforce) a cancel or a failure: the installed version gets back what was moved from it, and the partial goes)
|
||||
if undo_reuse(work):
|
||||
rmtree(work)
|
||||
raise
|
||||
try:
|
||||
os.replace(work, final)
|
||||
except OSError:
|
||||
if undo_reuse(work):
|
||||
rmtree(work)
|
||||
raise SetupError("Could not finish installing into %s." % final) from None
|
||||
try:
|
||||
os.remove(os.path.join(final, REUSED))
|
||||
except OSError:
|
||||
pass
|
||||
paths.write_small(os.path.join(final, "installed.txt"), tag)
|
||||
write_parts(final, parts)
|
||||
elif not os.path.isfile(os.path.join(final, "parts.txt")):
|
||||
@@ -258,8 +477,45 @@ def install(status, progress=lambda f: None, tag=None, repair=False, cancel=None
|
||||
return tag, final
|
||||
|
||||
|
||||
def unpack_parts(status, progress, tag, parts, work, downloads, cancel):
|
||||
"""Each part into work: moved over from the installed version, or downloaded (when not there already), checked and unpacked"""
|
||||
n = len(parts)
|
||||
for i, p in enumerate(parts, 1):
|
||||
if cancel and cancel():
|
||||
raise SetupError("cancelled")
|
||||
if reuse_part(tag, p, work):
|
||||
status("Part %d of %d unchanged: kept from the installed version" % (i, n))
|
||||
continue
|
||||
zpath = os.path.join(downloads, p.sha + ".zip")
|
||||
if not (os.path.isfile(zpath) and sha256_file(zpath) == p.sha):
|
||||
download_part(status, progress, p, i, n, zpath, cancel)
|
||||
status("Unpacking part %d of %d..." % (i, n))
|
||||
progress(-1)
|
||||
dest = work if p.dest == "." else os.path.join(work, p.dest.replace("/", os.sep))
|
||||
os.makedirs(dest, exist_ok=True)
|
||||
try:
|
||||
with zipfile.ZipFile(zpath) as z:
|
||||
check_zip(z, dest, "Part %d of %d" % (i, n)) # ((Triforce) every name, and the room, before the first file)
|
||||
z.extractall(dest)
|
||||
except (OSError, zipfile.BadZipFile) as e:
|
||||
raise SetupError("Unpacking part %d of %d failed: %s" % (i, n, e)) from None
|
||||
if p.frm != "-" and p.to != "-": # (llvm-mingw-<version>-ucrt-x86_64 -> llvm-mingw)
|
||||
src, dst = os.path.join(dest, p.frm), os.path.join(dest, p.to)
|
||||
if not os.path.exists(dst):
|
||||
try:
|
||||
os.replace(src, dst)
|
||||
except OSError:
|
||||
raise SetupError("Could not set up part %d of %d (renaming %s)." % (i, n, p.frm)) from None
|
||||
if not os.path.isfile(os.path.join(work, "HookshotBuilder", "tools", "nx2", "make_game.py")):
|
||||
raise SetupError("The build tools are missing after unpacking. Your antivirus may have removed them: allow the folder %s "
|
||||
"and try again." % paths.base())
|
||||
|
||||
|
||||
def download_part(status, progress, p, i, n, zpath, cancel):
|
||||
http, part = Http(), zpath + ".part"
|
||||
# ((Triforce) room for the size the list names, asked before the first byte: a full disk stopped the download half way with
|
||||
# Windows' own words for it. The part is written from its first byte at each try, so no address or mark is kept beside it)
|
||||
room(os.path.dirname(zpath), p.size, "part %d of %d" % (i, n))
|
||||
for attempt in range(len(WAITS) + 1):
|
||||
if attempt:
|
||||
status("Part %d of %d: GitHub didn't answer (%s). Trying again in %d s..." % (i, n, http.status or "no answer",
|
||||
|
||||
@@ -89,12 +89,21 @@ static void status(const wchar_t *fmt, ...) {
|
||||
PostMessageW(g_win, WM_STATUS, 0, 0);
|
||||
}
|
||||
|
||||
/* (Triforce) the unfinished version's folder while one is put together, and the versions folder: what fail() undoes (undo_reuse) */
|
||||
static wchar_t g_part_dir[MAX_PATH], g_versions_dir[MAX_PATH];
|
||||
static int undo_reuse(const wchar_t *versions, const wchar_t *install);
|
||||
static void rmtree(const wchar_t *dir);
|
||||
static void setup_lock_release(void);
|
||||
|
||||
static void fail(const wchar_t *fmt, ...) {
|
||||
wchar_t msg[1024];
|
||||
va_list ap;
|
||||
va_start(ap, fmt);
|
||||
vswprintf(msg, 1024, fmt, ap);
|
||||
va_end(ap);
|
||||
/* ((Triforce) the installed version gets back the folders that were moved from it, and the unfinished one goes) */
|
||||
if (g_part_dir[0] && undo_reuse(g_versions_dir, g_part_dir)) rmtree(g_part_dir);
|
||||
setup_lock_release(); /* ((Triforce) undone: the next setup may begin) */
|
||||
if (g_headless) say_error(msg), ExitProcess(1);
|
||||
MessageBoxW(g_win, msg, L"Hookshot Builder setup", MB_OK | MB_ICONERROR);
|
||||
PostMessageW(g_win, WM_DONE, 1, 0);
|
||||
@@ -318,6 +327,29 @@ static int part_folder(const Part *p, wchar_t *rel, int n) {
|
||||
return 1;
|
||||
}
|
||||
|
||||
#define REUSED_NOTE L"reused.txt"
|
||||
/* ((Triforce) undo_reuse's note: rel added to <install>\reused.txt, the installed version's tag first when it is new; written
|
||||
through to the disk. 0: it could not be written (the folder is then not moved) */
|
||||
static int reuse_note(const wchar_t *install, const char *was, const wchar_t *rel) {
|
||||
wchar_t note[MAX_PATH];
|
||||
char line[2 * MAX_PATH + 160];
|
||||
swprintf(note, MAX_PATH, L"%ls\\" REUSED_NOTE, install);
|
||||
const int fresh = GetFileAttributesW(note) == INVALID_FILE_ATTRIBUTES;
|
||||
HANDLE f = CreateFileW(note, FILE_APPEND_DATA, 0, NULL, OPEN_ALWAYS, FILE_FLAG_WRITE_THROUGH, NULL);
|
||||
if (f == INVALID_HANDLE_VALUE) return 0;
|
||||
int n = fresh ? snprintf(line, sizeof line, "%s\n", was) : 0;
|
||||
const int k = WideCharToMultiByte(CP_UTF8, 0, rel, -1, line + n, (int)sizeof line - n - 2, NULL, NULL);
|
||||
DWORD put = 0;
|
||||
int ok = k > 0;
|
||||
if (ok) {
|
||||
n += k - 1;
|
||||
line[n++] = '\n';
|
||||
ok = WriteFile(f, line, (DWORD)n, &put, NULL) && put == (DWORD)n && FlushFileBuffers(f);
|
||||
}
|
||||
CloseHandle(f);
|
||||
return ok;
|
||||
}
|
||||
|
||||
/* (players asked for less in %LOCALAPPDATA%: the downloaded zips are not kept beside the unpacked version any more. So that an update
|
||||
still downloads only what changed, a part whose checksum the installed version lists in its parts.txt (and whose folder it has)
|
||||
is moved over from it instead of downloaded: that version is deleted after the install anyway) */
|
||||
@@ -343,9 +375,106 @@ static int reuse_part(const wchar_t *versions, const char *tag, const Part *p, c
|
||||
wcscpy(parent, to);
|
||||
wchar_t *slash = wcsrchr(parent, L'\\');
|
||||
if (slash) *slash = 0, mkdirs(parent);
|
||||
/* ((Triforce) written down before it moves: the installed version is the rollback point, undo_reuse) */
|
||||
if (!reuse_note(install, was, rel)) return 0;
|
||||
return MoveFileW(from, to) != 0;
|
||||
}
|
||||
|
||||
/* (Triforce) The folders reuse_part moved from the installed version into the unfinished one (install), moved back: what a
|
||||
failed, closed or cut-off setup leaves is the version that was installed, whole. Before, they stayed in <tag>.part, which
|
||||
the next run deleted with them, and the version still installed had lost them. Each goes back only where the installed
|
||||
version has none; the note (<tag>.part\reused.txt: the installed version's tag, then a folder a line) is removed once all
|
||||
went back. 1: nothing is left to move back */
|
||||
static int undo_reuse(const wchar_t *versions, const wchar_t *install) {
|
||||
wchar_t note[MAX_PATH];
|
||||
swprintf(note, MAX_PATH, L"%ls\\" REUSED_NOTE, install);
|
||||
HANDLE f = CreateFileW(note, GENERIC_READ, FILE_SHARE_READ, NULL, OPEN_EXISTING, 0, NULL);
|
||||
if (f == INVALID_HANDLE_VALUE) return 1;
|
||||
char buf[8192];
|
||||
DWORD got = 0;
|
||||
ReadFile(f, buf, sizeof buf - 1, &got, NULL);
|
||||
CloseHandle(f);
|
||||
buf[got] = 0;
|
||||
char *next = buf, *was = next_line(&next), *rel;
|
||||
int ok = 1;
|
||||
if (!was || !list_name_ok(was, 0)) return DeleteFileW(note), 1; /* (no version named: nothing was moved yet) */
|
||||
while ((rel = next_line(&next)) != NULL) {
|
||||
if (!list_name_ok(rel, 1)) continue;
|
||||
wchar_t wrel[MAX_PATH], from[MAX_PATH], back[MAX_PATH], parent[MAX_PATH];
|
||||
MultiByteToWideChar(CP_UTF8, 0, rel, -1, wrel, MAX_PATH);
|
||||
swprintf(from, MAX_PATH, L"%ls\\%ls", install, wrel);
|
||||
swprintf(back, MAX_PATH, L"%ls\\%hs\\%ls", versions, was, wrel);
|
||||
if (GetFileAttributesW(from) == INVALID_FILE_ATTRIBUTES || GetFileAttributesW(back) != INVALID_FILE_ATTRIBUTES) continue;
|
||||
wcscpy(parent, back);
|
||||
wchar_t *slash = wcsrchr(parent, L'\\');
|
||||
if (slash) *slash = 0, mkdirs(parent);
|
||||
if (!MoveFileW(from, back)) ok = 0;
|
||||
}
|
||||
if (ok) DeleteFileW(note);
|
||||
return ok;
|
||||
}
|
||||
|
||||
/* (Triforce) undo_reuse for every unfinished version (versions\*.part) that holds a note. Before, only the <tag>.part of the
|
||||
release being set up now was undone: a setup of another release cut off before left the installed version's folders in its
|
||||
own <tag>.part, and the installed version ran without them. Each is tried; 1: nothing is left to move back, 0: left names the
|
||||
first folder whose folders could not all go back */
|
||||
static int undo_every_reuse(const wchar_t *versions, wchar_t *left, size_t cap) {
|
||||
wchar_t pattern[MAX_PATH], dir[MAX_PATH], note[MAX_PATH];
|
||||
WIN32_FIND_DATAW fd;
|
||||
swprintf(pattern, MAX_PATH, L"%ls\\*", versions);
|
||||
HANDLE h = FindFirstFileW(pattern, &fd);
|
||||
if (h == INVALID_HANDLE_VALUE) return 1;
|
||||
int ok = 1;
|
||||
do {
|
||||
const size_t n = wcslen(fd.cFileName);
|
||||
if (!(fd.dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY) || n < 6 || _wcsicmp(fd.cFileName + n - 5, L".part")) continue;
|
||||
swprintf(dir, MAX_PATH, L"%ls\\%ls", versions, fd.cFileName);
|
||||
swprintf(note, MAX_PATH, L"%ls\\" REUSED_NOTE, dir);
|
||||
if (GetFileAttributesW(note) == INVALID_FILE_ATTRIBUTES || undo_reuse(versions, dir)) continue;
|
||||
if (ok) wcsncpy(left, dir, cap - 1), left[cap - 1] = 0;
|
||||
ok = 0;
|
||||
} while (FindNextFileW(h, &fd));
|
||||
FindClose(h);
|
||||
return ok;
|
||||
}
|
||||
|
||||
/* (Triforce) One setup at a time for a setup folder (base), its sweep (undo_every_reuse) included. Two at once each undid the
|
||||
other's <tag>.part: the second one's sweep gave the installed version back the folders the first had just moved over, and
|
||||
the first finished without its Python. The lock is a named mutex (Windows lets go of it when its holder ends, however it
|
||||
ends), named as setup_parts.lock_name and Reuse.cs LockName name it: Local\HookshotBuilder-Setup-<a 64-bit FNV-1a of the
|
||||
folder's full path with '\' between its parts, none at its end, a to z in upper case>. Local\: this user's session */
|
||||
#define SETUP_LOCK_WAIT_MS 20000
|
||||
static HANDLE g_setup_lock;
|
||||
static void setup_lock_name(const wchar_t *base, wchar_t *name, size_t cap) {
|
||||
wchar_t full[2 * MAX_PATH];
|
||||
const DWORD got = GetFullPathNameW(base, 2 * MAX_PATH, full, NULL);
|
||||
if (!got || got >= 2 * MAX_PATH) wcsncpy(full, base, 2 * MAX_PATH - 1), full[2 * MAX_PATH - 1] = 0;
|
||||
size_t n = wcslen(full);
|
||||
for (size_t i = 0; i < n; i++) if (full[i] == L'/') full[i] = L'\\';
|
||||
while (n && full[n - 1] == L'\\') full[--n] = 0;
|
||||
unsigned long long h = 14695981039346656037ULL; /* (a name, not a secret) */
|
||||
for (size_t i = 0; i < n; i++) {
|
||||
const unsigned c = full[i] >= L'a' && full[i] <= L'z' ? full[i] - 32u : full[i];
|
||||
h = (h ^ c) * 1099511628211ULL;
|
||||
}
|
||||
swprintf(name, cap, L"Local\\HookshotBuilder-Setup-%016llx", h);
|
||||
}
|
||||
/* 1: this setup holds base (or no lock can be made at all: nobody is refused), 0: another setup held it for wait_ms */
|
||||
static int setup_lock_take(const wchar_t *base, DWORD wait_ms) {
|
||||
wchar_t name[96];
|
||||
if (g_setup_lock) return 1;
|
||||
setup_lock_name(base, name, 96);
|
||||
HANDLE m = CreateMutexW(NULL, FALSE, name);
|
||||
if (!m) return GetLastError() != ERROR_ACCESS_DENIED; /* (it is there, and belongs to a setup run as administrator) */
|
||||
const DWORD r = WaitForSingleObject(m, wait_ms);
|
||||
if (r == WAIT_OBJECT_0 || r == WAIT_ABANDONED) return g_setup_lock = m, 1; /* (abandoned: its holder ended, it is ours) */
|
||||
CloseHandle(m);
|
||||
return 0;
|
||||
}
|
||||
static void setup_lock_release(void) {
|
||||
if (g_setup_lock) ReleaseMutex(g_setup_lock), CloseHandle(g_setup_lock), g_setup_lock = NULL;
|
||||
}
|
||||
|
||||
/* The checksums of the parts this version holds in folders of their own (reuse_part reads them at the next update) */
|
||||
static void write_parts(const wchar_t *install, const Part *parts, int nparts) {
|
||||
wchar_t list[MAX_PATH], rel[MAX_PATH];
|
||||
@@ -536,6 +665,14 @@ static DWORD WINAPI work(LPVOID unused) {
|
||||
const wchar_t *tb = _wgetenv(L"HOOKSHOT_SETUP_BASE");
|
||||
if (tb && *tb) wcsncpy(base, tb, MAX_PATH - 1), base[MAX_PATH - 1] = 0;
|
||||
}
|
||||
/* ((Triforce) this folder for this setup alone until the builder starts, its sweep included (setup_lock_take): another one
|
||||
waits for it, a while) */
|
||||
if (!setup_lock_take(base, 0)) {
|
||||
status(L"Another Hookshot Builder setup is installing. Waiting for it to finish...");
|
||||
if (!setup_lock_take(base, SETUP_LOCK_WAIT_MS))
|
||||
fail(L"Another Hookshot Builder setup is installing into %ls right now. Wait for it to finish (or close it), then run the "
|
||||
L"setup again.", base);
|
||||
}
|
||||
swprintf(downloads, MAX_PATH, L"%ls\\downloads", base);
|
||||
mkdirs(downloads);
|
||||
|
||||
@@ -618,12 +755,21 @@ static DWORD WINAPI work(LPVOID unused) {
|
||||
swprintf(exe, MAX_PATH, L"%ls\\%ls", install, wrun);
|
||||
for (wchar_t *q = exe; *q; q++) if (*q == L'/') *q = L'\\';
|
||||
|
||||
{ /* ((Triforce) a setup cut off before, of this release or another: what it moved over from the installed version goes
|
||||
back first, before the installed version is looked at; a part folder whose folders could not all go back is kept, not
|
||||
deleted with them) */
|
||||
wchar_t left[MAX_PATH];
|
||||
if (!undo_every_reuse(versions, left, MAX_PATH))
|
||||
fail(L"The last setup could not be undone: %ls still holds folders of the installed version. Close every Hookshot window "
|
||||
L"and run the setup again.", left);
|
||||
}
|
||||
if (repair || GetFileAttributesW(marker) == INVALID_FILE_ATTRIBUTES || GetFileAttributesW(exe) == INVALID_FILE_ATTRIBUTES) {
|
||||
/* (unpacked into <tag>.part and renamed once complete: a setup stopped half way never leaves a version that looks installed) */
|
||||
swprintf(install, MAX_PATH, L"%ls\\%ls.part", versions, wtag);
|
||||
rmtree(install);
|
||||
rmtree(final_dir);
|
||||
mkdirs(install);
|
||||
wcscpy(g_versions_dir, versions), wcscpy(g_part_dir, install); /* (fail() undoes from here on) */
|
||||
for (int i = 0; i < nparts; i++) {
|
||||
Part *p = &parts[i];
|
||||
wchar_t zip[MAX_PATH], wsha[65];
|
||||
@@ -721,7 +867,13 @@ static DWORD WINAPI work(LPVOID unused) {
|
||||
if (GetFileAttributesW(staged_exe) == INVALID_FILE_ATTRIBUTES)
|
||||
fail(L"The builder is missing after unpacking (%ls). Your antivirus may have removed it: allow the folder %ls and run the setup again.", staged_exe, base);
|
||||
if (!MoveFileW(install, final_dir)) fail(L"Could not finish installing into %ls.", final_dir);
|
||||
g_part_dir[0] = 0; /* (installed: nothing to undo; the note goes) */
|
||||
wcscpy(install, final_dir);
|
||||
{
|
||||
wchar_t note[MAX_PATH];
|
||||
swprintf(note, MAX_PATH, L"%ls\\" REUSED_NOTE, final_dir);
|
||||
DeleteFileW(note);
|
||||
}
|
||||
write_small(marker, tag);
|
||||
write_parts(install, parts, nparts);
|
||||
}
|
||||
@@ -732,6 +884,7 @@ static DWORD WINAPI work(LPVOID unused) {
|
||||
}
|
||||
prune(versions, downloads, tag, parts, nparts);
|
||||
rmtree(downloads); /* (installed: its zips are not kept; an update moves unchanged folders over instead, reuse_part) */
|
||||
setup_lock_release(); /* ((Triforce) installed: the builder this starts may start a setup of its own) */
|
||||
|
||||
status(L"Starting the Hookshot Builder...");
|
||||
wchar_t cmd[4096], dir[MAX_PATH];
|
||||
|
||||
@@ -2,6 +2,6 @@
|
||||
<assembly xmlns="urn:schemas-microsoft-com:asm.v1" manifestVersion="1.0">
|
||||
<assemblyIdentity type="win32" name="Hookshot.BuilderSetup" version="1.0.0.0"/>
|
||||
<dependency><dependentAssembly><assemblyIdentity type="win32" name="Microsoft.Windows.Common-Controls" version="6.0.0.0" processorArchitecture="*" publicKeyToken="6595b64144ccf1df" language="*"/></dependentAssembly></dependency>
|
||||
<application xmlns="urn:schemas-microsoft-com:asm.v3"><windowsSettings><dpiAware xmlns="http://schemas.microsoft.com/SMI/2005/WindowsSettings">true</dpiAware></windowsSettings></application>
|
||||
<application xmlns="urn:schemas-microsoft-com:asm.v3"><windowsSettings><dpiAware xmlns="http://schemas.microsoft.com/SMI/2005/WindowsSettings">true</dpiAware><longPathAware xmlns="http://schemas.microsoft.com/SMI/2016/WindowsSettings">true</longPathAware></windowsSettings></application>
|
||||
<trustInfo xmlns="urn:schemas-microsoft-com:asm.v3"><security><requestedPrivileges><requestedExecutionLevel level="asInvoker" uiAccess="false"/></requestedPrivileges></security></trustInfo>
|
||||
</assembly>
|
||||
|
||||
@@ -108,7 +108,10 @@ second game.
|
||||
7. **The CPU saturates first.** Ryzen 5 7600X, 6 cores / 12 threads: the game's threads, the replay, captures and
|
||||
pipeline compiles share it, and extra helper threads have lowered fps before. Thread counts are measured, never
|
||||
raised by default (`HOS_VCORES`, `RX_VK_CAPTURE_THREADS`, `HOS_THREAD_TIMES=<seconds>`, the `HOS_SAMPLE` profiler
|
||||
with `scripts/prof.py`).
|
||||
with `scripts/prof.py`). The one exception, by the owner's choice (2026-11-15): the Replay helper threads row is
|
||||
Automatic unless set, none with four cores or fewer, and its threads are kept only where trials in play measure the
|
||||
frames shorter, or, without a trial, where the replay is plainly the frame's limit by a margin over the GPU's time
|
||||
(the rule in `gpu/vk/replay_threads.h`, fitted to the owner's logs; a trial that finds them slower holds it).
|
||||
|
||||
Also inherited, and game-specific in MKW: MKW's frame-pacer hooks at fixed guest addresses are gated on MKW's build
|
||||
id (`hos_game_is_mkw`); OoT gets its own once its pacer is found, never by address coincidence.
|
||||
@@ -260,6 +263,6 @@ dump, the runtime made game-neutral, the stage-1 tools checked on MKW's dump as
|
||||
repeat with a delay that keeps ~45 fps. The MKW session's mirror page skip fired 0 times here and was removed.
|
||||
Not an instrument's doing: with nothing attached, photographed from outside the window (37 s, ~2200 photographs): 16 with `RX_VK_FRAMEGEN=0`, none at the user's 4x. With
|
||||
`RX_VK_SUBMIT_DELAY_US`: 7 at 300, 0 at 1000; a repeat at 1000 with shots still had 7 events against 50-60. So it is a race with the game's threads: less overlap, fewer. The delay is a test, not a fix.
|
||||
- The builder's first full build: 43 minutes at below-normal priority, 0 failures; the folder starts from a double click and reaches the title at 60 fps.
|
||||
- The builder's first full build: 43 minutes at below-normal priority (the builders' start then; since the performance audit's OS-57 normal, below normal with `HOOKSHOT_LOW_PRIORITY=1`), 0 failures; the folder starts from a double click and reaches the title at 60 fps.
|
||||
- Render conditions here gate single lights and read reports earlier in the same stream, so they cannot be read at record time.
|
||||
|
||||
|
||||
@@ -1,7 +1,7 @@
|
||||
# Hookshot (oot-recomp)
|
||||
|
||||
**Hookshot** is the port's name (as Skybound is TOTKPort's and BlueShell is mkw-recomp's); the maintainer's repository
|
||||
folder stays `oot-recomp`. The game builds as `hookshot.exe` + `hookshot_vk.dll` (`hookshot` + `hookshot_vk.so` on Linux).
|
||||
folder stays `oot-recomp`. The game builds as `Triforce.exe` + `triforce_vk.dll` (`triforce` + `triforce_vk.so` on Linux).
|
||||
|
||||
A static recompilation of the Switch 2 *The Legend of Zelda: Ocarina of Time* remake to native x86-64 Windows and Linux:
|
||||
the game's code lifted to C (NotRexed's a64-kit, Rexcomp's AArch64 kit), the Horizon OS and NVN2 implemented
|
||||
@@ -76,17 +76,55 @@ screen. It costs some GPU time on cards that cull quickly, so it is off unless s
|
||||
| Switch | What it does |
|
||||
|---|---|
|
||||
| `HOS_HEADLESS` (set) | No window: the renderer draws offscreen, for screenshots and frame dumps |
|
||||
| `HOS_WINDOW_SIZE=<w>x<h>` | The window's size in pixels (1280x720 by default, wider for an ultrawide picture) |
|
||||
| `HOS_WINDOW_SIZE=<w>x<h>` | The window's size in pixels (1280x720 by default, wider for an ultrawide picture; 64 to 16384 a side) |
|
||||
| `HOS_WINDOW_NOACTIVATE` (set) | The window opens behind the others and does not take the focus (Windows) |
|
||||
| `HOS_FULLSCREEN=1` | Starts in fullscreen; the Insert menu's Fullscreen row is then left alone |
|
||||
| `HOS_ALT_MENU=1` | (Triforce, 11-22, after Casual Gamer's #dev-fixes post) A lone Alt is left to Windows, as before: it enters the menu loop on the window's thread, which during the start-up screen is the one waiting for the shaders, and the screen froze until a key was pressed. By default the window, which has no menu bar, answers that command itself (`host/display.c` `keymenu_dropped`, `tests/test_alt_menu.py`); Alt+Space's system menu, Alt+Enter, F10 and Alt+F4 are as they were |
|
||||
| `HOS_AUDIO=0` | No sound device is opened (the mixer still runs); `scripts/run-vk.ps1` sets it unless `-Sound` |
|
||||
| `HOS_AUDIO_EFFECTS=0` | The game's plain reverb, delay and biquad filter effects on its mixes and its splitter destinations' biquad filters are accepted and left out, as before they were emulated (the sound unfiltered and at full level: the "Volcano eruption audio bugged" ticket); the I3DL2 reverb still runs. By default they run on the mix buffers in the order the game added them (`host/nn_audio.c` "mix effects"; `tests/test_audio_mix_effects.py`) |
|
||||
| `HOS_LANGUAGE=<code>` | The game's language for this run, a code from `language.txt` |
|
||||
| `HOS_OUTPUT_RES=<0..3>`, `HOS_HANDHELD=1`, `HOS_ASPECT=<ratio>` | Over the options file: the output resolution (0 2048x1152, 1 2560x1440, 2 3840x2160, 3 1920x1080), the handheld profile, a picture wider than 16:9 (2.3889 for 3440x1440) |
|
||||
| `HOS_OUTPUT_RES=<0..3>`, `HOS_HANDHELD=1`, `HOS_ASPECT=<ratio>` | Over the options file: the output resolution (0 2048x1152, 1 2560x1440, 2 3840x2160, 3 1920x1080), the handheld profile, a picture wider than 16:9 (2.3889 for 3440x1440; the Ultrawide row's 21:9 is 2.3704, 64:27, and its 32:9 3.5556) |
|
||||
| `HOS_DYNRES=0`, `HOS_REALTIME_STEP=0`, `HOS_TEXT_SPEED=<0..3>` | Over the options file: Resolution 100% (non-zero: Auto), Game speed Per frame (1: Real time), Text speed |
|
||||
| `RX_VK_FRAMEGEN=0`, `RX_VK_FRAMEGEN_MULT=<2..8>`, `RX_VK_PACE=0`, `RX_VK_UPSCALE=<0..2>` | No frame generation whatever the menu says; over the options file: its multiplier, frame pacing, the upscaling filter |
|
||||
| `RX_VK_FIRST_START_AUTO=1`, `RX_VK_FIRST_START_AUTO=<MB>` | (Triforce, 11-22, the performance audit's PRESENT-46, after Source A's; off by default) A first start sized to the card: with no options file, `RX_VK_LOW_SPEC` not asked and `HOS_DYNRES` not set, a weak card starts with the Render resolution row at 50% in place of Auto, or at 33% for an integrated GPU beside 4 processor threads or fewer (`gpu/vk/first_start.h`, `gpu/vk/vk_menu.cpp` `menu_first_start`, called from `mkwvk_init` once the device is made). Weak is Source A's tier 1: an integrated GPU (a handheld's is one; the `SteamDeck` variable is not read, it took desktops for handhelds), a card without mesh shaders, or 4.5 GB of video memory or less. Only that row is set: the upscaler, the textures and the output stay, and nothing is saved, so each start chooses again until the player saves options. A number over 1 plays a card of that many MB (the lab: `=4096` on a large card). Not set or `=0`: as before. Off until the owner decides what a first start on a weak card gets; unmeasured (Source A states no gain and tried it on one RTX 4070 Ti). One `menu: first start:` line says the choice |
|
||||
| `HOS_STEP_STEADY=0` | (Triforce, on by default since the lab's A/B; `=0` the mean of the last five as before) The game's real-time step kept steady, after NX2navi's steady step (`host/step_steady.h`, `host/nn_vi.c` `realtime_step`): the mean of the last 12 present intervals without their longest and shortest in place of the mean of the last five (`HOS_REALTIME_WINDOW`), the paced period itself where that mean is within 3% of it, and what the display showed and the game was not given repaid at a 64th a frame, 10% of the step at most (a quarter second owed at most). A cap the game holds steps it by the period exactly, and a hitch is paid back over most of a second instead of in the next five frames (a 50 ms frame at 240: 200 frames of at most 10% more, where the mean of five gave five steps of 13 ms). Game time still follows the clock |
|
||||
| `RX_VK_FG_KEEP=1` | (Triforce, 11-21, the performance audit's PRESENT-18) Frame generation's images are kept once made, as before. The default lets them go after 120 presents in a row with frame generation off (about two seconds at 60, so a row stepped through keeps them): the built-in interpolation's frames and made-frame sets, which stayed in VRAM until a resize after it was turned off or left for NVIDIA's own, and AMD FSR 4 frame generation's four images of the picture without the HUD once that row is left. The display thread is drained and the GPU idle first; they are made again at the next frame that needs them. The log says how many MB went (`gpu/vk/framegen_idle.h`, `vk_framegen.cpp` `framegen_idle`) |
|
||||
| `RX_VK_CAPTURE_TABLES=0` | (Triforce, 11-21, the upscaler audit's UPS-02) The stored draws' two tables as `std::unordered_map`, emptied and filled every frame, as before. While NVIDIA DLSS, AMD FSR 4 or AMD FSR 3.1 upscales, every plain draw into the scene stores its positions and is drawn again into the motion vectors; the count of a model's draws and the frame before's draws by their key took a node from the heap for every stored draw every frame (about 1,400 of each in the owner's log, whose status line said 0.65 ms of CPU a frame for the stored draws in all). The default keeps their memory from frame to frame, so a frame the size of the one before takes nothing from the heap, and gives the next frame's list of draws its room at once. The status line's `ms CPU a frame` is the measure (`gpu/vk/capture_tables.h`, `vk_temporal.cpp` `capture_motion`). Not built or run in the game when it was written |
|
||||
| `RX_VK_REDRAW_STATE=0` | (Triforce, 11-21, the upscaler audit's UPS-03) Every stored draw drawn again into the motion vectors sets the primitive topology, the primitive restart and its index buffer before it draws, as before. The default sets each only where the draw's differs from what the pass has set so far (its first draw sets all three): about 900 draws a frame are drawn again in the owner's log, nearly all of them triangle lists without restart, so two commands a draw are left out, and the index buffer's where two draws in a row bind the same one at the same offset (`gpu/vk/redraw_state.h`, `vk_temporal.cpp` `capture_motion`). Not built or run in the game when it was written |
|
||||
| `RX_VK_DLSS_MODE_SLACK=0` | (Triforce, 11-22, the performance audit's PRESENT-38) DLSS's quality mode found strictly, as before: a render size is a mode's only inside the range NGX answers. Not set: a size outside a range by a rounding (a 64th of the range's largest size, 2 pixels at least: `gpu/vk/dlss_checks.h` `mode_for`) is that mode's, so the Render resolution row's 33% at 2560x1440 (854x480, NGX's is 853x480) draws with DLSS and not the game's own upscale; and the feature there is is let go only once a mode took the new size. The lab's `dlssstrict` phases switch it within a run |
|
||||
| `RX_VK_DLSS_MODEL_SETTLE=<frames>` | (Triforce, 11-21, the performance audit's PRESENT-37) The DLSS model row steps K, M, L, NVIDIA's default and applies live: going from K to L made a feature for M on the way, tens of ms each, and NGX keeps a released feature's video memory for some seconds. A new preset is taken once the row has held it for `<frames>` frames, 30 when not set; the feature already made draws on meanwhile, and with no feature the row's preset is taken at once. `0`: at once, as before (`gpu/vk/dlss_checks.h` `PresetSettle`, `vk_dlss.cpp` `prepare`) |
|
||||
| `RX_VK_DLSS_CNN_VRAM=1` | (Triforce, 11-22, the performance audit's PRESENT-36, after Source A's and Source B's; off by default) On a card with less than 7 GB of video memory the DLSS model row's K (the default) asks NGX for preset E, NVIDIA's last CNN model, and the row reads "CNN (E)" (`gpu/vk/upscaler_rows.h` `cnn_for_vram`, `gpu/vk/vk_dlss.cpp` `wanted_preset`, `dlss_cnn_vram`). M, L and NVIDIA's default stay as chosen, `RX_VK_DLSS_PRESET` goes over it, and cards of 7 GB and more are not touched. Source A measured it on a 6 GB laptop RTX 3050 (1080p, DLSS Performance, one run a preset): 15.75 to 14.20 ms of GPU a frame, NGX's own memory 85 to 58 MB. Off here, as in Source B: preset E ghosts more on grass and thin objects at the smallest render sizes and NVIDIA marks it deprecated, so the picture is the owner's to look at first. A DLSS runtime without E falls back to NVIDIA's default; `RX_VK_DLSS_LOG=1` shows the preset NGX took. One `vk: NVIDIA DLSS: ... MB of video memory: the K row draws with preset E` line says it is in use; `RX_VK_TEST_CARD_MB=6144` plays a 6 GB card |
|
||||
| `RX_VK_DLSS_FEATURE_MAX=1` | (Triforce, 11-21, the upscaler audit's UPS-01) NVIDIA DLSS's feature made for the largest render size its quality mode allows, whatever the Render resolution row says, as before. NGX answers that largest size as the output's own, so at a fixed share the feature was made for 2.25 times the pixels it was ever given at Quality (4 at Performance, 9 at Ultra Performance), and the chain's colour, depth and motion images with it; the owner's log has DLSS's own video memory at 1069 MB for such a feature with preset M, where NVIDIA's table has 456 MB at 4K, and NVIDIA's guide says the cost of presets L and M follows the size the feature is initialized for. The default makes a fixed share's feature for its own render size; the row's Auto (the game's dynamic resolution), and any run in which the render size is seen changing under one output, make it for the mode's whole range as before (one feature made again when that is first seen). DLAA's is the output's either way. The `feature N made` line says which, and the status line's `its own ... MB` and `its images ... MB` are the measure (`gpu/vk/dlss_feature.h`, `vk_dlss.cpp` `prepare`). Not built or run in the game when it was written |
|
||||
| `RX_VK_DLSS_PREWARM=1` | (Triforce, 11-22, the performance audit's PRESENT-32, after Source A's: its first step; off by default) NVIDIA DLSS at the output's size takes the scaled quality mode whose range holds that size, and only NVIDIA DLAA the one-size mode (`gpu/vk/dlss_checks.h` `dlaa_feature`, `gpu/vk/vk_dlss.cpp` `prepare`). With the Render resolution row at Auto the game starts at the output's size and lowers it some frames later: before, a one-size feature and then a second one made in play (Source A's save load-in: 119 ms, then 29.5 ms 17 frames on); with the switch one feature, for its mode's whole range. Off by default because DLSS at full size then draws with the scaled mode's network, not DLAA's: a picture to compare first. The rest of the entry is not built: making the feature, the chain's images and a first evaluation ahead in a light frame (a title or loading screen) needs the first gameplay upscale's sizes and colour kind before it has happened, which only a run with DLSS can settle. `tests/test_dlss_prewarm.py` runs the decisions over a load-in |
|
||||
| `RX_VK_FRAMEGEN_REFRESH=1` | (Triforce, 11-21, the upscaler audit's UPS-05; off unless set) The Frame generation row's 2x to 8x make no more frames for each of the game's than the display's refreshes that fall in one frame of the game: the game's recent frame time over the display's period, rounded up, a tenth of a refresh forgiven, never more than the row's number and never fewer than the game's own frame. Without it the row's number is made whatever the display shows: the owner's log has 5x on a 165 Hz display with the game near 66 frames a second, 331 frames a second presented, each present call 2.2 to 3.0 ms (0.03 ms without frame generation) and the display thread's line `0 frames shown on time, 1655 late`. Auto and AMD FSR's row are not touched. A `vk: frame interpolation: N frame(s) for each of the game's` line says each change (`gpu/vk/framegen_refresh.h`, `vk_framegen.cpp` `framegen_record`). Not built or run in the game when it was written |
|
||||
| `RX_VK_FRAMEGEN=0`, `RX_VK_FRAMEGEN_MULT=<0..9>`, `RX_VK_PACE=0`, `RX_VK_UPSCALE=<0..2>` | No frame generation whatever the menu says; over the options file: the Frame generation row's value (0 Off, 1 Auto, 2 to 8 the multiplier, 9 AMD FSR frame generation: a lab run's pin for each generator; `gpu/vk/framegen_rows.h` `env_row`), frame pacing, the upscaling filter |
|
||||
| `RX_VK_FG_OWN_MOTION=0` | (Triforce, FG-4) Frame generation's motion the camera's alone, as before, for comparing. Without it, while NVIDIA DLSS or AMD FSR 4 upscale, the depth and motion frame generation takes (AMD's, NVIDIA DLSS frame generation, FSR 3.1's interpolation) are made from the upscaler chain's frame, with what moves by itself drawn into the motion (`gpu/vk/vk_temporal.cpp` `fg_inputs_from_chain`, `gpu/vk/vk_fsr3_inputs.cpp` `chain_motion`): also where the gameplay network does not translate on the card and the upscaler draws it in its place (RDNA 3), where they were never made before and AMD's frame generation made no frame |
|
||||
| `RX_VK_DLSSG_OWN_MOTION=1` | (Triforce, 11-22) NVIDIA DLSS frame generation is given the upscaler chain's motion, as before (`RX_VK_FG_OWN_MOTION` above). By default it is given the camera's alone, from FSR's inputs: with the chain's the owner's flicker of the whole picture while the camera moved came back on 11-22, with frame generation at 2x and DLSS at 50%, and it stayed with the upscaler audit's commits set back; with `RX_VK_FG_OWN_MOTION=0` it was gone. DLSS frame generation finds what moves by itself with its own optical flow. AMD's frame generation and FSR 3.1's interpolation keep the chain's (`gpu/vk/fg_motion.h`, `vk_fsr3_inputs.cpp` `chain_motion`). Not built when it was written: the same as the run with `RX_VK_FG_OWN_MOTION=0`, for DLSS frame generation alone |
|
||||
| `RX_VK_DLSSG_CLIP=0` | (Triforce, 11-20) NVIDIA DLSS frame generation is given the game's own camera matrices, as before, for comparing. Without it they are in the clip space of the depth it is given (FSR's copy, 1 at the near plane): the game's clip z runs from -1 at the near plane to 1 at the far one, and with its matrices that depth read as a clip z put what is far next to the camera, which showed as a flicker of the whole picture while the camera moved (`gpu/vk/dlss_math.h` `depth_clip_projection`, `gpu/vk/vk_dlssg.cpp` `frame`) |
|
||||
| `RX_VK_FG_HUDLESS=0` | (Triforce, FG-2) No picture without the HUD is taken for AMD's frame generation, so its library makes no frame (the game's frame alone): a test. Without it the gameplay upscale's output is taken before the HUD is drawn into it, sRGB-encoded as the game's last pass encodes it, and given as `HUDLessColor` (`gpu/vk/vk_framegen.cpp` `framegen_hudless_note`, `hudless_for`); a present without it shows the game's frame alone; with HDR Native or RenoDX's RGBA16F scene AMD's library is not asked and the renderer's own makes the frame, as at 2x |
|
||||
| `RX_VK_FRAMEGEN_FINE=0` | Frame generation's own made frames (Auto, 2x to 8x) the old way: a vector per 16x16 pixels in steps of 4 pixels, a blend of four of them per pixel, a pixel alike in both frames taken as still. Without it, a vector per 8x8 pixels to a fraction of a pixel, settled among its neighbours, and one motion per pixel (`gpu/vk/vk_framegen.cpp` `fine_vectors`, `shaders/fg_fine_*.comp`), which keeps moving straight lines straight and thin lines drawn (`tests/test_framegen_lines.py` measures both) |
|
||||
| `RX_VK_PACE_SMOOTH=0` | (Triforce, on by default since the lab's A/B; `=0` off) The Frame pacing row on the plain present too, without frame generation: with the row on, a picture of the game's that comes sooner than seven eighths of the recent interval after the last one waits for that, its blit on the queue and the queue's lock let go; a late one goes at once, and none waits more than 50 ms (`gpu/vk/present_smooth.h`, after NX2navi's `pace_smooth`). Without it the row spaces only frame generation's frames, as before. With `RX_VK_PRESENT_THREAD=0` the present is on the game's presenting thread, so a wait also holds that thread's next frame: what that does to the rate under the cap is for a measurement before it is on. A `vk: frame pacing` line every 1200 pictures says how many waited and how long |
|
||||
| `RX_VK_PRESENT_THREAD=1` | (off by default since the lab showed no gain) (Triforce, on by default; `=0` as before) The game's presents go out from a present thread of their own (`gpu/vk/present_thread.h`, after NX2navi's present thread): the presenting thread records and submits the blit as before (the frame fence, the swapchain's image, the overlays, the bug capture's copy and the frame log's: the same commands in the same order on the queue), then hands the pacing's wait (`RX_VK_PACE_SMOOTH`) and the present call to that thread and goes on to its next frame; the next present waits first until the one handed over is out, so one is in hand at most. Not with frame generation, NVIDIA DLSS frame generation or AMD Anti-Lag 2 (presents and stages of their own); NVIDIA Reflex's present markers carry their frame's id. The `vk: a frame's present` line every 120 presents says how many went out from it, the presenting thread's time in the present and how much of it was the wait for the present before, and the pacing's wait and the present call (on whichever thread); the stall line's present is the presenting thread's, with that wait named |
|
||||
| `RX_VK_ACQUIRE_FENCE=0` | (Triforce, on by default; `=0` as before) The swapchain's image waited for on the CPU, not on the replay's queue (`gpu/vk/vk_replay.cpp` `present_image`, `image_free`): the acquire signals a fence, the present waits for it, at most 250 ms, before it records anything for the image, and the blit's submit waits for nothing. AMD's Windows driver returns from the acquire at once, the image not yet free: with the window covered (Alt+Tab, another window or OBS in front) the blit's wait for the acquire held the queue, every later submit queued behind it, and the game froze for as long as the cover. An image not free in time is held (`vk: swapchain image ... held`): no frame is shown, the game goes on, and the next presents look at it again without waiting until it is back. The wait counts as the pacing's in the present line. With DLSS frame generation the acquire's semaphore as before |
|
||||
| `RX_VK_SURFACE_RECOVER=0` | (Triforce, 11-22, the performance audit's PRESENT-45, after Source A's; on by default) A lost window surface is not handled, as before: after `VK_ERROR_SURFACE_LOST_KHR` (a display driver reset, a monitor change, a remote session) every acquire fails without a word and the picture stays frozen until a restart. Not set, an acquire, a present or a swapchain's creation that answers it lets go of what hangs on the surface, in order: the present thread's present out, the device idle, the overlays' views, the images' semaphores, the swapchain, then the surface (`gpu/vk/vk_main.cpp` `invalidate_lost_surface`); and the next present's `create_swapchain` first makes a surface for the window again with the startup's calls (`open_window_surface`), only while the window has a size (a minimized window waits), one try each 250 ms at most, with the presents' own queue chosen again for it. A loss seen on the present thread is only noted there and handled by the next present. Nothing changes while no surface is lost (every new statement is behind that result or a missing surface, and the startup's own code is as it was), nor while NVIDIA DLSS frame generation runs (the swapchain is Streamline's). `vk: the window's surface is lost` and `... is made again` lines say it. Never seen on a real driver here, and the window's path is not in the scripted (headless) runs: `tests/test_present_surface_lost.py` runs the code against a model of a driver that loses its surface at each call |
|
||||
| `RX_VK_REFLEX_LIMIT=1` | (Triforce, off by default) With the NVIDIA Reflex row on and no frame generation, Reflex's sleep holds the Game frame rate cap (the sleep mode's minimum interval) where each frame starts, in place of the game's own wait; and with DLSS frame generation, Streamline's Reflex holds it the same way (`gpu/vk/vk_dlssg.cpp` `pace`; options_version 7: the Reflex frame limit row is gone, and this is its only switch until that path has been played on an RTX 40/50 card; `vk: dlssg: frame limit:` lines say who holds it): one wait a frame, timed by the driver (`gpu/vk/vk_reflex.cpp` `reflex_pace`, the rule `gpu/vk/dlssg_pace.h`'s; after NX2navi's `low_latency`). The game's own wait holds a movie, a frame RivaTuner holds, a window that showed nothing for 200 ms, and the rest of the run once over half of 60 frames came sooner than three quarters of the period, or when the driver refuses the interval. `vk: NVIDIA Reflex: frame limit:` lines say who holds it; the Full overlay's Frame limit line says NVIDIA Reflex after the rate. |
|
||||
| `HOS_PACE_HANDOFF=0` | The game's own wait paces every frame, with DLSS frame generation's Reflex too and whether RivaTuner Statistics Server runs: the renderer is never asked whether another limiter holds a frame, and RivaTuner gets no limit from the game (`host/nn_vi.c` `pace_handed`, `gpu/vk/vk_pace.cpp`) |
|
||||
| `HOS_MEM_QUERY_USED=0` | (Triforce, 11-19) `nn::os::QueryMemoryInfo` answers the memory in use as the program and the system's part only, as before. By default it is the program, the heap and every page the game mapped beyond it, at most the total, as Horizon answers (`host/hos_mem.c` `query_used`): the game's resource cache sizes what it may still take from it, and without it kept every area it ever loaded (12.5 GB of its 9809 MB after four areas). Needs the census on (`HOS_MEM_CENSUS`). `HOS_MEM_QUERY_SLACK_MB=<MB>` (1024 by default): that much less of the game's pages answered as in use, room for its cache beyond a Switch's (our game needs ~3.2 GB of regions at a save load where a Switch leaves ~2.8; with 0 it evicted in every scene) |
|
||||
| `HOS_OBJ_CACHE=<n>` | (Triforce) The per-thread cache that finds a guest nn::os object (mutex, event, condition variable ...) by its address (`host/nn_os_sync.c` `obj_lookup_cached`, `tests/test_obj_cache_size.py`): by default 16,384 entries, four ways to a 64-byte set picked by murmur3's finalizer, the oldest way replaced; 256 KB a thread, allocated at its first lookup. The direct-mapped cache missed at nearly every step of the game's walk over some 4,000 mutexes 0x80 apart. n: 256 to 65,536 entries (down to a power of two); `0`: the direct-mapped 4,096 entries, as before. Every 10 s the log's `sync: object cache` line counts the lookups and misses |
|
||||
| `HOS_NO_OBJ_CACHE=1` | No per-thread object cache at all (either kind): every nn::os call looks its object up in the shared table |
|
||||
| `HOS_GUEST_MEMFD=0`, `=1` | (Linux) The guest window as a memfd's, its pages handed to the GPU as dma-bufs (udmabuf), where `/dev/udmabuf` can be opened: amdgpu checks every page of a host-pointer import again at each submission, 15-45 ms a submission in Kokiri. (Triforce, 11-21) Also where udmabuf's `size_limit_mb` cannot be read (Flatpak and other sandboxes hide `/sys/module`): its default, 64 MB, is assumed and the log says so; only a limit that reads 0 keeps the window anonymous. The log's `mem: the guest window is a memfd's` line. `=0`: anonymous memory, host-pointer imports; `=1`: the memfd all the same |
|
||||
| `HOS_MEM_TRACE=1` | (Triforce, lab) Logs every address region and page call of the game with its size and caller (`memtrace:` lines; a region's with its call stack) and every `QueryMemoryInfo` answer |
|
||||
| `RX_VK_WB_EXCL_CLIP=0` | (Triforce, 11-22; on by default with `RX_VK_CAP_ONLY_GPU=1`) A written binding is copied back to guest memory whole at the submission's end, as in 0.6.0. By default the parts of it in memory the capture leaves out (heap memory between the game's pools inside an imported chunk) are not: the collection does not stamp the CPU's writes there, so the mirror's bytes there are stale, and the write-back diff copied them over the game's newer ones. A pointer table's bare entry (its 64-byte head marked written, `gpu/vk/vk_draw.cpp`) did so over a heap block 0x18 past its base: the 0.6.0 crashes at area changes in guest 0x803b44e0 (a block header of 0x9ed00000001 or 0x1aa88000013fe), 0x80a30a00 and 0x80a31030 (`gpu/vk/wb_excl_clip.h`, `gpu/vk/vk_memory.cpp` `gpu_buffer_bind`; `tests/test_wb_excl_clip.py`). The first 20 clipped write-backs are logged (`vk: write-back of ... not copied back`) |
|
||||
| `HOS_FPS_STEADY=0`, `=1`, `=2` | (Triforce, off by default: the Frame pacing row's two Steady values, over them) 1 any rate, 2 steps of the display's refresh: the game's frames paced to a period this PC holds in the scene, never shorter than the Game frame rate cap's (`host/pace_steady.h`, `host/nn_vi.c` `steady_period`). The rate is chosen by probes (`pace_steady_probed`): 32 frames at the cap, unpaced, the last 24 of them measured, at the start, 2 s after a load, at least every 12 s, and when a quarter of the last 64 frames ran late (2 s after the last probe at the soonest, twice that each time a probe changes nothing, up to 12 s); the rate becomes the first the last three probes' 90th percentile fits in with 2% to spare where that is slower, or the first it fits in with 5% to spare where that is faster. A frame paced slower does more work (the lab's Market: 11.6 ms unpaced, 15.5 ms paced at 55 fps), so the frames' own measure as they run chased the rate down; frames late now only bring a probe on, and the rate is never under what a probe shows the PC holds. `HOS_FPS_STEADY_PROBE=0`: no probes, the rule below as before. Without probes: what a frame took is its present interval less the longest of the pacing's waits in it (the acquire's, on the presenting thread or another, and the present's hold), and what it needed the interval less all of them (OoT's main thread and Presentation Thread take turns; where the waits add up to the whole interval, the longest: `pace_steady_measure`); a quarter of the last 64 frames that took more than the period lowers it at once to the first rate their 90th percentile fits in with 2% to spare, and once it has held 3 s since a raise (6 s since a lowering, doubled for each lowering within 30 s of a raise, up to 120 s, and back to 6 s at a load) it rises to the fastest rate the 90th percentile of what the last 512 frames needed fits in with 5% to spare (one rate at a time where the game acquires on another thread). A frame that took over 2.5 times the median of the last 64 (hitches among them), a present interval over 250 ms, and the frames of the 2 s after such a load (the scene is not its measure yet) are left out; the first lowering after a load waits for 128 frames of the new scene and takes their 90th percentile. The acquire's wait starts the frames a period apart (the steady step then gives the game the period as its time step) and the present holds each picture of the game's to the period after the last, less a 64th of it (`gpu/vk/present_smooth.h` `present_smooth_paced_at`, with or without the Frame pacing row; not with frame generation). Another limiter holding the frames (RivaTuner, Reflex) puts the cap back while it does. `vi: steady frame rate:` lines at each change say what the frames needed, and every 5 s what the rule was given (what the frames took and needed at p50 and p90, the present interval's p50, the waits a frame by kind, the frames kept and left out by reason, how many of the last 64 kept were late, and the probes and what the last three took unpaced); the `vk: frame pacing` line counts the present's holds |
|
||||
| `HOS_FPS_STEADY_PROBE=0` | (Triforce, with `HOS_FPS_STEADY` or the Steady frame rate row) No unpaced probes: the rate by the frames' own measure as they run, as before test31 (`host/pace_steady.h` `pace_steady_probed`, `host/nn_vi.c` `steady_period`); the frames' work grows when they are paced slower, and that rule slides down |
|
||||
| `HOS_PACE_SPIN=0`, `HOS_PACE_SPIN=<microseconds>` | (Triforce) The game's own frame wait on the high-resolution timer for the whole wait, as before, which wakes up to about 0.5 ms after the frame's time. By default the timer takes the wait up to 0.6 ms before the frame's time and the rest is yielded away on the clock, so the frame starts on its time (`host/pace_wait.h`, `host/nn_vi.c` `pace_sleep`); a number is another margin. Every 5 s in which frames waited a `vi: frame waits` line says how long after their time the waits ended, on average and at most |
|
||||
| `HOS_RTSS=0` | RivaTuner Statistics Server is left out entirely: nothing of it is read or loaded and its profile is not touched; a line `triforce_rtss_profile.txt` holds from a run that did not close is only reported in the log (`gpu/vk/rtss_limit.cpp`, *The options file* below) |
|
||||
| `HOS_RTSS_DLSSG=1` | With DLSS frame generation RivaTuner is tried as the limiter too, at the cap times the multiplier in effect, with every safeguard (off: the game's own wait or Reflex, `gpu/vk/pace_rule.h` says why) |
|
||||
| `RX_VK_DEVICE=<n>` | The graphics card listed as `vk: device <n>` in the log, over the menu's GPU device |
|
||||
| `HOOKSHOT_NO_UPDATE_CHECK=1` | No look for a newer release at the start |
|
||||
| `HOS_SINGLE_INSTANCE=0` | (Triforce) No lock on the game folder: as before, a second copy of the game starts on a folder one already runs from. By default the first holds a lock named after its game folder (`host/hos_instance.h`), taken before the log is opened; a second copy waits up to 5 s for one that is ending (20 s after the menu's Restart), then says in a message box which process holds the folder and ends (a `--prepare-shaders` run or `HOS_HEADLESS`: an `@@error` line, and the prepare ends as one with no Vulkan). Other folders (the lab's copy, the owner's game) never meet. The log's `instance:` line says what this run did |
|
||||
| `HOS_TRACE=1` | Every Horizon call logged, as `--trace` does |
|
||||
| `HOS_SAVE_EXPORT_DIR=<folder>` | Where the System tab's Export save rows write their zip, in place of the Downloads folder (set in the environment; `tests/test_save_share.py` uses it) |
|
||||
| `HOS_SAVE_LOCK=0`, `TRIFORCE_SAVE_LOCK=0` | (Triforce) The save's lock between the game and Setup: the game holds `<user>\save.lock` open without sharing from its first save mount to its end, and Setup's Convert saves holds it across its copy (`host/save_tx.h` `save_tx_lock`, `setup/Guard.cs` `SaveLock`). A running game makes Setup refuse the copy; a game started during one waits for it (30 s at most) before it reads its save. `HOS_SAVE_LOCK=0`: the game takes none; `TRIFORCE_SAVE_LOCK=0`: Setup takes none and only looks at the running programs, as before |
|
||||
| `RX_BACKEND=<name>` | Unset or `vulkan`: `hookshot_vk.dll`; any other value but `eden`: the WebGPU renderer, when the exe was built with one |
|
||||
|
||||
### Input for scripted runs
|
||||
@@ -128,8 +166,17 @@ case-sensitive, and an unknown one is ignored.
|
||||
| `HOS_THREAD_TIMES=<seconds>` | Every interval, the threads that used over 5% of a core |
|
||||
| `HOS_FPS_LOG` (set), `HOS_PRESENT_TRACE=<csv>` | The game's frame limiter, VRR flags and nominal rate every 300 presents; a CSV line per game present |
|
||||
| `RX_VK_FRAME_LOG=<csv>` | A CSV line per present: frame, seconds, frame time in ms |
|
||||
| `RX_VK_FRAME_TRACE=1` | A line a frame in `triforce_frames.csv` beside the game (the lab's scenarios keep it as `frames.csv`): present to present, the waits for shaders and pipelines, the uploads, and the replay's parts (`memsync_ms` ... `tex_lookup_ms`: the draws' parts are sampled and summed over every thread that records). (Triforce, wp20, `gpu/vk/frame_trace.h`) Then each frame split as the stall line splits a long one: `replay_ms` (the submits and the present phase), `present_ms`, `dispatch_ms`, `qsubmit_ms`, `par_wait_ms` (the replay thread's waits for its helpers' segments), `outside_ms` (the frame less the replay and the present: the game's own update), `game_wait_ms` (the game's fence waits), `captures_ms` (the captures at its submits), `wb_wait_ms` and `wb_held_ms` (the waits for write-backs, and their part after the writer's recording was done: `RX_VK_WB_WAIT_REC`), `qs_bg_ms` (the queue submits beside a background import), `gpu_waits_ms` ((Triforce, 11-22, the performance audit's SYNC-27) every thread's time in calls that block on the GPU, `RX_VK_GPU_WAITS` below). `python -I scripts/frame_rank.py <run folder or csv>` ranks what made the frames over the 99th percentile and over 1.5x the median by these, per window (the Market, the tour's stays, the loads) |
|
||||
| `RX_VK_SETTLE_GATE=1`, `=2` | (Triforce, 11-22, off by default; the performance audit's SYNC-29, after Source A's; `gpu/vk/settle_gate.h`) A CPU read that lets the stopped queues whose waits are met go on (the readback's look, which `RX_VK_RB_SETTLE=1` turns on, and the two looks made for a range a texture may be in) takes the order's lock whenever a queue is stopped, and that lock is held through whole replays and presents. With the gate a site looks only while a queue is stopped **and** the CPU has signalled an event since the last finished look, or a stopped queue waits on an event whose value is read from guest memory; any other met wait was met by a signal in the stream, and the thread that replayed it looked before it let the lock go. `1`: the readback's look; `2`: the texture sites' too (new ground: those looks order the exposure meter's readback, so `2` needs the lab's flicker count). Unset or `0`: every look as before. The `RX_VK_LAB_AB` knobs `settlegate` and `texgate` are `1` and `2` in their phases (`rbsettle+settlegate` is the pair the audit asks for). The 3000-readback line and the texture catch-ups' 300-call line count the looks made and skipped |
|
||||
| `RX_VK_READBACK_LATE=<rule>[,<rule>]`, `RX_VK_LOG_READBACK_RANGES=1` | (Triforce, 11-22, the performance audit's SYNC-19, after Source A's; `gpu/vk/readback_late.h`) Ranges the CPU reads back **without** the wait for their write-back, from what guest memory holds (the last write-back made): a rule is a size in hex (`3980`: every range of 0x3980 bytes, the culling camera block Source A reads late by default) or `newest` (a range whose writer is the newest submission handed out: the game cannot have waited for it). **Unset: no rule**, every readback waits as before, until the lab has run it where the waits are long (a handheld, an AMD iGPU, a weak discrete card); `=0`: none either. The `RX_VK_LAB_AB` knob `rblate` is `3980` in its phases. A late read hands the game the frame before's values, and when the CPU resets a range read late the pending write-back lands over its reset (the exposure meter's flashes): `newest` can catch that meter, so it is never a default. `RX_VK_LOG_READBACK_RANGES=1`: with every 3000-readback line, the 12 ranges that waited longest (readbacks, waits, ms, how often their writer was the newest submission, how often read late), to find such ranges on other hardware; nothing is noted without it |
|
||||
| `RX_VK_WB_WAIT_REC=1`, `=2` | (Triforce, off by default; the handoff's §12.9, `gpu/vk/wb_rec_wait.h`) A wait for a write-back still on the GPU (a capture's late pages on the game's submitting thread, a read or binding of a page put off) ends once the recording its submission was finished in is done on the GPU, not when the submissions' timeline reaches it: that timeline is held below any submission stopped at an event wait, and the lab's Market had 20-46 ms waits of the game's thread with the GPU 4 submissions behind. Whatever the switch, the wait for the recording comes first and the rest is counted: the stall line's `N of them X ms after their writer's recording was done`, the `late:` line's `ms of the wait after the recording that finished it was done`, the frame trace's `wb_held_ms`. Off: the captures' late-page wait orders something not known yet (`RX_VK_LATE_NEW`), so it is tried with the lab's flicker count first. (Triforce, 11-22, the performance audit's SYNC-122, after Source B's) The reads' side: a readback whose write-back is not recorded yet waited on the submissions' timeline with no bound and no second look; it is counted now (the 3000-readback line's second line), and the stall line names the held part by caller (takes, reads, binds, readbacks). The variable is a mode: `1` as above, and such a readback waits in 1 ms slices, looks its writer's recording up between them and ends with it; `2` is the check, where no wait ends earlier and the readback's slices count the part after the recording was done (what `1` would not have waited); unset, `0` or any other text: off, one wait as before. The check of the pages a later write-back changes (the audit's SYNC-101) is not built. `RX_VK_TL_REC=1` and the lab knob `tlrec` ended a capture's wait at a submission's first part, before its burst's write-backs were recorded: they now end it at the recording the submission was finished in, as `=1` here does; `RX_VK_TL_REC=2` keeps the old lookup |
|
||||
| `RX_VK_LATE_INLINE=1` | (Triforce, 11-22, the performance audit's MEM-80, after Source B's; `gpu/vk/late_inline.h`; **off unless set**, not yet run in the game) A late page (one the CPU wrote while a submission still on the GPU writes it back) is waited for and taken again by default (`RX_VK_LATE_QUEUE`: at the queue's next submit). `RX_VK_LATE_NEW=1` (knob `latenew`) turns four no-wait rules on, each of which brought brightness steps back at the Castle Town save. This turns on the partial rule alone (knob `lateinline`): a page only part of which a submission on the GPU writes back has its other parts synced from the capture, no wait; pages of pools with no mirror, texture storage, pages written back whole and pages whose write-backs' ranges are not known are waited for as by default. The partial rule, here and under `RX_VK_LATE_NEW`, now asks which ranges are still on the GPU against the progress the capture read before its copies, not a second read at the take: a write-back done between the two counts as in flight, and its range is not synced from the older copy. `RX_VK_LOG_LATE` shows the pages waited for |
|
||||
| `RX_VK_PASS_TIMING=1` or `2` | GPU time per pass (2: per program too), logged every 600 presents |
|
||||
| `RX_VK_SLOW_LOG=<ms>` | The replay's times for each frame that took longer than this |
|
||||
| `RX_VK_STALL_LOG=<ms>` or `0` | On in every run (`gpu/vk/stall_watch.h`): every 600 of the game's frames a `vk: frame times` line (the longest frame and how many were over 33, 50, 100 and 250 ms, present to present and in the replay's own time), and for each frame of 50 ms or more a `vk: stall` line of where its time went: the replay's phases, memory sync's parts, the waits for write-backs still on the GPU (how far ahead, and whether behind a present's blit), the waits for shaders or pipelines, the uploads, the pools imported where they were bound, the GPU behind, the present (its image, fence, call and Reflex's sleep), the game's fence waits, the rest, and a pool's trial in guest memory if one is under way (`RX_VK_DEMOTE_QUICK`); and the new textures left for the next frame and whether the frame's picture was held (`RX_VK_HOLD_MS`), whose holds the 600-frame line counts too. At most 3 stall lines a second and 1500 a session. `<ms>`: the stall line's threshold; `0`: neither line |
|
||||
| `RX_VK_LOCK_CENSUS=1` | (Triforce, 11-22, the performance audit's THREAD-68, after Source B's; off unless set) The lock census: who holds the lock a thread waits for. A lock that is busy when a thread comes for it notes the waiter, the thread that held it and how long the wait was, and every 10 s a thread of the census's own logs the 12 pairs that waited longest, as `vk: lock census: <waiter> <- <holder> (<lock>) x ms/s (n/s, y us each)`. Counted: `replay_mu`, `sync_order_mu`, `held_mu`, and in `vk_memory.cpp` `pools_mu`, `sync_mu` and `cap_mu`. Threads are named by their descriptions (the game's threads and the renderer's own have them). Off, each lock is taken exactly as before; on, a free lock is one try and a store, and only a busy one reads the clock. The holder is read without the lock, so one that lets go at that moment can be misnamed. With `RX_VK_LAB_AB` naming a knob `lockcensus` it counts only in that knob's phases. Only the counting of Source B's is here, not its sampling of the holders' stacks (`gpu/vk/lock_census.h`) |
|
||||
| `HOS_STALL_HOST=0` | (Triforce, 11-22, the performance audit's OS-51, after Source B's; on whenever the stall line is) The stall line names a long frame's time on the replay's side and calls the rest outside the replay and the present. A present that comes `RX_VK_STALL_LOG`'s threshold (50 ms) or more after the last one also logs the host's part of that frame, a `vi: stall, the host's part` line: the game's file reads, writes, flushes and save commits (how many, MB, their time and the longest one), the discards, zero-fills and decommits of memory it let go and the give-back's passes, the waits for a virtual core summed over the game's threads, the process's page faults (soft and hard together) and its CPU time in the kernel and in its own code. Causes that did not happen are left out; `none of the host's causes` rules them out. At most 3 lines a second and 1500 in a session, as the stall line. It costs a clock read before and after each counted call and two system calls a present. `0`: none of it; so with `RX_VK_STALL_LOG=0` (`host/stall_host.h`, `host/mkw_glue.c` `hos_stall_host_presented`) |
|
||||
| `RX_VK_SLOW_TAKE_MS=<ms>` or `0` | (Triforce, 20 by default) A submit-time capture take that long or longer (`gpu/vk/vk_memory.cpp` `take_capture`): a `vk: slow capture take` line with where its time went (the marks and late pages, the pools lock, the retires, the stamps, the texture shadow, the buffer kept or given back), the first 200 of them. A Linux RTX 4090 near the Market stuttered with `memory sync [capture taken 30 to 110 ms]` while the 600-take average said 1.3 ms. `0`: none |
|
||||
| `HOS_PGO=<file>` | In a run of an `RX_PGO=gen` build: `<file>.reset` clears the counters, `<file>.dump` writes them to `<file>`. It is read before `hookshot_env.txt`, so set it in the environment |
|
||||
|
||||
### The GPU
|
||||
@@ -137,25 +184,253 @@ case-sensitive, and an unknown one is ignored.
|
||||
| Switch | What it does |
|
||||
|---|---|
|
||||
| `RX_VK_VALIDATION` (set) | The Khronos validation layer |
|
||||
| `RX_VK_CHECKPOINTS=0` | No NVIDIA checkpoints (on by default; after a lost device they name the draw it stopped at) |
|
||||
| `RX_VK_HANG_DUMP=<file>` | Every thread's stack to the file once no frame was shown for `RX_VK_HANG_SEC` seconds (20) (Windows) |
|
||||
| `RX_VK_CHECKPOINTS=0` | No NVIDIA checkpoints (on by default; after a lost device they name the draw it stopped at). (Triforce) The boundaries between a frame's work have markers too: the write-back diff pass, each recording's start, uploads and end, the present's blit and its menu (`gpu/vk/checkpoint_kinds.h`); the report reads the two stages' last markers together, says what followed the last one reached, and logs the switches, the queue's timeline and the mirrors' piece binds the submits wait for. (P4) Each marking thread numbers its own markers in a ring of its own (15 of 4096; threads after the 15th share a 16th), no atomic shared by the recording threads; the report names a marker's ring and reads the markers around it in that ring |
|
||||
| `RX_VK_HANG_DUMP=<file>` or `0` | The hang watch (Windows) is on in every run: once no frame was shown for `RX_VK_HANG_SEC` seconds (20) with the game's window in front, the top of every thread's stack goes into the log and the whole stacks into `triforce_hang.txt` in the game folder, once a stall and at most three times a run. `<file>`: that file instead, and every second counts whatever window is in front. `0`: no watch (Triforce) The same report can be asked for at any time: put a file named `triforce_hang.txt.now` in the game folder (the report's name with `.now`); within a second the watch takes it away and writes every thread's stack, for a game that goes on showing frames and gets nowhere (a loading screen that never ends). |
|
||||
| `RX_VK_NOOPT` (set) | Pipelines built without the driver's optimisations |
|
||||
| `RX_VK_MESH_OPT=1` or `=0` | (Triforce) Task and mesh pipelines on AMD. Unset: built without the driver's optimizer on AMD's own drivers (it hung the GPU minutes into Kokiri on an RX 6600) and with it on RADV (without it, the foliage culling's task programs timed out the GPU's ring at the save load on a Steam Deck); the driver's part is made once more on RADV with mesh shaders. `=1`: with the optimizer everywhere. `=0`: without it on every AMD driver, RADV too, as before |
|
||||
| `RX_VK_GPU_CLOCK=0`, `=peak`, `=standard` | (Triforce, Linux) The Steam Deck's GPU held at amdgpu's peak stable power state while the game runs (`gpu/vk/vk_gpu_clock.cpp`): the Deck's automatic clock dropped while the frame waited on the CPU or a readback, and the GPU work after it ran slow (Kokiri handheld on a Deck OLED: 11.1 to 16.2 fps at the same 15 W; up to 4 W more where the APU was under its cap, so shorter battery life). Unset: only on the Deck's APU, and only while `power_dpm_force_performance_level` is `auto` (a level set by hand is left alone). The render node VK_EXT_physical_device_drm names is opened and the state set on an amdgpu context of the renderer's own (no root); the kernel gives the level back at the process's end, a crash too. The log's `vk: GPU clock:` line says what was done (AMD only). `=0`: nothing opened; `=peak` (or `=1`), `=standard`: that state on any AMD GPU, whatever the level |
|
||||
| `HOS_HALF_RATE_SHADING=1` | (Triforce, Linux) Keeps `RADV_FORCE_VRS_CONFIG_FILE`, which Steam's Game Mode gives every game (its Half Rate Shading). By default it is dropped before Vulkan loads when it names a file that exists: while it does, RADV on RDNA 2 and newer puts a forced shading rate into every pipeline's key, whatever the file holds, so the pipelines prepared in Desktop Mode or by the builder all compiled again (over 25 minutes to the title on a Deck). The slider then does nothing for this game, little loss: RADV keeps fragment programs that read their position at full rate, and nearly all of this game's do. `RADV_FORCE_VRS` is left alone. The log's boot: line says what was done. Can be set in `triforce_env.txt` |
|
||||
| `RX_VK_COMPUTE_PIN=all`, `lanes` or `shared` | (Triforce, the performance audit's PIPE-52, after Source A's and Source B's) Which compute programs' pipelines ask for subgroups of 32 lanes (`gpu/vk/compute_pin.h`, `gpu/vk/vk_draw.cpp` `pin_compute`). `all`, the default on every driver but Mesa's Intel one: every compute program, as before. `lanes`, the default on Mesa's Intel driver: only the programs that can tell how many lanes run beside them (a vote, a lane mask, the lane's index, a shuffle, FSWZADD, a matrix multiply, a barrier of Subgroup scope in the module, or shared memory in a workgroup of 32 invocations or fewer); the rest get the driver's choice, which Mesa's Intel compiler builds in a fraction of the time and memory (Source A, an Iris Xe: 0.8-1 GB over 23-25 s pinned, at most 300 MB unpinned). `shared`: as `lanes`, and any shared memory pins, whatever the workgroup. The log has `vk: compute programs held to 32 lanes: ...` where it is not `all` by default, and the pinned and free counts after the precompile. An unpinned program that did need 32 lanes computes wrongly: `=all` is the old pipelines. Not run on an Intel GPU here |
|
||||
| `RX_VK_FALLBACK_PHI_LIMIT=<n>` | (Triforce, the performance audit's PIPE-56, after Source A's) On Mesa's RADV for GCN 1 to 4 GPUs (HD 7000 to RX 500: RADV without 16-bit storage input and output) and on Mesa's Intel driver only: a mesh-fallback program with a compute stage of more than `<n>` `OpPhi` instructions is not handed to the driver; it is left out as a program the driver refused is, with a log line (8 at most), and its draws are not drawn (`gpu/vk/fallback_size_policy.h`, `gpu/vk/vk_draw.cpp` `fallback_compute_pipelines`). Unset: 50,000 on GCN (Source A: RADV for GFX6-8 had not built a 79,947-phi stage after 45-90 s and was killed at 25 GB), 20,000 on Mesa's Intel driver. `0`: no limit. No other driver has a limit, whatever is set. Not run on either driver here |
|
||||
| `RX_VK_FALLBACK_CALL_LIMIT=<n>` | (Triforce, PIPE-56) The same for a stage's `OpFunctionCall` instructions. Unset: 200 on Mesa's Intel driver (Source A, an Iris Xe: a 251-call stage unfinished after 27 minutes at 14 GB), none on RADV for GCN. `0`: no limit |
|
||||
| `RX_VK_NO_SPILLING_RATE=1` | (Triforce, the performance audit's PIPE-55, after Source A's; Linux build only) The Linux build sets `shader_spilling_rate=2` in its environment before the Vulkan instance is made (`gpu/vk/mesa_spilling.h`, `gpu/vk/vk_main.cpp` `create_instance`), unless the player has set that variable. Only Mesa's Intel shader compiler reads it: a program whose registers do not fit is built in fewer, larger steps (Source A, an Iris Xe: one compute program not built after 20 minutes and 13 GB, 32 s and 1.7 GB at rate 2). `=1` sets nothing, so the driver keeps its own rate. On Mesa's Intel driver the log has `vk: Mesa's Intel compiler spills at rate ...`. The Windows build sets nothing: under Proton, add `shader_spilling_rate=2` to the launch options. Not run on an Intel GPU here |
|
||||
| `RX_VK_DRIVER_FAULTS=0` or `2` | (Triforce, 11-22, the performance audit's PIPE-57, after Source A's; `gpu/vk/driver_faults.h`) On AMD's own Windows driver only (`VK_DRIVER_ID_AMD_PROPRIETARY`), unless `0`: a program whose pipeline build faulted inside the driver is not asked of the driver again this run, and its key is appended to `triforce_driver_faults.txt` beside the executable (one line: vendor, driver version, program) by the fault's handler, before a fault during the start's shader work restarts the game, so the next start with that driver skips it instead of faulting again (`vk: driver faults: N programs ... kept out of it`; delete the file to try them again, another driver version ignores its lines). And one build of a program is inside the driver at a time: a second waits outside it, a probe of the caches is told a compile is required. `2`: on every driver (the lab's); `0`: nowhere. On other drivers nothing of it runs (`gpu/vk/vk_draw.cpp` `fault_quarantine`, `build_graphics`, `guarded_graphics`) |
|
||||
| `RX_VK_PASS_BARRIER=game` or `full` | The barrier a draw pass starts with (`gpu/vk/pass_barrier.h`). `game`: a pass right after another, with nothing recorded between (no barrier of the game's, no copy, clear, uniform update or dispatch), waits for the earlier work in its fragment stages only, so its vertex, task and mesh work may overlap the earlier pass's fragments; a pass the replay itself split off for a hazard (`a pass of its own` in the log) and `RX_VK_BARRIER_EVERY_DRAW`'s keep a full one. (Triforce, P4) "Nothing recorded between" is the command buffer's own count (`RX_VK_BARRIER_CONTEXT`), so beside the Replay helper threads too. `full`: a full barrier at every pass start, as before. Unset: `game` with NVIDIA's driver, `full` on every other card. The log's `vk: pass barriers:` line says which, and the frame statistics' `pass starts ordered in their fragment stages only` how many a frame |
|
||||
| `RX_VK_CHUNK=<n>` | (Triforce, 11-22, the performance audit's SYNC-66, after Source B's; `gpu/vk/barrier_rules.h` `part_draws`) How many draws and dispatches a part of a long submission has in a serial replay: the GPU starts on the first part while the rest is recorded, and each split costs a queue submit, two full barriers and the part's bookkeeping (about 27 a frame at 120). Unset: the driver's default, **120 on every driver** until the lab has measured one (Source B cuts at 128 where the GPU idles while the CPU records and at 4096 on desktop drivers); `0`: whole submissions, never split; `120`: today's size whatever a later default says. The size was read once into a static: now the `RX_VK_LAB_AB` knob `chunk:<n>` changes it in its phases while the variable is unset (`RX_VK_LAB_AB=chunk:480,chunk:240` with `RX_VK_PARALLEL=0`). The log's first `vk: a submission's parts:` line says the size in force and the driver's default |
|
||||
| `RX_VK_SPLIT_BARRIER=pass` or `full` | (Triforce) A long submission goes to the GPU in parts of 120 draws and dispatches (`RX_VK_CHUNK`); `gpu/vk/barrier_rules.h`. `pass`: a pass open at a part's split goes on in the next part, begun again with the same targets behind a barrier of its attachments' outputs alone and with what its draws sampled and wrote, so the GPU does not drain between two draws of one of the game's passes; where the part's end records anything (its uniform updates, its write-backs) or an event's write is held, the full barriers as before. A parallel replay's segments are not split into parts: each begins and ends with a full barrier, so nothing goes on across a segment's edge. `full`: a full barrier at each part's end and the next one's start, as before. Unset: `full` on every card ((Triforce, 11-22, SYNC-66) this row said `pass` with NVIDIA's driver; `pass` has been off unless set since 11-16, when the lab's Market showed no gain from it: `barrier_rules.h` `split_in_pass`). The log's `vk: a submission's parts:` lines say which and, every 3,600 splits, how many went on in their pass and how many of those parts had uploads (whose own full barriers order them after the part before all the same) |
|
||||
| `RX_VK_GPU_WAITS=0` | (Triforce, 11-22, the performance audit's SYNC-27; on by default; not yet run in the game) The host's calls that block on the GPU are made without being timed, as before. By default each is timed where it is wrapped (`gpu/vk/vk_work.h`, `gpu/vk/gpu_wait_kinds.h`) and summed by kind: `vkWaitForFences` and `vkWaitSemaphores` with a timeout above 0 (fences, timelines), `vkQueueWaitIdle` and `vkDeviceWaitIdle` (idle), `vkGetQueryPoolResults` with the wait bit (queries). Polls count nothing, a wait inside a wait counts once, and the low-latency sleep's wait is left out as pacing. After each 120-present line the log has `vk: host waits for the GPU: ... ms a frame (fences, timelines, idle, queries; all threads)`, which counts what that line's "GPU-bound wait" (the recording slots' waits alone) does not, and the frame trace has `gpu_waits_ms`. The cost is two clock reads a blocking wait. The `RX_VK_LAB_AB` knob `nogpuwaits` turns the count off in its phases |
|
||||
| `RX_VK_ORDER_WAKE=1` | (Triforce, 11-22, the performance audit's SYNC-59; off unless set; not yet run in the game) A queue stopped at a wait for an event goes on when a submit finds the wait met. A signal of the CPU's (`nvnEventSignal`) wakes no one: the stop waits for the order thread's next `Sleep(1)`, or a game thread's wait yields up to 3 ms and then forces it. With the switch (`gpu/vk/vk_replay.cpp` `sync_watch`, `event_signal_queued`) the order thread waits on an auto-reset event, 1 ms at most, which the signal sets while a queue is stopped: one `SetEvent`, no lock and no replay on the signalling thread. And a game thread's wait for stopped work that sees a signal pending lets the queues whose waits are met go on first, as `sync_settle` does; nothing is forced. The order line counts the wakes and those that let work on: under about one a second, leave it off. The `RX_VK_LAB_AB` knob `orderwake` turns it on in its phases |
|
||||
| `RX_VK_RELEASE_QUEUE=1` | (Triforce, 11-22, the performance audit's MEM-85, after Source B's idea; `gpu/vk/release_queue.h`; **off unless set**, not yet run in the game) Without the replay thread (`RX_VK_ASYNC` unset), a texture or program the game finalizes is let go by its own thread: held for the replaying thread only while a queue is stopped at a wait (PR 430), otherwise after `sync_drain` and the replay's lock, which a replay holds through whole submissions. With this on, every such release goes to a list the replaying thread applies at the start of the next submission's replay or present when no queue is stopped (after the stopped work otherwise, as PR 430's); a texture's release counts the game's writes from the write generation of its call, so writes after the call are still uploaded. The `RX_VK_ASYNC` job path is not changed. The `RX_VK_LAB_AB` knob `relq` turns it on in its phases. On in every run, a count: every 600 releases a `vk: releases` line when one of them waited 0.5 ms or more for the replay or its order (how many, the time in all, the longest) or went to the list. Not built: the entry's `rel_wait_ms` column in frames.csv |
|
||||
| `RX_VK_UB_BATCH=1` or `2` | (Triforce, 11-22, the performance audit's SYNC-21; off unless set; not yet run in the game) A recording's uniform updates are written into guest memory at its end between two full barriers, one `vkCmdUpdateBuffer` an update (about 80 a frame) and nothing between two updates of the same bytes. `=1` (`gpu/vk/ub_batch.h`, `gpu/vk/vk_replay.cpp` `flush_ub_batched`): each update's bytes go into the recording's stream (256 bytes at least) and each run of updates into one buffer is one `vkCmdCopyBuffer` of many regions; a run ends where an update would write bytes the run already wrote, or where the destination buffer changes, and the next starts behind a full barrier, so overlapping updates reach memory in the stream's order. A piece that is not 4-byte aligned becomes a region too (it was a CPU write at record time); with the stream full a piece is written as before. `=2`: written as before, the overlaps counted (by guest address). Both log a line every 600 flushes: updates, runs ended by an overlap and by another buffer, copies and regions. Look at the counts first: a barrier at every run end costs GPU time if they are common. The `RX_VK_LAB_AB` knob `ubbatch` is `=1` in its phases |
|
||||
| `RX_VK_SPLIT_BEFORE_DRAW=0` or `1` | (Triforce, the performance audit's REPLAY-14, after Source A's) A part of a long submission (and a helper thread's segment) ends only right before a plain draw that will draw (`NOP_DRAW`, `_IDX`, `_INDIRECT`; rendering on, no rasterizer discard), not after any draw or dispatch (`gpu/vk/barrier_rules.h` `split_before_draw`, `cut_ok`; `gpu/vk/vk_replay.cpp`). On by default on Intel's GPUs only: on an Arc B580 a command buffer that begins with mesh draws drew its later plain draws as exploded triangles (Source A: 40 shard frames of 1,121, then 0 of 1,101). `=0` off, `=1` on with any GPU. A run of mesh draws or dispatches is then not cut, so a part can be longer than `RX_VK_CHUNK`. Not run on an Intel GPU here |
|
||||
| `RX_VK_SPLIT_UBFLUSH=0` or `1` | (Triforce, REPLAY-14) Whether a part's end writes the submission's uniform updates so far into the game's memory (`split_ub_flush`). On by default, as before, except on Intel's GPUs without mirrors of the game's pools (integrated ones), where only the submission's last part writes them, as an unsplit submission does: written at each part's end they reach the memory later parts' other reads and the game's CPU see (Source A: white blocks in 1 run of 3 on a Meteor Lake iGPU, then 0 of 3). `=0` off, `=1` on with any GPU. The log's `vk: a submission's parts: ended only right before a plain draw ...` line says both |
|
||||
| `RX_VK_EVENT_BARRIER=late` or `full` | (Triforce) An event signal in the game's command stream (`gpu/vk/barrier_rules.h`). `late`: nothing is recorded at the signal and its pass goes on (the word's buffer is bound there, as before); the event's word is written right after the next full barrier on its command buffer (the game's next barrier, a dispatch's or pass start's), so after all the work before the signal, and a recording's end writes what is still held behind a full barrier before its write-backs copy the words to the game's memory; a wait the replay finds met for an event whose write is held gets a full barrier first. The game sees the word at that barrier rather than at the signal. `full`: the pass ended, a full barrier and the write at each signal, as before. Unset: with the pass barriers' game mode (`RX_VK_PASS_BARRIER`: NVIDIA's driver). Not with the parallel replay. The `vk: event signals:` line says which |
|
||||
| `RX_VK_TAIL_EARLY=1` or `=0` | (Triforce, on by default since options_version 6; over the Debug tab's Start next frame early row, `tail_early=0` to turn it off; read at the start; after NX2navi's `tail_early`) The early tail (`gpu/vk/tail_early.h`, `gpu/vk/vk_replay.cpp`): the replay cuts a frame's recording at the gameplay upscale's dispatch (the game's network, or DLSS, DLAA, FSR in its place) and records the rest, the upscaler and the interface, as one command buffer; its event signals are held. At the recording's end a tail that wrote no binding (its inline uniform updates aside), bound no guest memory itself, ran no program of the game's that reaches guest memory through the page table or a texel buffer, and wrote no texture back has its event words written in its upload command buffer, after the scene's write-backs and ahead of its own commands: OoT's present poll then sees its frame's event while the GPU still runs the tail. Any other tail has the words written at its end, as before. Also with the parallel replay (its segments end at that dispatch) and with the lazy write-back. Not with `RX_VK_BENCH` or a frame dump. A `vk: the early tail` line at the start says on, and every 1200 tails how many went early and why the others did not. Not yet run in the game |
|
||||
| `RX_VK_DISPATCH_BARRIER=game` | (Triforce) Off unless set, not yet run in the game: a dispatch right after another, nothing recorded between, records no full barrier of its own (not an indirect one, nor a program a stand-in may take over), as NX2navi's `barrier_game` does; the game's own barriers order its dispatches. Only the game shows whether that is always so (a one-frame band or flicker where it is not): a switch for the lab. A `vk: RX_VK_DISPATCH_BARRIER=game:` line every 20,000 dispatches it left without one. (wp14, P4) It no longer stands aside while the Replay helper threads record (most frames at the Market, so the lab's earlier run of it measured little): "nothing recorded between" is the command buffer's own count of its commands (`gpu/vk/cmd_work.h`, `RX_VK_BARRIER_CONTEXT`), a segment being a command buffer of its own |
|
||||
| `RX_VK_BARRIER_CONTEXT=0` | (Triforce, P4; after NX2navi's per-context barrier state) What the barrier rules ask, "was anything recorded into this command buffer since its last full barrier, its last pass's end, its last dispatch", is answered by each command buffer's own count of its commands (`gpu/vk/cmd_work.h`), where one count of every thread's commands served before (an atomic add on every command by every recording thread). So the rules hold while the Replay helper threads record: the left-out repeats of a full barrier, `RX_VK_PASS_BARRIER`'s pass starts and `RX_VK_DISPATCH_BARRIER`. A command buffer's begin starts its state afresh; each segment begins and ends with a full barrier, so no rule relaxes across a segment's edge. `RX_VK_SPLIT_BARRIER` asks also whether anything went to the queue between two parts (every submit is counted). The rules' defaults are as they were. `0`: the one count, and the rules standing aside while the helper threads record, as before. The `vk: barrier state:` line at the start says which; the frame statistics' `full barriers recorded, ... left out` and `pass starts ordered in their fragment stages only` show the difference |
|
||||
| `RX_VK_BATCH_IDLE=1` | (Triforce, 11-22, off by default; the performance audit's SYNC-09, after Source A's; `gpu/vk/barrier_rules.h` `batch_idle`) With held submits (`RX_VK_BATCH`, the row below: on by default with Mesa's RADV only), a submit that joins the held batch sends it at once when the GPU has finished every batch sent; while the GPU is busy, submits still group. A held batch otherwise waits for a wait, a present, the 2 ms poll or the cap of 32 with the GPU idle beside it. Source A has this on wherever batching is, and off where guest memory stayed host-pointer imports (each submit there pays amdgpu's page check); here it is **off unless set on every driver** until a Linux RADV machine has run it, and no guard for host-pointer imports is built. Nothing changes without held submits (NVIDIA, AMD's and Intel's Windows drivers). The 120-present `vk: batched submits:` line counts its flushes as `idle` |
|
||||
| `RX_VK_BATCH=parts`, `1` or `0` | Held submits (`gpu/vk/vk_replay.cpp` `batch_way`). `1` (the default with Mesa's RADV): every submit held until something needs the GPU's results or the queue's order. (Triforce) `parts`, off unless set: only a whole submission's recordings are held, and only while the game's submit is being replayed; a part goes at once with those held before it, as do 4 held (`RX_VK_BATCH_MAX`), and the held go when the replay of the game's submit returns. Neither changes the command buffers or their order on the queue. `0`: each submit at once (the default elsewhere). The `vk: batched submits` lines say which and how many command buffers went in each queue call |
|
||||
| `RX_VK_PEND_HELD=0` | (Triforce, on by default) A flush for a fence wait or a progress poll looks at a flag before the queue's lock (`dev.present_mu`) and returns when nothing is held (always, with batching off): it took the lock to find nothing, behind a present or a late page's fix waiting on the GPU. A held submit's recording is made known to the waiters only once it is held with the flag raised, so a waiter that finds it flushes it (before, it could flush nothing and wait for the held signal until the next poll, present or cap); a recordings' wait whose recording the recordings' timeline has passed returns without a flush. The batched submits' line counts the looks that found nothing held. `=0`: the lock every time and the registration first, as before; `RX_VK_LAB_AB=nopendheld` the same in its phases (`tests/test_flush_nothing_held.py`) |
|
||||
| `RX_VK_IMPORT_HIGH=<percent>` or `0`, `RX_VK_IMPORT_LOW=<percent>` | The idle imports' valve (`gpu/vk/vk_memory.cpp` `imports_give_back`): past `HIGH` percent (85) of the system heap's budget, the imported pools bound least lately, idle for 600 submissions and without a mirror, are given back down to `LOW` percent (75; 10 to 95). On by default where that budget is under 12 GB (on Windows only with the memory handling on); `RX_VK_IMPORT_HIGH=85` turns it on at any budget, `=0` off altogether. Setting either keeps the whole budget as the measure (no cap, below) |
|
||||
| `RX_VK_IMPORT_CAP_MB=<n>` or `0` | (Triforce, the performance audit's MEM-22, after Source A's) What the valve above measures the heap against, where that is less than the budget (`gpu/vk/import_budget.h`). Unset: 4,500 MB with Mesa's RADV at a budget of 12 GB or more, where the valve was off and live imports piled up for the session (Source A, an RX 9070 XT with 32 GB: 221 live imports and 7.4 GB after two hours against 50-64 and 2.3-2.7 GB capped); no cap on any other driver or at a smaller budget. `<n>`: that cap on any driver and budget, which turns the valve on. `0`: no cap, RADV's behaviour before. A pass against the cap gives back 1 GB at most. Not run on RADV here |
|
||||
| `RX_VK_FOLIAGE_SHADOW_CULL=<layer>[f]` | (Triforce, after NX2navi's 754ded9) On from layer 1 unless set (`off`: as the game draws them): the draws of mesh programs into the sun's shadow cascades (depth only, square, wider than 768) from that layer of the cascades on (0 to 2; NX2navi's is 1) that cull neither face are drawn with their back faces culled (the front ones with `f`, as `1f`): the foliage's double-sided cards draw half their fragments there (`gpu/vk/foliage_shadow.h`). Not quite the same picture: a lone card facing away from the sun casts no shadow in the middle and far cascades (where cards stand in pairs or clumps the others cover it; the near cascade is drawn as the game asks). The log says `vk: RX_VK_FOLIAGE_SHADOW_CULL` at the start and names the first draws it culled (`vk: foliage shadow cull:`) |
|
||||
| `RX_VK_FOLIAGE_SHADOW_VRS=<layer>` | (Triforce, after NX2navi's 754ded9) On from layer 2 unless set (`off`: the extension not asked for, no pipeline changed): the same draws of native mesh programs, from that layer of the cascades on, shade one fragment per 2 x 2 pixels (`VK_KHR_fragment_shading_rate`, where the device has `pipelineFragmentShadingRate` and native mesh shaders; elsewhere nothing changes): the leaves' alpha test runs a quarter as often, depth still per texel. The rate is dynamic state only in the pipelines of native mesh programs with no colour target and a depth one (and in their two shader libraries, keyed apart), which the driver builds again once: on such a device the driver's part of the preparation makes those libraries too (and the `RX_VK_LIB_CHECK` sample asks for them), once more at the first start after the update, before the game (the other libraries come back from the cache), and the light precompile builds each depth-only pipeline of a native mesh program once more; their keys, the program cache and every other pipeline stay as they were. The log says `vk: the far shadow cascades' foliage at 2 x 2:` at the start and names the first draws (`vk: foliage shadow 2 x 2:`) |
|
||||
| `RX_VK_FOG_CULL=1` | (Triforce, after NX2navi's 0379b00) Off unless set: the volumetric fog's first compute pass skips, as the game does, the froxels behind a coarse depth table of the frame before and writes them as empty, which the fog's history smears into dark trails beside near objects (worst at the screen edges, in foggy places such as the Lost Woods and Kokiri Forest). By default those froxels are computed too: the comparison is made never true in that one module, found by its shape in the SPIR-V (`gpu/vk/spv_fog.h`: a block of five image writes, four of (0,0,0,1) and last 1/255, and the one branch to it an ordered greater-than decides). It is applied where the module is made (`make_module`), outside the translation stamp: the program cache's records stay as they were translated, so nothing is translated again, and only that one compute pipeline is new to the driver's caches. With the precompile keeping none (`RX_VK_PRECOMPILE_KEEP`, less RAM), its record of the compute programs built has the patch and the matcher's version in its key (`gpu/vk/vk_draw.cpp` kFogPatch times `spv_fog::kMatcher`, raised whenever the matcher finds modules it did not: a record written by the matcher before 11-17's, which patched none of Triforce's fog modules, is built again), so the first start with the patch, or with a new matcher, builds the listed compute programs once more before the game (every one but the fog's a cache hit, the fog's compiled into the play cache, `mkw_vk_pipelines_play.bin`, or the main one where there is none, and saved before the record says built) rather than compiling the fog's at its first dispatch; with `RX_VK_FOG_CULL=1` that key is the one it was. A little more GPU time in the fog's pass. The log says `vk: fog: module <crc>, the volumetric fog's first pass:` and what was done (also for a near miss: the exit found but not its comparison) |
|
||||
| `RX_VK_FIFO` (set), `RX_VK_IMMEDIATE` (set) | The swapchain's present mode for one run, over the Insert menu's VSync row: FIFO, as with vsync forced by a driver; immediate, where the surface lists it. Unset, the row decides (`gpu/vk/present_mode.h`): On is mailbox where the driver has it (immediate on Windows where it has none, FIFO elsewhere), Off immediate wherever the surface lists it |
|
||||
| `RX_VK_SKIP_PROG=<prefix>,...` | Leaves out the draws of these programs (default `53509044401d78d2`; `-` leaves out none) |
|
||||
| `RX_VK_PIPE_SKIP=0`, `1` or `2` | Default/`1`: wait for shader translation; a pending pipeline can be skipped only for attachment-only draws whose targets hold an earlier frame. `0`: wait for every pipeline too. `2`: restore the old skip for untranslated draws with persistent targets; this can omit writes before the shader is known. Translated memory-writing draws still wait |
|
||||
| `RX_VK_TRANSLATE_WAIT_MS=<ms>` | (Triforce, 11-22, the performance audit's PIPE-30, after Source A's; `gpu/vk/translate_wait.h`) Off by default: a draw whose program is still translating waits for it, as before. Set (1 to 100, digits only; 8 to try): a draw into targets that hold an earlier frame, of a program whose key's translated variants write no memory (noted at each new pipeline and at the precompile's loads), waits at most that many ms a frame in all, the translation hurried but not run on the draw's thread, then is left out; a spent budget leaves the next one out at once. A key never translated, one with a variant that writes memory, and a draw into new targets wait as before; `RX_VK_PIPE_SKIP=0` and `=2` keep their meaning. The log says how many waited and were left out (`vk: programs still translating, so far: ...`). More draws left out, mostly in load frames: for a lab to weigh (`gpu/vk/vk_draw.cpp` `draw`, `translate_note`) |
|
||||
| `RX_VK_PIPE_DEPTH_BUDGET=1` | (Triforce, 11-22, the performance audit's PIPE-29, after Source B's; `gpu/vk/pipe_wait.h` `depth_budget`, `choose`) Off by default on every driver, Intel's too (the entry would have it on for Intel's Windows driver: that waits for the owner's word). Set: a depth-tested draw into a pass drawn once whose shaders write no memory, outside a samples-passed count and not into targets kept from an earlier frame (the item icon capture's), waits for its pipeline from the frame's wait pool (`RX_VK_PIPE_FRAME_WAIT_MS`, 50 ms) and is then left out; that pipeline is then left out at once until it is built (the list emptied every 600 presents). The first 16 left out are named, and a line counts them (`vk: depth-tested draws into passes drawn once, so far: ...`). A render into a new depth target left out can stay missing until it is drawn again: the trade Source B made for an Arc B580's 11-13 s freezes (`gpu/vk/vk_draw.cpp` `draw`, `pipe_depth_budget`) |
|
||||
| `RX_VK_PIPE_FRAME_WAIT_MS=<ms>`, `RX_VK_PIPE_ONESHOT_BUDGET=1`, `RX_VK_PIPE_SKIP_NEW=0` | A draw whose pipeline is not ready (`gpu/vk/pipe_wait.h`, NX2navi's rules). A full-screen draw (at most 6 vertices over half the pass or more: a lighting, fog or tone mapping pass) into targets that hold an earlier frame is left out only where the hold of the last picture takes its frame (`RX_VK_HOLD_PIPELINES`); where it cannot (at its bound, resting after one, or off) the draw waits for its pipeline up to this many ms a frame in all (50, at most 500), then 40 ms more, and is left out after both; its job goes to the front of the workers' queue meanwhile. `0`: left out, as before. `RX_VK_PIPE_ONESHOT_BUDGET=1` (off; NX2navi has it on): draws into targets drawn once and draws whose shaders write memory wait from the same budget and are left out once it is spent, instead of waiting until their pipeline is made. `RX_VK_PIPE_SKIP_NEW=0`: a pass whose colour target is drawn for the first time is kept by its depth target alone, as before (by default its draws wait, as for any target drawn once: the game's item icon capture). The log's `vk: pipelines not ready` line counts them |
|
||||
| `RX_VK_PIPE_FAST_WORKERS=0`, `RX_VK_PIPE_LINK_MS=<ms>` | A pipeline whose stage libraries were cold has those libraries built by a worker first and is linked from them, ready with that link, and its whole optimised build follows as a job of its own (`gpu/vk/pipe_wait.h` `libs_first`, NX2navi's `pipe_fast_workers`); a draw that waits for it runs the libraries and the link, not the whole build. `0`: the whole build first and the cold libraries after it, as before. `RX_VK_PIPE_LINK_MS=<ms>` (0 by default, at most 100; NX2navi's 6): the links made on the draws' threads in a frame take at most this long; a new pipeline after that is linked by a worker, its draws left out or waited for meanwhile (needs the fast pipeline link and the first switch on) |
|
||||
| `RX_VK_PRECOMPILE=0` or `1` | The shader preparation before the title off or on (on in builds made with `RX_RELEASE=1`, as the builders' are) |
|
||||
| `RX_VK_TASK_MESH_FALLBACK=0` or `1` | Task + mesh draws through the compute fallback (1) or the driver's mesh shaders (0). By default the fallback is on for AMD's Navi 21 and Navi 22 on Windows (RX 6800, 6800 XT, 6900 XT, 6950 XT, 6700 XT), where the native path timed out the driver (issue 996) |
|
||||
| `RX_VK_PREPARE_THREADS=<n>` | The threads the shaders are translated and their driver's part made on before the title, 1 to 64. Unset: every processor, a thread past half of them for each GB of RAM free at the start, and half in a process started above normal priority (the log's "the shaders' preparation runs on" line says how many) |
|
||||
| `RX_VK_WARM_COMPUTES=listed` or `all` | (Triforce, 11-22; `gpu/vk/warm_computes.h`, `tests/test_warm_computes.py`) The compute pipelines the shader preparation's translation pass has the driver compile. `all`: every compute program's texture variants (about 9,700), as before. `listed`: only the programs play dispatched (`triforce_computes.bin`, about a tenth of them; the precompile builds the same ones), the others translated as before and compiled at their first dispatch; with no list yet (nothing played) every one. Unset: `listed` on Intel's GPUs (an Arc A750 spent 5175 s of a 5858 s pass on them, an RTX 4070 SUPER 170 s on the whole pass), `all` on the others. No stamp holds it |
|
||||
| `RX_VK_FALLBACK_SHARED_RECORDS=0` | (Triforce, 11-21, the performance audit's MEM-154) With the compute fallback for task and mesh shaders, each recording thread keeps buffer pairs: 64 MB of records, the same size for every program, and a payload sized by the task program. A payload that fits no pair now makes a pair that takes the newest pair's records and only a new payload; before, 64 MB more of records each time, kept for the session. The log's `vk: mesh fallback: a buffer pair` lines say made or shared. `0`: its own records each time, as before (`gpu/vk/vk_draw.cpp` `fallback_buffers_of`) |
|
||||
| `RX_VK_TASK_MESH_FALLBACK=0` or `1` | Task + mesh draws through the compute fallback (1) or the driver's mesh shaders (0). By default the fallback is on for AMD's Windows driver for RDNA 2 (every RX 6000 card and that generation's integrated GPUs: the driver whose `vk: vendor fixes` line names the "AMD proprietary shader compiler"), where the native path timed out the driver on the RX 6700 XT and RX 6800 XT (issue 996) |
|
||||
| `RX_VK_FALLBACK_LOADED=0` | (Triforce, 11-22, the performance audit's PIPE-62, after Source A's) On the task and mesh fallback, the precompile is handed only the fallback programs whose records are still on the disk, as before. The default also hands it every fallback program already in memory that is ready and not failed (`gpu/vk/vk_shader.cpp` `program_cache_load_some`, in the part outside the translation stamp), so that their compute pipelines are built on the precompile's workers before the title: in a start that prepares, `state_list_prepare` has read the state lists' programs in before the precompile asks, and those got no pipelines until their first draw, which waited up to 50 ms on the recording thread and was left out until they were built. The log says how many: `vk: N mesh fallback programs read in earlier in this start are made ready with the rest`. Only GPUs on the fallback (no mesh shaders, RX 6000 and RDNA2 integrated GPUs on Windows, the menu's Compute row, `RX_VK_TASK_MESH_FALLBACK=1`); the preparing start's load grows by the compiles moved out of play |
|
||||
| `RX_VK_MESH_AHEAD=0` | (Triforce, on by default, after NX2navi's `mesh_ahead`; `=0` as before) An indirect mesh draw's records are copied into a ring as `{count, 1, 1}` before it, outside the pass, so each one ended its pass and the next draw began it again (OoT's foliage, grass and leaves: about 16 a frame at the Market). Now the first indirect mesh draw of a run copies the records of the indirect mesh draws after it as well, in the same three barriers, and those stay in the pass: a run is the indirect mesh draws after it with only state, draws, counter reports and render conditions between (`gpu/vk/vk_replay.cpp` `mesh_run_kind`); a dispatch, a barrier, a copy, a clear, a uniform update, new targets or an event ends it, so no work of the game's between can write the records after they were copied. A draw takes only the entry copied for it (its op, records, count and stride); at most 32 draws and 1 MB of the ring a run (`gpu/vk/mesh_ahead.h`). A `vk: indirect mesh draws` line every 20,000 conversions says how many took their records from an earlier draw |
|
||||
| `RX_VK_NO_PROGRAM_CACHE` (set) | Neither reads nor writes `hookshot_programs.bin` and the pipeline notes |
|
||||
| `RX_VK_PROGRAM_CARRY=0` | The program cache (`triforce_programs.bin`) is good for one build of the translator only, as before: any change to its files, a log line or a host option's number in `gpu/vk/vk_api.h` too, translates every program again, and the pipeline cache, the pipeline notes and the prepare's record go with it. By default the cache is carried over an update that changed nothing of what translates (`gpu/vk/program_carry.h` has the rule, `tests/test_program_carry.py` holds it): the translation stamp (`nx2/build.py` `translation_stamp`) is the translator's files without `vk_api.h`, without the parts of `vk_shader.cpp` and `vk_bank.cpp` marked as outside it, and without comments, blank lines and spaces at lines' ends. The build log's `build: translation stamp` line says whether it is the last release's (`nx2/released_translation_stamp.json`, which players' caches are for, and which `tests/test_program_carry.py` holds this source to) or which stamped files moved it against that release; a second line, in brackets, compares it with the previous build in the same folder (its record: `translation_stamp.json` in the build folder, written once the objects are compiled). Both are about the build's part of the stamp: the runtime adds its own settings to it (`disk_stamp`). The log's `vk: program cache` line at the start says whether the programs were kept, carried over an update of the translator (and from which build), or translated again (and how many the cache had); while they are translated again a line now and then counts them. Switching it either way translates every program once more |
|
||||
| `RX_VK_PREPARE_RECHECK=0` | (Triforce, 11-22, after Casual Gamer's #dev-fixes post) Both as before: a start whose record (`triforce_prepared.bin`) says the shaders are prepared takes its word, and the empty mark of a start that died in the driver's part without stage libraries (`triforce_prepared.bin.warming`) leaves that part out on any GPU and driver. By default (`gpu/vk/prepare_again.h`, `vk_draw.cpp` `prepare_run`): where the record says nothing to do, the program cache's head and size are read, and one made again since for another GPU or other settings (another stamp, or its head alone) holds none of the shaders: `preparing again` (logged), the stale record removed first; a cache carried over an update keeps the stamp and passes. And the mark holds what the part is made for (the GPU, the driver, the programs), as the libraries' journal does: another stamp's mark is another driver's, and the part goes on (`tests/test_prepare_again.py`) |
|
||||
| `RX_VK_CACHE_SET=0` | (Triforce) No earlier set of the shader caches is kept or put back (`gpu/vk/cache_set.h` has the rule, `gpu/vk/vk_cache_set.cpp` the files, `tests/test_cache_set.py`). By default, when the caches' stamp changes and the build of the translator did not (a setting of the translation: the menu's Robust buffers row, the mesh shaders row, the GPU, an `RX_VK_` switch of the translation), the set there (the program cache and its index, the pipeline cache and the play file with their `.for`, the prepare's record with its `.check`, `.warming` and `.warming.skip`, the pipeline notes and what is kept beside them by the stamp) is renamed to `<name>.<stamp>.<ext>` when the disk has twice the set and 10 GB free, and `triforce_cache_set.kept.txt` lists it. One set is kept. A start whose stamp is the kept set's puts every piece of it back (a set with a piece missing is deleted), and the set it found is kept in its place; a start of another build deletes it. `triforce_cache_set.txt` names what the set there was made for, and the log's `vk: shader caches:` lines say which part of the stamp changed. A set put back while the player's `triforce_pipeline_states.bin` is not the one it was kept with leaves `triforce_cache_set.states.txt`, and the prepare before the title then goes through the state lists, so that what was met under the other set is built before the game. Files of a kept set's names that no record names (a start that ended in a swap) are deleted at the next start. `RX_VK_CACHE_SET=0` also deletes a kept set; `RX_VK_PCACHE_KEEP_OLD` and `RX_VK_NO_PROGRAM_CACHE` leave it off too (a kept set of this build stays, another build's is deleted) |
|
||||
| `RX_VK_GPL=1` or `0`, `RX_VK_GPL_KEEP_LINKS` (set) | The fast pipeline link (`VK_EXT_graphics_pipeline_library`; `gpu/vk/gpl_rules.h`), on by default since options_version 5: a new pipeline is linked from its program's two shader libraries where the driver has them compiled, and the optimised build replaces it. The Debug tab's Fast pipeline link row is the same without an environment to set (`fast_pipeline_link=1`); the switch goes over the row. With it the preparation has its driver's part ("Compiling shaders"), which now goes on after a start that died in it, leaving out the programs that start had in the driver's hands (`triforce_prepared.bin.warming` and `.warming.skip`), up to three starts in a row. `RX_VK_GPL_KEEP_LINKS`: the linked pipelines the optimised ones replaced are kept for the run, as before, instead of destroyed 600 presents and 10 s later. `RX_VK_NO_WARM` (set): no driver's part; `RX_VK_GPL_SYNC` (set): a cold library is compiled in the draw; `RX_VK_GPL_NO_CACHE` (set): the libraries are not kept in the pipeline cache |
|
||||
| `RX_VK_LIB_CHECK=0` | (Triforce, after NX2navi's e10ef60) On unless `0`. At a start whose record says the driver's part is made for these shaders and this driver, a sample of 12 kept program variants has its two stage libraries made as that part makes them, asking the caches only where the device can (the answer a draw's link gets; without the flag, `RX_VK_GPL_SYNC`, their time against a second making, which comes from the cache), twice, before the game (`gpu/vk/lib_check.h`). When more than half are cold (`mkw_vk_pipelines.bin` lost or refused, or the libraries built otherwise now without `kWarmRecipe` counted up) the driver's part is made again at once, as after a driver update, with its journal and skip list; warm, nothing more happens (some 50 library lookups). `triforce_prepared.bin.check` records that it did: a start that finds them cold again right after leaves them to the draws, as before (each try holds a start for minutes; remove the file to try again). A making again whose cache could not be saved is recorded the same way. A cache file over its bound is not made again at all. The log's `vk: stage libraries:` line says what was found |
|
||||
| `RX_VK_BUILT_CHECK=0`, `=2` | (Triforce, 11-22, the performance audit's PIPE-21, after Source A's and Source B's; `gpu/vk/built_check.h`) On unless `0`. A light precompile (`RX_VK_PRECOMPILE_KEEP`: none kept) records what it built for the caches in `<notes>.built`, and later starts left all of it to its first draw or dispatch on the record's stamp alone. Now, before the record is trusted, up to 24 of the recorded notes are asked of Triforce's own pipeline caches as a draw asks them (without compiling, where the device has `pipelineCreationCacheControl`; without it each is built twice and timed: a first build over 4 times the second and 3 ms more was compiled), and with the record's compute key up to 6 of the listed compute programs as a dispatch asks them. Under 16 answers nothing is said; half of them or more not there: the caches lost them, and everything noted and listed is built again before the title and recorded anew (`vk: precompile's record: ...`). Lost again right after that (`<notes>.built.check`): left to the draws, as before, with a hint about the driver's shader cache size. Also counted: play builds of recorded pipelines that took 40 ms or more and recorded compute programs a dispatch did not find (`vk: precompile's record: in play so far, ...`, at most every 10 s while it grows). `2`: built again at every start (the lab's). `0`: neither the check nor the counter, as before (`gpu/vk/vk_draw.cpp` `built_check_cold`, `precompile_run`, `built_play_build`) |
|
||||
| `RX_VK_SPV_PACK=0` | Every shader module kept for the whole run, as before. By default the SPIR-V of graphics and mesh programs is kept packed in memory (`gpu/vk/spv_store.h`, to about 0.37 of its size) and their shader modules, which the driver otherwise holds for the whole run, are made only for the time a pipeline is created with them, on the thread that builds it. Compute programs, the programs the preparation lets go and the programs a draw reads from the cache itself keep their modules. The log's `vk: shader code packed` lines say what it comes to and what it costs |
|
||||
| `RX_VK_PCACHE_RAM=1` | Off by default (in its one session on, v0.4.0-test9, it gave back nothing the working set showed). With it, once the PC has had under 1.5 GB of RAM available at two looks a second apart, the pipeline cache (when it is 256 MB or more) is destroyed once no pipeline creation is using it (`gpu/vk/pcache_gate.h`): the pipelines built after that go without it and it is not saved again that session, so `mkw_vk_pipelines.bin` keeps its last save. The log's `vk: RAM short ...: the pipeline cache ... is let go` line says when and what the working set did. Whatever this switch says, the cache's save in play waits while a copy of it would not fit in the RAM available with 2 GB to spare (the save at the game's close does not), and the builder of the predicted pipelines behind the game rests while under 3 GB (or a tenth of a PC of more than 30 GB) is available (`gpu/vk/mem_parts.h`) |
|
||||
| `RX_VK_WRITERS_WAIT_MS=<ms>` | (Triforce) How long the game's close waits for the files the renderer is writing on threads of their own (`gpu/vk/file_writers.h`; 2000 by default, `0`: no wait and nothing refused, as before). The close ends the process through ExitProcess, which stops every other thread where it is: the pipeline cache's save in play, the order file, a bug capture's or a debug picture, the hang report, the predicted pipelines' list, the frame log. At the close (`HOS_OPT_EXITING`) no such write begins any more and the ones under way are waited for; one still going at the bound is cut, and is one of the files written whole beside the old (`.part` / `.tmp`, then put in place: the prepared record `triforce_prepared.bin` and the `.for` files are now too), or a list whose reader takes whole records only. The log's `vk: exit: the files being written ... were waited for` line says how long, and what was cut |
|
||||
| `RX_VK_OPTIONS_MERGE=0` | (Triforce) The menu writes its own values over `triforce_options.ini` in place at every change, as before. By default (`gpu/vk/ini_merge.h`, `vk_menu.cpp` `save`) each line it writes is merged by its key with the file as it is then: a line changed outside the game while it runs (a text editor, a tool) and not changed in the game since the game read or wrote the file is written as the file has it, so a change made outside and one made in the menu within the same second both stay (a row changed on both sides is the game's); the game takes such a line at its next start. The file is written whole beside the old one and put in its place (in place only where another program holds it open so that it cannot be replaced). The log's `menu: options: ... changed in triforce_options.ini outside the game` line names such lines |
|
||||
| `RX_VK_MEM_PARTS=0` | No `vk: memory, the parts` line. By default every 600 presents it gives the working set, the game's own memory committed and in RAM (a sample of its pages), the rest of the working set in RAM by what it is (a sample too: the program's code, the driver's code, other code, mapped files and views, write-combined or uncached memory, the heaps and other private memory), the shader modules alive and the packed stages, the pipeline cache, the pipelines and stage libraries kept (how many a draw has had, how many were given back) and the allocator's blocks in system memory (`gpu/vk/mem_parts.h`), and what its walks and samples took. It is written on a thread of its own that nothing waits for; the guest window's walk is made again every eighth line |
|
||||
| `HOS_MEM_ZERO_MARK=1` | (Triforce, 11-22, the performance audit's MEM-13; off unless set; not yet run in the game) Freed guest pages that a Vulkan import holds are zeroed in place, not discarded (`host/hos_mem.c` `discard_pages` and `FreeMemoryPages`, `host/hos_svc.c` UnmapPhysicalMemory). The renderer does not see a bare memset, so its mirror, its captures and the write-back's shadow keep the freed bytes, and a later write-back could put them back over the zeros. With the switch the fill (`host/hos.h` `hos_guest_fill`) marks the pages written, after the memset. Off by default until it is measured: the marked zeros (up to 5.2 GB over a tour) are pages the captures and the mirror syncs then copy, which could hitch an arrival. To measure: two runs a side, the log's "capture ... copied (MB)" and "mirror syncs ... dirty pages", and the frames over 50 ms in each arrival's first 5 seconds |
|
||||
| `RX_VK_SWAP_RESERVE_MB=<MB>`, `RX_VK_TEST_SWAP_MB=<n>` | (Linux; Triforce, 11-22, the performance audit's MEM-141, after Source A's; off unless set, not yet run on Linux) A memory pressure hold that counts the swap. The RAM rules look at `MemAvailable` alone, which a PC with swap keeps up by pushing the game's pages out (a 16 GB Steam Deck's swap filled while it said little). With the reserve set (Source A's is 2560), every 30 presents the PC is short when `MemAvailable` is under 1536 MB or `MemAvailable` and `SwapFree` together are under the reserve, and until both are 256 MB clear of their lines a hold is on: a capture buffer is kept as a spare only up to 32 MB of spares, the staging arenas the GPU is done with are freed as it begins and then after 5 s unused, `malloc_trim` is asked every 600 presents from 64 MB free, idle imports are given back (in all of one hold at most the bytes the PC was short at its worst, still only after 3,600 idle submissions), and an integrated GPU's texture budget is less by what the PC is short, never under 512 MB, coming back 256 MB a look. A PC without swap, or whose swap cannot be read, is never short by the sum. The log's `vk: swap hold: on` and `off` lines say when and what each gave. `0` or not set: no hold, as before; never on Windows. `RX_VK_TEST_SWAP_MB=<n>` plays that much swap free, as `RX_VK_TEST_AVAIL_MB` plays `MemAvailable` (`gpu/vk/linux_pressure.h`, `gpu/vk/vk_memory.cpp` `swap_hold_step`) |
|
||||
| `HOS_MEM_CENSUS=<seconds>` or `0` | (Triforce) The guest memory census (`host/guest_census.h`), on by default every 5 seconds, on a thread of its own: a `mem: census` line with what the runtime handed out of the guest window by kind (the program and the system region, the heap, the address regions and the game's own pages in them, what lies outside them: the physical maps and first touches), the game's own count against the 9809 MB it is told it has, the pages the game let go that stay committed (zeroed under an import, or discarded with a freed region) and what the give-back gave back, the process's commit and working set, and the PC's commit and RAM; and a `mem: census, the renderer` line with the Vulkan back end's big holders (its imports, mirrors, VRAM in use, the allocator's blocks in system memory, the pipeline cache, the pipelines and stage libraries kept, the shader modules). A count of bits a 4 KB page, kept as each place commits, maps or lets go of guest pages. `0`: no line (and, with the give-back off, nothing kept) |
|
||||
| `HOS_MEM_GIVE_BACK=1` or `2` | (Triforce) Off by default, not yet run in the game: the pages the game let go that stay committed because an import held them (`FreeMemoryPages`, `UnmapPhysicalMemory`) or because they were discarded with a freed address region are decommitted once nothing of the renderer imports them, from 4 GB up (below it pages are handed out without a commit), up to 256 MB a census pass, in runs of 16 MB at most; the renderer answers and runs the decommit with its import locks held (`gpu/vk/vk_memory.cpp` `guest_release_unimported`), so no import can lock them between. `1`: every pass; `2`: only while the PC is short (its commit past 85% of its limit, or under 1.5 GB of RAM available). The census line's `given back` says how much |
|
||||
| `HOS_MEM_MAP_RESULT=0` | (Triforce) `MapPhysicalMemory` answers success even where Windows refused the commit, as before. By default it answers Horizon's LimitReached (`0x10801`), which the SDK turns into its out-of-memory result, and the log says `mem: MapPhysicalMemory ...: the commit failed` |
|
||||
| `RX_VK_PRECOMPILE_KEEP=1` or `0` | The precompile keeps every pipeline it builds for the session (1), or keeps none of them: they are built for the caches and let go, and a draw that first needs one builds it from there (0). Unset or `auto`: kept where the RAM available at the precompile and what the game holds already (less what reading the pipeline cache added, except on AMD's driver: `own_for_keep`) come to 22 GB or more, on a PC of 20 GB or more (`gpu/vk/mem_parts.h` keep_all_pipelines); the log's `vk: precompile:` line says which and why. (Triforce, 11-22, after Casual Gamer's #dev-fixes post) Unset (not `auto`): on a discrete GPU in a PC of under 20 GB every pipeline is kept while the game is told its mapped pages are used (`HOS_MEM_QUERY_USED`, `keep_all_small_pc`): a 16 GB laptop, ~0.6 GB more private memory, frames of 1 s or more 15 and 11 in two tours -> 3 and 6. A card the driver allows under 6.5 GB still leaves them to their first draw first; integrated GPUs keep the RAM rule; `auto`: the RAM rule, as before |
|
||||
| `RX_VK_TEX_PIPES_AHEAD=0` | (Triforce, on by default; `=0` as before) The texture converters' compute pipelines (the GPU ASTC decoder, the BC encoder, the unswizzle and the swizzle back: `gpu/vk/vk_texture.cpp` `texture_compute_pipeline`, `texture_pipelines_warm`) made through the pipeline cache kept across runs, and all four before the game: in the precompile, right after the cache is read, or at the start when the preparation was skipped at the build. Each was made without a cache inside the first upload that needed it, which with an empty driver cache (a first start, a new driver) waited for the compile, a `vk: slow upload` of 100-320 ms. The ASTC decoder and the BC encoder only where the GPU does not sample ASTC. The `vk: texture converters ready in` line says what it took |
|
||||
| `RX_VK_PIPE_RELEASE=0`, `RX_VK_PIPE_RELEASE_EVERY=<presents>` | Every kept pipeline stays for the session, as before. By default, while the PC's RAM is short (under 1.5 GB available at two looks in a row), every 300 presents up to 512 of the kept pipelines no draw has had yet are given back to the driver (`gpu/vk/pipe_release.h`); a draw that needs one later has it built again by a worker. The log's `vk: RAM short: ... kept pipelines no draw has had given back` lines say how many and what the working set did |
|
||||
| `RX_VK_PCACHE_CARRY=0` | The pipeline cache's file carried over a stamp change (`gpu/vk/pcache_carry.h`): at an update's preparation (Setup's "Preparing shaders", `HOOKSHOT_PREPARE_ONLY`) a `mkw_vk_pipelines.bin` made for other shaders is read all the same when it is this device's and driver's, within its bound, and at most 1.25 times its base (the size of the last set built fresh, kept beside it in `mkw_vk_pipelines.bin.base` with the carries since): the pipelines of shaders the update did not change are hits instead of hours of compiling. The cap is over every carry since the base was built, not a quarter more each time, and a carried cache may never pass the file's bound either. The preparation looks at a carried cache's size every 10 s while it builds and at each save: past the cap (or the bound), nothing more is built into it, it is not written, `mkw_vk_pipelines.bin.nocarry` names the new stamp (a preparation begun again after a Cancel starts fresh), and the preparation builds the driver's part again into a fresh cache: the compute pipelines the carried run built, from the program cache, then the stage libraries, nothing translated again (from the driver's own cache for what it had just compiled), which gets a new base. A carried preparation that is stopped (a Cancel, a closed window) saves its cache under the new stamp with its history marked open in `.base`: the preparation begun again reads it as carried still, capped the same (or, when `.nocarry` names the stamp, not at all), until a carried preparation goes through to its end. The log says `pipeline cache carried over a stamp change (N MB, base M MB)`. `0`: a file for other shaders is never read, as before (the rule of 11-14). The play file keeps its own rule |
|
||||
| `HOS_DRIVER_CACHE_ROOM=0` | (Linux; Triforce, 11-22, the performance audit's PIPE-97, after Source A's and Source B's; not yet run on Linux) The game builds the pipelines its notes predict (about 41,400) into the driver's own shader cache on disk, and NVIDIA's and Mesa's caches hold 1 GiB unless told otherwise and trim themselves to it: part of a full preparation is dropped and compiled again in play. By default the game sets `__GL_SHADER_DISK_CACHE_SIZE` (NVIDIA) and `MESA_SHADER_CACHE_MAX_SIZE` (Mesa: RADV, the Steam Deck) to 10 GiB before Vulkan loads, for its own run and the builder's `--prepare-shaders` run; a size already in the environment stays, and the cache only grows by what is built. The log's `boot: room in the drivers' shader caches:` line gives both values. `HOS_DRIVER_CACHE_ROOM=0`, in the environment or `triforce_env.txt`: neither is set, as before. Nothing is done on Windows, where NVIDIA's driver takes its size from its control panel (`host/hos_main.c` `driver_cache_room`) |
|
||||
| `HOS_DRIVER_CACHE_FOLDER=0` | (Linux; Triforce, 11-22, the performance audit's PIPE-98, after Source A's; not yet run on Linux) NVIDIA's Linux driver keeps no shader cache at all when `__GL_SHADER_DISK_CACHE_PATH` names a folder that is not there: it neither makes it nor falls back to its own folder, and every prepared pipeline is compiled again at each start. By default the game makes the folder the variable names, with its parents, for the user alone, before Vulkan loads (the builder's `--prepare-shaders` run too); the value is never changed, nothing is made when the variable is not set, and a `boot: __GL_SHADER_DISK_CACHE_PATH=<path> is not a folder and cannot be made one` line says when it could not. `HOS_DRIVER_CACHE_FOLDER=0`, in the environment or `triforce_env.txt`: nothing is made, as before. The driver files its cache by the program's file name: the Linux binary stays `triforce` (`tests/test_driver_cache_folder.py` pins `nx2/build.py`'s name), and `__GL_SHADER_DISK_CACHE_APP_NAME` is never set (`host/hos_main.c` `driver_cache_folder`) |
|
||||
| `RX_VK_PCACHE_CARRY_PROJECT=0` | The early drop of a carried pipeline cache (`gpu/vk/pcache_carry.h`). When the stage libraries are built into a cache carried over a stamp change, every 32nd program of the warm's order is built first, the others waiting for it, and the cache's growth over that sample, per program, times all the programs, is where it ends at this rate. With at least 1500 libraries in the sample, a projected growth of at least twice the room it had when they began (its cap, or the bound, less its size then) drops the carry at once, exactly as past its limit (not written, `.nocarry`, the driver's part built again into a fresh cache), a few minutes in rather than when it gets there: an update whose file was built by an older translator misses nearly every library and filled the cache for 34 minutes before it was let go. The sample is spread over the whole order, not its head: the notes' programs go first and the rest in the program cache's file order, and hits and misses come in stretches of it, so a head that misses while the rest hits would have been dropped for nothing. The notes' programs wait for the sample (a few minutes cold). The log says `the pipeline cache carried over a stamp change would grow to about N MB at this rate`. The compute pipelines' pass and the precompile keep the look past the limit alone. `0`: dropped only once past its limit, and the warm in its own order, as before |
|
||||
| `RX_VK_PCACHE_MAX_MB=<MB>` | The bound on the pipeline cache's file, `mkw_vk_pipelines.bin` (`gpu/vk/pcache_bound.h`): a file over it is not read (the cache starts empty and the next save writes a new one), and data over it is not written (the file keeps its last save). Unset: an eighth of the PC's RAM, 2048 MB at least and 8192 MB at most. `0`: no bound, as before |
|
||||
| `RX_VK_PCACHE_LIBS_ONLY=1` or `2` | (Triforce, 11-22, the performance audit's PIPE-78, after Source A's and Source B's; `gpu/vk/pcache_libs.h`) Off unless set. `1`: when a fresh pipeline cache file begins (none read for these shaders), on NVIDIA's own driver with 20 GB of RAM or less, 512 MB or more in the driver's own cache folder (GLCache) and the stage libraries kept in our cache, the file keeps the stage libraries and the compute pipelines alone: whole graphics pipelines are left to the driver's cache and there is no play file, so a file of ~650 MB instead of 1.1-1.6 GB is read and held. Recorded in `<file>.lean` for that file's life (a full file stays full until the next fresh one); not decided in a start that only prepares. `vk: pipeline cache: libraries only (...)`. `2`: lean at every fresh file whatever the PC (the lab's). Whole pipelines then load 3-4 times slower from the driver's shared, capped cache (`gpu/vk/vk_draw.cpp` `pcache_libs_decide`, `build_graphics`, `pcache_play_open`) |
|
||||
| `RX_VK_PCACHE_CARRY_MB=<MB>` | (Triforce, 11-18) The carry of the pipeline cache's file over an update that changed the translation (`gpu/vk/pcache_bound.h` `carried`): a file for other shaders is read all the same when it is no larger than this (1536 MB unset), leaves room under the bound for as much again and the PC has 31 GB of RAM or more (Windows), so the driver finds what it compiled for the shaders that did not change (AMD's driver keeps no cache of its own for these). `0`: never carried, the rule of 11-14 alone |
|
||||
| `RX_VK_PCACHE_PLAY_FILE=0`, `RX_VK_PCACHE_FOLD_MB=<MB>` | The pipeline cache's play file (`gpu/vk/pcache_play.h`, NX2navi's play cache): once the preparation is done, what a run compiles goes into a second cache with a file of its own, `mkw_vk_pipelines_play.bin` beside the first (a creation asks the first cache without compiling, then the play cache), and a save writes that small file; `mkw_vk_pipelines.bin` is written only when the preparation built into it. Before, a save after a few new pipelines wrote the whole file again (2.6 GB at the owner's start after its precompile). At a start whose play file has reached `RX_VK_PCACHE_FOLD_MB` (64) it is folded into the first cache, that session builds into it as before, and its first save writes it and removes the play file. Needs `pipelineCreationCacheControl`. `0`: no play file, as before |
|
||||
| `RX_VK_PCACHE_OWN_PASSES=0` | (Triforce, 11-22, the performance audit's PIPE-95, after Source A's) The back end's own compute passes (FSR 3.1's dozen and its HDR output pass, the anti-aliasing's, the texture converters') are made with a bare creation into the main pipeline cache, as before. The default makes them the way a dispatch's pipeline is made (`gpu/vk/vk_draw.cpp` `shared_cache_compute`, `cached_compute`; `gpu/vk/pcache_play.h`): the main cache asked without compiling, a miss compiled into the play cache, a driver fault caught (the pass is then not made). Bare, they grew the main cache by a little, whose file is written in play only after 64 MB of growth, so they waited for the exit's save of the whole main file (650 MB fresh, 2 to 3 GB with the stage libraries), and an exit that cut that write short had them compiled again at the next start, inside the first upscaled frame. Without a play cache (no `pipelineCreationCacheControl`, or `RX_VK_PCACHE_PLAY_FILE=0`) only the guard is new |
|
||||
| `RX_VK_WARM_SAVE_MIN=<minutes>` | (Triforce) Checkpoint saves of the pipeline cache during the shader preparation's long builds (`gpu/vk/warm_checkpoint.h`, `tests/test_warm_checkpoint.py`): while the driver's part (the stage libraries, about two hours cold on an RX 7900 XT) and the precompile build, `mkw_vk_pipelines.bin` is saved every 10 minutes once at least 1000 were built since the last save, so that a restart, a crash, a power cut or a kill goes on from the last checkpoint (what it saved comes back as hits) instead of from nothing. Each goes through the preparation's own save: the file's bound, a carried session's cap with its history kept open as a stopped one's (`RX_VK_PCACHE_CARRY`), the file written whole with its `.for` and `.base`; it is left out while a copy of the cache would not fit in the RAM available with 2 GB to spare or the disk lacks its size and 1 GB (asked again a minute later), fetched with room for what is built meanwhile, written to the disk before it takes the old file's place, given up when the window is closed or Setup's Cancel is pressed (the stop's own save writes the cache), due only while the part moves (a stuck one goes silent), and it does not count the part as made (the next start still runs it). Beside the file, `.own` says it holds this part's pipelines alone, so that a part over the file's bound after a crash ends as a new file over the bound, not as an older build's file made again from nothing. The log says `vk: the driver's part so far saved for a restart: N MB in X s`. Not on Intel unless `RX_VK_PCACHE_PLAY_SAVE=1` (its driver crashed fetching the cache while pipelines were built with it); `RX_VK_PCACHE_PLAY_SAVE=0`: on no card. `0`: saved at the end alone, as before |
|
||||
| `RX_VK_PRECOMPILE_SAVE_S=<seconds>` | (Triforce, 11-22, the performance audit's PIPE-23, after Source A's; `gpu/vk/precompile_save.h`) A precompile that keeps its pipelines and built for this many seconds or more (30 by default, at most a day) saves the pipeline cache before the title, and says what that wrote: `vk: precompile: N s of building: the pipeline cache saved before the title (... KB in ... ms)`, or why nothing was. It was first saved in play (present 300, after 32 more builds; on Intel only at a clean close), so a crash, a kill or a power cut early in play cost the whole build again at the next start. The save is the play file's alone while there is a play cache (`gpu/vk/pcache_play.h`): the main file, 2 to 3 GB with the stage libraries, is not written before the title. Not when the game went on while workers still build, after a stop, or in a start that only prepares (those save by themselves); the precompile that keeps nothing saves before its record, as before. Also: a preparation without stage libraries counts the compute pipelines it builds into the main cache for that file's save, and saves the cache at its end before its record says the part is made (it ended unsaved). `0`: none of this, as before (`gpu/vk/vk_draw.cpp` `precompile_end_save`, `prepare_compute_pipeline`, `prepare_run`, `pipeline_cache_save`) |
|
||||
| `RX_VK_FETCH_BOUNDS=0` or `1`, `RX_VK_NAN_BITS=0` or `1`, `RX_VK_GATHER_OFFSET=0` or `1` | The translation's vendor and driver workarounds (`gpu/vk/vk_vendor.h`) off or on for one run, whatever the GPU; the log's `vk: vendor fixes` line says which are on. By default: out-of-bounds texel fetches answered with 0 on every GPU but NVIDIA's; NaN tests on a value's bits on AMD's older Windows shader compiler (RX 5000 and RX 6000); gathers nudged by 1/512 of a texel on AMD's and Intel's drivers. They change the shaders' translation (one preparation of the shaders) |
|
||||
| `RX_VK_MIRROR=0`, `RX_VK_MIRROR_MB=<MB>` | No VRAM mirrors of the game's memory pools (on for discrete cards); their budget |
|
||||
| `RX_VK_VRAM_MB=<MB>` | The texture cache's budget (default 70% of what the driver allows, at most 6 GB) |
|
||||
| `RX_VK_NAN_SELECTS=0` | (Triforce, after NX2navi's e38f42a, issue #156) Where the NaN tests are made on the bits (`RX_VK_NAN_BITS`: by default AMD's older Windows shader compiler, RX 5000 and RX 6000), a float select of a constant whose bits are a NaN (an integer mask the translation types as a float) is also made a select of the bits, so that compiler cannot fold it into stray black or white blocks (`gpu/vk/spv_nan_select.h`: 32-bit selects, and 16-bit ones in a module that has a 16-bit integer type). It is applied where a module is made (`make_module`), outside the translation stamp: no program is translated again, nothing changes on any other card, and on those cards only the modules that have such a select are new to the driver's caches. So that those are compiled before the game rather than at their first draw or dispatch, the driver's part of the preparation, the light precompile's record and the predicted pipelines' record have a stamp of their own there: the first start on those cards goes through them once more, everything but those modules coming back from the caches (the variants on the driver's part's skip list are carried over). `RX_VK_NAN_SELECTS=0` leaves the selects as they are (the NaN tests on bits stay). The log says `vk: NaN selects: module <crc>, <n> float select(s) of a NaN constant made select(s) of the bits` for the first four modules and every 64th |
|
||||
| `RX_VK_VB_RANGE=0`, `RX_VK_MESH_VB_REBIND=1` | A draw's vertex streams (`vk_draw.cpp` `draw`). `RX_VK_VB_RANGE=0`: all 16 bound again whenever one changed, as before v0.6.0 (by default only the ones that changed, `vertex_bind_range`). `RX_VK_MESH_VB_REBIND=1` (off by default: about 3% at the Market in the lab, and its tickets not reproduced): after a mesh pipeline's bind the next vertex draw binds all 16 streams again with their strides (by default they stay as they were bound, as before: a mesh pipeline has the strides static, and its bind leaves them undefined for the next pipeline that has them dynamic, as it does the topology; `tests/test_draw_binds.py`) |
|
||||
| `RX_VK_MIRROR=0`, `RX_VK_MIRROR_MB=<MB>` | No VRAM mirrors of the game's memory pools (on for discrete cards); their budget (unset: by the card, `gpu/vk/mirror_budget.h`: a quarter of the video memory on a discrete card over 4.5 GB, at most 3 GB; an eighth on cards of 4.5 GB or less and on integrated GPUs up to 6.5 GB; the log's `vk: VRAM mirrors' budget` line. `=768` is what a 6 GB card had before) |
|
||||
| `RX_VK_MIRROR=0`, `RX_VK_MIRROR_MB=<MB>` | No VRAM mirrors of the game's memory pools (on for discrete cards); their budget. Unset, the budget goes by the card (`gpu/vk/mirror_budget.h`): a quarter, at most 3072 MB; an eighth on cards of 4.5 GB or less and on integrated GPUs up to 6.5 GB. Discrete cards between 4.5 and 6.5 GB had the eighth before 11-21: `RX_VK_MIRROR_MB=768` is a 6 GB card's old size. (Triforce, 11-22, the performance audit's MEM-51) Without mirrors (`RX_VK_MIRROR=0`, Intel's integrated GPUs) the replay no longer notes what only a mirror's write-back reads (`gpu/vk/vk_memory.cpp` `wb_notes_off`: the report words of `memory_cpu_owned`, the written pages of `shadow_written_to` and `guest_cpu_wrote`): those lists grew to 65,536 words and 8 million pages, were dropped and grew again. No switch of its own: with mirrors on everything is noted as before |
|
||||
| `RX_VK_DEMOTE_QUICK=0` | A pool the CPU and the GPU both write is tried in guest memory (`vk_memory.cpp` `demote_trials`); by default a trial whose GPU time a frame is over 3 times its baseline and 5 ms more, after 3 frames or more, ends then, settling or measured, and is judged as a slower one is (mirrored again, measured again: `gpu/vk/demote_watch.h`). `0`: every trial runs its 10 settling and its measured frames, as before (in the owner's logs, some twenty frames of 100 ms for a pool at 106 ms against 7.4) |
|
||||
| `RX_VK_IMPORT_CHUNK_MB=<MB>` | The game's memory is imported in whole chunks of its address space, 16 MB each (`0`: by the game's own pools). (Triforce, 11-22, the performance audit's MEM-16, after Source A's) Unset, the chunk is 4 MB where the mirrors are on and whole on an NVIDIA card (NVIDIA before Turing by default, a card without sparse residency buffers, `RX_VK_MIRROR_PIECES=0`): a whole mirror copies all of its import into video memory, the padding too, and Source A measured about a quarter less mirror memory with 4 MB for about half again as many imports. With the mirrors in pieces, without mirrors, and on AMD's and Intel's cards it stays 16 MB (an import stalled the replay 20 to 75 ms on an AMD card). The log says `vk: imports in chunks of N MB` once. An imports-ahead file written at 16 MB is read as it is. `16`: as before on every card |
|
||||
| `RX_VK_IMPORT_AHEAD=1` or `0` | On by default with NVIDIA's driver (`=0` off). Off for test12 to test14: the GPU hung at the title in 6 of 12 lab runs; the cause, the game's freed pages decommitted under an import the background thread had not finished (`bg_held`), is closed and the lab's A/B with it was clean in 10 of 10 (`gpu/vk/import_ahead.h`). The ranges of the game's memory a run imported on a binding's own thread (8 to 17 ms each on an RTX 5090), whole as they were imported (their span over the game's pools included; the first version kept 16 MB chunks, which a binding across their ends imported again as one), are kept in `triforce_imports_ahead.txt` beside the game, at most 256, and the next runs import them on the background import thread as soon as the game's pools reach them, 4 at a look, at each change of the pool table and every 8 submissions (`gpu/vk/import_ahead.h`), each range once a session, and none while the PC's RAM is low (under 3 GB available, or a tenth of the RAM: the RAM's rule gives idle imports back then, and the two took turns on the same ranges in the owner's test9 session). A file of the first version's chunks is read as it is. Unset: off (it was on with NVIDIA's driver only, as `RX_VK_ASYNC_IMPORT` is); `1`: on any card; `0`: off. (Triforce) A race that fits the hang is closed since: the host decommitted the game's freed pages under an import the background thread had locked and not yet put in the pool table (`vk_memory.cpp` `bg_held`; `RX_VK_BG_IMPORT_HOLD=0`: not, as before, for the A/B); it stays off until a lab A/B with `=1` hangs no more. The log's `vk: imports ahead` lines count them and name the ranges sent, `vk: the game's pages ... are freed under a background import` the frees that window met; the stall line's "pools imported where bound" is what is left |
|
||||
| `RX_VK_IMPORT_TOUCH=0` | Off: the pages of a range of the game's memory are not read before the driver imports it. On by default, on Windows, for the imports made on the background thread (`gpu/vk/import_touch.h`): the driver locks a range's pages inside `vkAllocateMemory`, faulting in the untouched ones, and a queue submit from another thread waits for it; the ranges imported ahead are up to 1 GB. Read first, outside the driver, the allocate of the 288 to 752 MB ranges went from 16 to 93 ms to 4 to 13 ms on an RTX 5090, and the longest queue submit beside an import in play from 71 to 93 ms to 10 to 12 ms. The log's `vk: import` lines give a slow import's parts: the first 12 of a session, and the first 48 of 64 MB or more. |
|
||||
| `RX_VK_IMPORT_LEAVING=0` | (Triforce, 11-22, after Casual Gamer's #dev-fixes post) The game's freed pages under an import already given back whose memory is not freed yet (an evicted pool, a retired pool's import given back) are discarded or decommitted, as before. By default they are zeroed in place, as under a live import (`gpu/vk/vk_memory.cpp` `memory_guest_imported`, `imports_leaving_overlap`): the driver keeps those pages locked until the GPU is past them, and a discard there lost the device (v0.4.0 on a 16 GB laptop: 2 tours of 5 without, 0 of 53 with). The log says `freed under an import still being given back` for the first 8 and every 64th (`tests/test_import_leaving.py`). |
|
||||
| `RX_VK_KEEP_RETIRED=1` or `0` | A retired pool's import (a pool merged into a larger one) kept for the session (1) or given back once the GPU is past it (0; `gpu/vk/vk_memory.cpp` `retired_imports_give_back`). Unset: given back on Linux and with NVIDIA's driver, kept on Windows with AMD's and Intel's (device losses right after they went back there, issues 17, 309, 571). (Triforce, 11-22, after Casual Gamer's #dev-fixes post) Unset, those kept are given back all the same past the import rule's high mark of the system heap's budget (`RX_VK_IMPORT_HIGH`, 85%) or for a minute after an import was refused for want of room (`gpu/vk/mem_parts.h` `retired_pressed`; the log says `retired imports given back after all`): on a Radeon 780M the retired imports filled that heap, and 24.9% of the draws were left out for want of their pools (0 with it; refused imports 16 -> 1). `=1` keeps them always, as before. |
|
||||
| `RX_VK_CAP_SPARE_MB=<MB>` or `0` | The submit-time capture's spare buffers, kept for the next captures (12 at most, `gpu/vk/vk_memory.cpp` `cap_spare_max`), held to that many MB; `0`: up to 768 MB, as before. Unset: 32 MB on a card the driver allows 6.5 GB or less, and (Triforce, 11-22, after Casual Gamer's #dev-fixes post) on a PC of under 20 GB of RAM (`gpu/vk/mem_parts.h` `spare_capped`; the log says `capture buffers: 32 MB of spares at most`): on a 16 GB laptop the spares held ~0.45 GB that Windows paged out and in again, capped 10-18% fewer page-ins. Not on AMD's cards, where a capture buffer made anew costs about 100 ms: they keep their spares |
|
||||
| `RX_VK_IMPORT_RAM=1` | The imports' second trigger (`gpu/vk/vk_memory.cpp` `imports_give_back`): with the PC's RAM short (under 1.5 GB available) every import unbound for ~10 s of play is given back, up to 1 GB a pass. (Triforce, 11-22, after Casual Gamer's #dev-fixes post) Off unless set: a 16 GB PC has under 1.5 GB available nearly all the time, so it fired every ~20 s and the imports it gave back came straight back (690 to 1216 imports in a 13-minute tour). The system heap's trigger (`RX_VK_IMPORT_HIGH`) stays as it is. |
|
||||
| `RX_VK_KEEP_RETIRED=wait`, `RX_VK_RETIRED_REFUSED=1` | (Triforce, 11-22, the performance audit's MEM-21, after Source A's; `gpu/vk/retired_give_back.h`; **off unless set**, not yet run in the game) A pool remade larger is retired, and once the GPU is past it its import is pinned RAM nothing binds. Unset, as before: given back at once on Linux and with NVIDIA's driver on Windows, kept for the session with AMD's and Intel's on Windows (given back at once there, an Arc B390, an Ally X and an RX 6600 lost the device right after); `RX_VK_KEEP_RETIRED=0` gives back at once and any other value keeps, as before. `wait`, on any driver: given back `RX_VK_RETIRE_SECS` (30) seconds after the pool retired, `RX_VK_RETIRE_MIN_SECS` (2) while the system heap is past 85% of its budget; with it the imports given back and not freed yet are left out of what the pressure give-back (`RX_VK_IMPORT_HIGH`) reads. `RX_VK_RETIRED_REFUSED=1`: also the retired pools whose mirror was refused (`RX_VK_NO_MIRROR_POOLS`, a pool over 1 GB, video memory refused), never given back before on any OS; on with `wait` unless `=0`. The `vk: imports:` line every 1200 mirror epochs says how many retired pools are still imported |
|
||||
| `HOS_CRASH_THREADS=0` | Off: a crash report has the faulting thread alone. On by default, on Windows (`host/crash_threads.h`): a crash or fatal error's report ends with one `CRASH: beside it, thread ...` line for every other thread, its host stack as module and offset, taken at the report's start on a helper thread; the threads running code come first, then the ones waiting in the system. For the crash while a save loads, which reads an object whose memory holds something else: the lines say what the other threads were doing. |
|
||||
| `HOS_CRASH_TEST_S=<seconds>` | (the lab's) A host thread reads address 8 that long after the start: a crash report on purpose, to look at one from a run in the game (its other threads' lines above). Not for play. |
|
||||
| `EDEN_NX2_GM_LEAN=1` | Only the global-memory helpers a program's instructions call are put into its SPIR-V (`gpu/vk/eden/shader_recompiler/backend/spirv/gm_lean.h`). A stage that reads or writes through a pointer no storage buffer was tracked to carried every helper, six and for a storing stage eighteen, each with a search over every storage buffer in it. Off unless set: the SPIR-V is then what it was. An `EDEN_` name: each setting has its own program cache. The translator's files are in the translation stamp, so a build with this code translates every program again once, whatever the switch says. |
|
||||
| `RX_VK_UPLOAD_AGAIN_MS=<ms>` | A texture uploaded before whose memory the game wrote again, met by a draw after the frame's uploads took this long (16), is uploaded at the next frame instead; its last bytes are drawn meanwhile, at most 8 frames in a row (`gpu/vk/upload_owed.h` `later`). New textures are never left so (theirs is `RX_VK_HOLD_MS`'s rule). `0`: every one at once, as before. The log's "left for the next frame" lines and the stall line count them. `RX_VK_UPLOAD_AGAIN_FRAMES=<n>`: a texture's upload is left so at most that many frames in a row (2: its last bytes drawn a frame or two at most; `8`: as before). With 16 ms nothing was ever left in the lab's runs (the frames with re-upload bursts took 9.8 ms of uploads); `3` spreads a burst over the frames after it, its textures a frame or two late meanwhile, which is why it is not the default |
|
||||
| `RX_VK_UPLOAD_TORN=0` | (Triforce, wp18-flash) A texture uploaded before, 1 MB or more, whose storage the CPU has rewritten since the copy its re-upload would read (a capture's copy of one of its whole pages, every 16th and the last, differs from guest memory at the draw's look) keeps its last bytes for that frame and is uploaded at the next frame with what that frame's capture took (`gpu/vk/upload_torn.h`): a copy taken while the game was still writing the texture was test29's white frames at the Market after a tour (BC6H bytes passing through, decoded to bright colours the bloom spread over the picture). At most one frame in a row, and not for a texture found so in the two presents before (one the CPU rewrites every frame is uploaded as before). Not a dispatch's upload. `0`: uploaded at once, as before. The log's `re-uploads of storage the CPU rewrote since its capture` lines count them, those found and those left (with `0` too: the measure of how often a copy may have been taken mid-write) |
|
||||
| `RX_VK_REUPLOAD_EARLY=0` | The textures the CPU keeps rewriting (uploaded again from its writes twice or more within 120 presents, 1 MB or more, sampled in the frame or the one before, not drawn into by the GPU) are looked at once on the replay thread before a parallel replay's segments begin, after the dry pass, their uploads in the recording before the segments: not made by a segment under the recording lock, where the other recording threads queued behind them (lab S2 after a tour: 4 to 12 such uploads of 28 to 84 MB in ~13% of the frames, the descriptor time 24 ms over the threads against 3). `0`: the segments' looks alone, as before. The log's `textures the CPU keeps rewriting, looked at before` line says how many and how long (`gpu/vk/reupload_early.h`) |
|
||||
| `RX_VK_ZERO_HOLES=0` | (Triforce, Linux, on by default; `=0` as before) The all-zero check of a big texture's storage (32 MB or more: cleared on the GPU when every byte is 0; `RX_VK_ZERO_UPLOAD=0` turns the check off) reads only what the guest memfd holds data for (`lseek` SEEK_DATA and SEEK_HOLE, `gpu/vk/guest_holes.h`), not its holes: a page the game never wrote reads as zero, but a read through the shared mapping made it RAM (the 288 MB shadow map at each Hyrule Field arrival, 200-300 ms on a Steam Deck). Where the window is not a memfd, or a seek fails, the whole storage is read. The `is all zero` and `is not all zero` lines say how many MB were read. `=0`: the whole storage always |
|
||||
| `RX_VK_UPLOAD_SAME=0` | (Triforce) A texture uploaded before whose storage the CPU wrote again is checked against the bytes its image holds (`gpu/vk/upload_same.h`): the units of its storage the writes reached (each level's and layer's GOB-block rows, as its upload reads them) whose bytes hash as at their last upload are not uploaded, and when that is all of them nothing is. The upload of the rest reads the copy the check hashed, so that what is kept is what the image got; any other upload and any GPU write forget it. Textures of 16 MB or less uploaded on the CPU (block-compressed, or under 8 MB), not sparse; one whose checks find nothing the same 4 times in a row is uploaded unchecked for its next 64 re-uploads. Lab S2 after some tours: the game rewrote a few big BC textures the Market samples with the bytes they had, 40 to 60 MB every 4 to 5 frames (the Market at 62-78 fps against 83-96, ~26 GB of uploads in its stay). `0`: every re-upload made as before. The log's `re-uploads checked against the bytes their images hold` line says what it looked at, found the same and uploaded. `RX_VK_UPLOAD_SAME_KEEP=0`: the check hashes every byte it looks at; by default a page the submit-time capture keeps a copy of is hashed once and its hash kept with the copy until a capture copies other bytes into it (about half of those textures' pages passed through other bytes between two captures) |
|
||||
| `RX_VK_UPLOAD_SAME_ASTC=0` | (Triforce) ASTC textures decoded on the GPU (kept as RGBA8, BC3 or BC1; 2D only) are checked too, by `RX_VK_UPLOAD_SAME`'s check: one unit a layer, its levels from the first one the image keeps, which is what the GPU decoder stages of a layer a write reached; the staging reads the copy the check hashed. Lab S2's Market after a tour: the game rewrote such textures with the bytes they had (1536 x 1536 ASTC 5x5 of 11 levels, 768 x 768 kept as BC3, 512 x 512), staged and decoded again unchecked, about 1.4 GB in 10 s (90 fps down to 72, the 1% low 57 to 33). `0`: ASTC re-uploads made as before, every layer a write reached staged and decoded. The `re-uploads checked` line says what of it was ASTC; the `re-uploads not checked` line every 600 frames counts the re-uploads made without the check, by route and by why |
|
||||
| `RX_VK_TWIN_SAME=0` | (Triforce) The parallel replay's twin (a texture an earlier segment samples that a later segment of the same submission uploaded, uploaded again in the earlier segment's uploads so that its draws see the bytes as submitted) uploads only what that submission's uploads into the image asked for (`gpu/vk/upload_same.h` `note_upload`; the whole storage when one of them was whole or the image was made again), and where every unit it reads has a hash kept by `RX_VK_UPLOAD_SAME`'s check, reads a copy of them hashed as copied: when the copy holds the bytes the hashes are of, the hashes are kept. Before, the twin uploaded the whole storage and, unchecked, forgot the hashes the later segment's checked upload had just kept, so the next rewrite of the same bytes was uploaded again, the twin with it, every few frames: lab S2's Market after a tour, about 97% of what was still uploaded unchecked in those loops, up to 6.9 GB in one stay. `0`: the twin uploads the whole storage and forgets the hashes, as before. The log's `the parallel replay's twins` line every 600 frames counts them, the bytes uploaded against the whole, and how many kept the hashes |
|
||||
| `RX_VK_UPLOAD_EARLIEST=0` | (Triforce) A parallel replay's upload of a texture, made by the worker of a later segment of the submission while an earlier one is still being recorded, is recorded into the init command buffer of the earliest segment still being recorded (submitted before that segment's uploads and commands; `gpu/vk/upload_earliest.h`, `vk_texture.cpp` `UploadRedirect`). Every segment before it has ended its recording and no longer samples the texture, and every one from it on finds the upload at or before its place in the queue, so no earlier segment uploads the texture again (the twin, `RX_VK_TWIN_SAME`): it is uploaded once, the same bytes. Lab S2's heavy Market loops (round 48) uploaded 2.1 GB of twins in their last minutes, the big streamed or new textures each uploaded once by a later segment and again by an earlier one; those twins took the frames' upload budget (`RX_VK_UPLOAD_NEW_MB`, `RX_VK_UPLOAD_NEW_MS`), and the new textures after them were left for the next frame, the picture held. Only an upload staged from the CPU as it is recorded is moved: not one read on the GPU (the GPU unswizzle, from the pool) or decided from live memory (the all-zero clear), which in the earlier segment's init would run before a segment between writes that memory on the GPU. Not for a texture the GPU drew (its contents are the GPU's; the CPU's writes over it, `RX_VK_CPU_OVER_GPU`, keep their own path) or a sparse one: those keep the segment's own uploads, and the twin. The uploads' GPU time in an init command buffer is not in the segments' GPU time. `0`: every upload in its segment's own uploads, the twin as before. The log's `uploads recorded into an earlier segment's init command buffer` line every 600 frames (with `0` too) counts them and the bytes, and how often and how long a segment's end waited for its init lock, each segment's own, held while such an upload (or a new image's first layout change) is recorded into its init (its submit waits as long) |
|
||||
| `RX_VK_SHADOW_STAGE=0`, `RX_VK_CAP_UNCACHED=1` | (on by default; the row is Triforce's, 11-22, with the performance audit's MEM-39) A capture buffer asks for cached host memory, but a driver can hand out write-combined memory, which the CPU reads at about 0.3 GB/s. Such a capture's texture pages are copied into a staging block in ordinary memory first, and to the buffer from there, and every CPU read of them uses the staging (`gpu/vk/vk_memory.cpp` `capture_make`, `cap_page_bytes`): the shadow's copies, and since MEM-39 the `RX_VK_TEX_SAME` compare of each copied texture page with its shadow copy, which still read the buffer. `RX_VK_SHADOW_STAGE=0`: no staging, every read from the capture buffer, as before. Where the buffers are cached (the 600-take `shadow staging` line counts them) nothing is staged and nothing changes; the first buffer that is not cached is logged with its memory type. Lab only: `RX_VK_CAP_UNCACHED=1` asks for the capture buffers as memory written in order, which is uncached where the driver has such a type, to make the slow case on any PC, and the `RX_VK_LAB_AB` knob `texsamemap` reads the buffer in the compare in its phases |
|
||||
| `RX_VK_SHADOW_KEEP=0` | (Triforce) A texture page the submit-time capture takes with the bytes its shadow copy holds (the copy the uploads read, `RX_VK_TEX_SAME`'s compare) keeps that copy as it is, not copied again, its kept hash with it (`gpu/vk/shadow_prev.h`). Lab S2's Market after some tours: about half of the ~400 texture pages a capture took, six captures a frame, held their copy's bytes. Only for a page the capture takes once; any other is copied as before, and the shadow holds what the capture holds either way. `0`: every page copied, as before. The log's `texture shadow (600 takes)` line counts the pages compared, left, given back and copied |
|
||||
| `RX_VK_SHADOW_PREV=0` | (Triforce) A texture page the capture takes with other bytes keeps the copy it had, and that copy's hash, beside the new one (`gpu/vk/shadow_prev.h` `Prev`), and a capture that brings those bytes back gives it back, hash and all: no copy, and the re-upload check after it finds the hashes kept (lab S2's Market after some tours: the other half of those pages went back and forth between two versions, and a check hashed each again after every change, 0.2 to 0.3 ms a frame). A kept copy that does not come back within 8 changes of its page, or within 64 captures, goes; at most 64 MB are kept. `0`: none kept, a changed page copied over as before. The `texture shadow` line says how many are kept and how many went |
|
||||
| `RX_VK_SHADOW_VERS` | (Triforce) The copies a texture page keeps beside its own in the shadow (`gpu/vk/shadow_prev.h` `Vers`): `RX_VK_SHADOW_PREV` keeps one, and a page whose captures bring more than two versions (A, B, C, A every few frames) is copied again for every version but the one before the last. `2` to `4` keep that many, newest first, and a capture that brings any kept version's bytes takes that one back, the copy it had becoming the newest, so each version a page cycles through while it is still among the kept copies is copied once: 4 KB a version, under the same 64 MB cap. Unset or `1`: one copy, as before. The `texture shadow` line counts the pages given a copy of their own, one before it, and two or more before |
|
||||
| `RX_VK_SHADOW_CHECK=1` | (Triforce) Each texture page `RX_VK_SHADOW_KEEP` or `RX_VK_SHADOW_PREV` left uncopied is compared with the capture's copy again after the capture is taken, and the hashes kept with the shadow's copies are hashed again: the `texture shadow check` line every 600 captures counts the copies that differ and the hashes that are not of their bytes, both of which must be 0. Slow: for the lab's checks. Unset or `0`: not checked |
|
||||
| `RX_VK_HOLD_MS=<ms>` | Hold the last picture (`gpu/vk/present_hold.h`, on by default): a new texture a draw samples after the frame's uploads took their budget (`RX_VK_UPLOAD_NEW_MS`, `RX_VK_UPLOAD_NEW_MB`) is uploaded at the next frame, the dummy drawn in its place, and that frame is not shown: the window gets the last complete picture again, the Insert menu, the counters and the notices drawn over it, until a frame is drawn with every texture complete. The game runs on underneath. No frame generation makes a frame from a held one, and the one shown after a hold starts its history anew (and the temporal upscalers theirs). `<ms>`: the hold's bound (250); at it the rest are uploaded at once and the next frame is shown whatever it holds, and no hold comes for 120 presents (twice as long after each hold at its bound that follows, up to 1920). `0`: no hold, no new texture left for later: the one frame that takes them all, as before. The log's `vk: hold:` lines, the stall line and the 600-frame line count them |
|
||||
| `RX_VK_HOLD_PIPELINES=0` | A frame with draws or dispatches left out because their program, pipeline or compute pipeline was still being made is held as one with a texture still owed its bytes (`RX_VK_HOLD_MS`'s hold, with its bound and its rests; NX2navi's frame hold, `gpu/vk/pipe_wait.h`): by default the window keeps the last complete picture instead of showing one that lacks a pass. `0`: such frames are shown, as before |
|
||||
| `RX_VK_UPLOAD_NEW_MS=<ms>`, `RX_VK_UPLOAD_NEW_MB=<MB>` | The new textures' budget a frame while the hold is on: once the frame's uploads took 12 ms or 64 MB, a new texture is left for the next frame (ASTC ones too; the frame's first upload always goes). 12 ms keeps an area change's frames near 16.7 ms with their draws and memory sync; 64 MB is what the GPU copies or decodes in a few milliseconds, so that no frame leaves it a backlog. `0`: that half of the budget off |
|
||||
| `RX_VK_UPLOAD_SPLIT_MB=<MB>` | A new layered ASTC texture whose texels are more than this (128) has its first upload in parts while the hold is on, the texture's layers over at most 6 frames (`gpu/vk/layer_split.h`): its draws sample the dummy and its frames are held until the last part is in; for a dispatch or at the hold's bound what is left goes at once. In the owner's test7 session one 1365 x 1365 x 90 texture kept as BC3 took one frame's upload command buffer 114.8 ms of GPU time. `0`: whole, as before. The stall line counts the parts |
|
||||
| `RX_VK_DUMMY_GEN=0` | A frame that left textures for later makes every texture view a draw resolved stale (`destroy_gen`), as before. By default only the looks that resolved to the dummy for want of their texture are (`texture_dummy_generation`, `vk_draw.cpp` `texture_view_of`); the stall line counts the views resolved anew |
|
||||
| `RX_VK_VIEW_KEEP=0` | Any texture's end (an eviction, a dead render target let go, a texture made BC3 from BC1, the compaction moving one) makes every texture look of every recording thread stale, and each is resolved anew when next met (the slot, the cache, the view, under the parallel replay's lock), as before. By default a thread that finds the destroy generation moved keeps the looks of the textures whose views did not go: those are noted by their creation number before the generation moves (`gpu/vk/view_keep.h`, `vk_draw.cpp` `texture_view_of`); never a look of no texture or no view, nor a flat texture's 3D copy. More than 1024 gone since a thread's last pass: all resolved anew. The log's `vk: texture looks kept` lines count them, the stall line the views resolved anew. After NX2navi's `tex_gen_undo` |
|
||||
| `RX_VK_DESC_KEEP=0` or `check` | The per-binding reuse of a draw's descriptors (`gpu/vk/desc_keep.h`, on by default): a draw that builds its descriptors takes, for each sampled texture, user uniform buffer, stage's constants and read-only storage buffer, what the program's last draw on that thread resolved it to, while everything it was resolved from is the same: the texture's handle, its slot's registration, the samplers', no texture destroyed or made again, no texture left for later and the same submission; the buffer's address and size, no inline uniform update, no pool moved or mirror made, the same recording. `0`: every descriptor resolved, as before (so also with `RX_VK_NO_DESC_REUSE`, `RX_VK_SAMPLER_CACHE=0` and the descriptor diagnostics). `check`: every descriptor resolved too, and one that comes out otherwise than what was kept for the same inputs is counted and logged (`vk: descriptor reuse check`): the way to try the rules. The frame statistics' `a frame's descriptors` line says how many were resolved and taken, how many pushes were whole, of the changed bindings or of nothing, and how many sets were written, through a template, or bound again |
|
||||
| `RX_VK_DESC_SIG=0` or `check` | The whole-draw skip by the signature of a draw's descriptor inputs (`gpu/vk/desc_sig.h`, on by default with the per-binding reuse): a draw whose bindings moved, of the program whose descriptors its last draw pushed in that command buffer, first reads what each of them is resolved from (a texture's handle and its slot's version, a user uniform buffer's address and size, a read-only storage buffer's address and size words, a driver bank's stage generation) and the stamps the per-binding reuse gives each kind; when that equals the signature the pushed descriptors were built from, and their build kept every one, the build and the push are left out (a build would take each of them again and push nothing). Push programs only, of the kinds the per-binding reuse keeps: not those with storage buffers they write, tables, texel buffers or storage images, not the mesh fallback or RenoDX's. `0`: built, as before. `check`: built too, and one that comes out otherwise than what is pushed is counted and logged (`vk: descriptor signature check`). The frame statistics' `a frame's descriptor signatures` line says how many draws read theirs and how many were kept whole |
|
||||
| `RX_VK_DESC_COUNTS=0` | (Triforce, 11-22, the performance audit's DESC-22, after Source A's; `gpu/vk/desc_counts.h`) `build_descriptors` walks the program's stages at every build to sum its buffers, images and texel views, as before. The default keeps the three sums packed on the program (`Program::desc_n`, 10 bits each, in the padding after `ready` so that `sizeof(Program)`, which the shader caches' stamp holds, stays what it is): stored at the program's first build once it is translated, read at every later one. A program with more than 1023 of a kind walks at every build. A build that went by kept counts and had one of its three vectors move is made again, walked, and that program walks from then on (one log line). The lab's knob `descwalk` (`RX_VK_LAB_AB`) is this switch in its phases |
|
||||
| `RX_VK_PROF_ALL=1` (the dispatch's pick clock) | (Triforce, 11-22, the performance audit's DESC-30, after Source A's and Source B's) Every profiling scope timed, and with it the clock pair around each compute dispatch's program pick, as before. Unset, the pick's time for the stall line is the change of the thread's exactly timed nanoseconds across the pick (its slow path is timed by `PROF_EXACT`), so a pick the memo answers reads no clock. DESC-30 has no switch of its own, the results being the same: besides the pick, `taau_key` answers from its list parsed once (it wrote the key out and walked the list as text several times a dispatch), `RX_VK_LOG_PUSH` formats its key only when set, the first-pipeline clock is read only when there is one to make, and `texture_frame` fills `RX_VK_LOG_BAKED`'s list only for that switch (read once) and finds its largest textures by a partial sort. The lab's knob `diagold` (`RX_VK_LAB_AB`) puts all of it back in its phases (`gpu/vk/vk_draw.cpp` `dispatch`, `taau_key`, `dispatch_diag_old`; `gpu/vk/vk_texture.cpp` `texture_frame`) |
|
||||
| `RX_VK_LAB_AB=e_draw:<ns>`, `e_dispatch:<ns>`, `d_present_after:<us>` | (Triforce, 11-22, the performance audit's REPLAY-51, after Source B's; lab knobs of the in-run A/B, `gpu/vk/lab_ab.h`, nothing without the variable) The `d_<site>` knobs wait once a frame; `e_draw` and `e_dispatch` wait the knob's nanoseconds at every draw or dispatch, first in `draw()` and `dispatch()` so the profiles leave it out, on whichever thread records, and count each wait. The frame's change against the time spun says how much of that path is on the frame's critical path (200 ns at about 3,050 draws a frame is 0.6 ms of work). A phase's end logs `vk: lab A/B: phase e_draw:<ns> over:` with its waits and ms a second. `d_present_after` is `d_present`'s control: the same wait just after the present's `drain(0)`, where `d_present` is just before it. Pin `RX_VK_PARALLEL` with the `e_` knobs (the log warns otherwise): Automatic moves draws between threads. Off: one relaxed load a draw. `tests/test_lab_ab_each.py` |
|
||||
| `HOS_OOT_SHADOW_STEP=<n>` | (Triforce, 11-21, the performance audit's GAME-05) The game gives each nearby character or object a square depth shadow map sized by its size on screen, 384 to 1,920 a side, so while anything moves each frame asks for depth textures of sides not seen before. The one read of the side is rounded up to a multiple of `<n>` (a power of two up to 256; 64 when not set): 25 sizes, which come round again, so the render targets' pool serves them. A map gets at most 63 more texels a side; its projection does not change. `0` or `1`: the game's own side, as before (`host/oot_shadow_step.h`, `host/oot_shadow_step.c`, `nx2/oot_patches.py` `patch_shadow_step`) |
|
||||
| `HOS_SHADOW_BUCKET=<step>` | (Triforce, 11-21, the performance audit's TEX-85) The game sizes each character's shadow map (square, D32, one level) from the camera's distance every frame and places a frame's render targets first fit in its target heap, so a map a few pixels larger moved every target after it, and each moved one is a new image to the renderer. Such a map wider than 384 is answered the storage size and alignment of the next multiple of the step up (384: sides 768, 1152, 1536, 1920); the texture's own size and layout do not change, only the bytes the game sets aside, by 1.8 to 5.3 MB a map. `0`: every map its own size, as before. Not set: 384, and 0 where the target heap is pinned (`HOS_RT_POOL_MB`, or a 5120-wide output's 900 MB). The log says the step and counts the stepped answers (`host/shadow_slot.h`, `host/nvn_native.c` `shadow_slot_layout`) |
|
||||
| `HOS_OOT_TEX_COMPACT_MS=<n>` | (the lab's) The game's texture data compaction (its `aglTextureDataMgr::Compaction` thread moves a texture's data in the game's heap, a cycle of about 18 frames; idle in a fresh Market, at work at every ask after a tour, where each move is a texture at a new address for the renderer to capture and check): at most one cycle started every n ms. `-1` (a test): none started. Unset or 0: as shipped. The cycles started are logged every 10 s either way ("oot: texture compaction in"). `host/oot_tex_compact.c`, `tests/test_oot_tex_compact.py` |
|
||||
| `HOS_OOT_TEX_ASK_MS=<n>` | (Triforce) The game's texture data manager asks for a compaction round (above) at most once every n ms: its step, every frame, asks whenever no round is asked for and its flag says so, and each ask let through can start a round that moves a texture to a new address. Unset (or not a number): 1000. `0`: every ask, as shipped. `off`: none. A round started always finishes; an ask held back is made again at the next step. The asks let through and held are logged every 10 s ("oot: texture compaction asks in"). `host/oot_tex_compact.h`, `nx2/oot_patches.py` `patch_tex_ask`, `tests/test_oot_tex_compact.py` |
|
||||
| `RX_VK_TEX_WARM=1` | Textures made and uploaded ahead, while the frames are light (`gpu/vk/tex_warm.h`; off unless set). A texture is made and uploaded where a draw first samples it, and the frame's budget for that (12 ms, 64 MB, the picture held) applies only to targets drawn in the frame before, so an area's first frame, whose targets are new, takes every texture of the area at once: 350 to 800 uploads of 250 to 390 MB, 100 to 180 ms, in the lab's tours. With the switch every registered slot is noted, and at the start of a game's submit that follows a light frame (under 400 draws and dispatches: a loading screen, a menu) the noted textures, 8 frames old at least, are made and uploaded, 6 ms a frame. Bytes the game writes afterwards are uploaded again by the write marks, as before. "vk: textures made ahead in light frames" counts them. `tests/test_tex_warm.py` |
|
||||
| `RX_VK_TEX_WARM_LIST=1` or `2` | (Triforce, 11-22, the performance audit's TEX-23, after Source A's; off unless set, and only with `RX_VK_TEX_WARM=1`) The textures made ahead by a list of what was drawn (`gpu/vk/tex_warm.h`). Every registered slot made ahead needs the 40% rule, and on an 8 GB card that rule stops the warming at once. With the switch, a BC or ASTC texture a draw uses within 90 s of the game making it is learnt: its shape and eight 64-byte samples of its bytes, at its storage's address. The list is kept beside the game as `triforce_textures_drawn.bin` (32,768 at most; saved every 300 presents when it changed; a file cut short or of another version is not read). At a light frame the ripe slots the list names are made while the textures and 256 MB more fit their budget and the render targets' pool is not being given back; `1`: the others after them, by the 40% rule, once no noted slot is left ripe; `2`: the listed alone. The look itself learns nothing. "vk: textures made ahead by the list" counts listed, unlisted, learnt and no-room. Unset or `0`: every registered slot as above. Not measured: lab proof first (the entry's own default is `1`). `tests/test_tex_warm_list.py` |
|
||||
| `HOS_MEM_LIMIT=1` or `<MB>` | (the lab's) The game's memory held to what it is told it has, as the console holds it. The game is told 9809 MB, but nothing was refused until the PC's own commit ran out, and it holds 11.7 GB by its own count in a fresh Market and 13.5 GB after a tour (the `mem: census` line), with its GPU heap within 15 MB of full and its texture compaction running every frame. `1`: `AllocateMemoryPages` and `MapPhysicalMemory` are refused past the 9809 MB, and `UsedMemorySize` is the game's count; a number above 1: refused past that many MB. Unset or 0: as before. What the game does when refused is what the lab is to find out. `host/hos_mem.c`, `tests/test_mem_limit.py` Refused, the game asks again thousands of times, each ask a region taken and freed; a region of the game's own freed with none of its pages in it is no longer discarded (`hos_region_discard_needed`), and the regions it frees with pages in them are discarded with the thread's virtual core given up. Those two were the stutter when walking with the limit on: in a three-area lab run the game's own part of the frame was over 8 ms in 24% of the frames before and in 0.3% after (as without the limit), 63.8 fps to 100.0. **`=1` does not fit this build:** the heap takes 6.9 GB of the 9.8 and an area's resources need 4.7 GB and more, so a save's load can starve and stay on its loading screen for good (the owner's, 2026-11-20, read with the hang watch's stacks: the game at 9,807 MB, its load threads idle). Use a size in MB: `12800` loaded and played four areas without reaching it. |
|
||||
| `RX_VK_DESC_SET_KEEP=0`, `RX_VK_DESC_TEMPLATE=0` | A program over the driver's push limit (32 descriptors on NVIDIA's, AMD's and Intel's drivers) has a set written for each draw that builds its descriptors. By default a set whose descriptors are all as those of the last one written for that layout in the recording is bound again instead, while no texture was destroyed or made again and no pool moved, and a set written is written through the program's descriptor update template (Vulkan 1.1). (Triforce, after NX2navi's `desc_template`) A program within the push limit has its descriptors pushed whole (in a new command buffer, or after another program at its bind point) through a push descriptor template of its own (`vkCmdPushDescriptorSetWithTemplateKHR`, one a program and bind point, from the same entries and packed array as a set's), not with the write structs the driver walks one by one; a push of the changed bindings only keeps its write structs. The frame statistics' `a frame's descriptors` line counts the pushes made through it. With `RX_VK_RECORDER` the packet copies the packed array (`gpu/vk/vk_recorder.h`). `RX_VK_DESC_SET_KEEP=0`: a new set each time; `RX_VK_DESC_TEMPLATE=0`: sets written and pushes made with the write structs; both as before |
|
||||
| `RX_VK_CREATE_NEW_MS=<ms>` | While the hold is on, a sampled texture seen for the first time once the frame's makes took this long (8), or its uploads their budget, is not made yet (`vk_texture.cpp` `create_later`): the draw samples the dummy, the frame is held, the next frame makes it. The frame's first make always goes; never a dispatch's. `0`: made at once, as before (a texture left for later was made all the same). The stall line counts the textures made, their time, those not made yet, and the images made |
|
||||
| `RX_VK_UNSWIZZLE_MT_MB=<MB>` | A block-linear level unswizzled on the CPU is split over several threads (a thread for each 2 MB, 8 at most) from this size (4); `16`: from 16 MB, as before |
|
||||
| `RX_VK_PROGRAM_READAHEAD=<n>` | After a draw reads a program from the program cache itself, or takes one read ahead, workers read the next `n` programs (32) and keep them made for the draws that come to them, 512 at most (`gpu/vk/program_readahead.h`). `0`: none, as before. The log's `vk: programs from the cache` lines and the stall line count them |
|
||||
| `RX_VK_PROGRAM_READ_UNLOCKED=0` | A draw that reads a program from the program cache itself, or waits for one the read-ahead's worker is reading, lets the shaders' lock go meanwhile, so that the replay's helper threads go on picking their own programs (they queued behind it: "the program picks waited 169 ms for the shaders' lock" at the Market's arrival); a second draw of that program waits for the first's result, and the record stays listed until it is read (`gpu/vk/program_read_once.h`). A program's release waits until no pick has the lock let go. `0`: read and waited for under the lock, as before. The stall line counts the waits for another draw's read, and the picks' waits to take the lock again with their waits for it |
|
||||
| `RX_VK_STAGE_CODE_CACHE=<MB>` | What a program pick made of a stage's bytes (the control section and the code area read from guest memory: its code key, the instructions lowered to Maxwell) is kept by those bytes, 64 MB at most, the least recently used going first: a program object met for the first time whose stages' bytes were met before (the game makes them again at loads) takes a copy instead of decoding them again under the shaders' lock (`gpu/vk/stage_code_cache.h`; an entry is taken only for the same bytes, byte for byte). The save's load-in frame said "the program picks read the stages' code 45.6 ms": a stage was some 60 us to read and decode, and is 3 us to find. `0`: none, each stage decoded again for each program object, as before. The log's `vk: the stages' code read for the program picks` lines count them |
|
||||
| `RX_VK_STAGE_CODE_LAZY=0` | A program pick reads the stages of a program object it meets for the first time without lowering their code: their keys, NVN stages and headers (what finds the program) come from the control sections and code areas, and the code is lowered (and kept, `RX_VK_STAGE_CODE_CACHE`) only when the pick translates a variant. A program found in memory, read ahead or in the program cache was made from stages that lower (its key names them); one not found has its stages lowered before it is translated, and should one not lower, the pick starts again from those that do: the program it was before (`gpu/vk/stage_code_cache.h`). The save's load-in frame lowered every stage of each program object met, 48 of a stage's 60 us, for programs all prepared before the title. Not with `RX_VK_REJECT_STAGES`, `RX_VK_REJECT_KEYS`, `RX_VK_OLD_LOWERING`, `RX_VK_DUMP_NATIVE` or `RX_VK_DUMP_FAILED`, which act in the lowering. `0`: each stage lowered as it is read, as before. The log's `vk: the stages' code read for the program picks` lines count the stages read without lowering and those lowered since |
|
||||
| `RX_VK_STAGE_CODE_UNLOCKED=0` | A program pick reads the stages of a program object it meets for the first time (guest memory, their keys, the stage code cache) with the shaders' lock let go, as a draw's own read of a program (`gpu/vk/program_read_once.h`: a program's release waits until no pick has the lock let go), so that the replay's helper threads go on picking their own programs meanwhile; two picks of one object at once each read it, and the first list put stands. `0`: read under the lock, as before. The stall line's "the program picks read the stages' code" counts the read and the lock's taking again |
|
||||
| `RX_VK_PROGRAM_ORDER=0` | Each session writes the order its draws first needed their programs in (read from the program cache, or translated), the first 4096, into `triforce_program_order.bin` beside the cache now and then (this session's first, then those of the sessions before it has not met). The next session's read-ahead workers, from the first draw on (the title and the file select), read those programs in that order whenever no draw asks them for anything, the whole order, and keep them made apart from the shelf until a draw takes them (`RX_VK_PROGRAM_ORDER_KEEP` below; `gpu/vk/program_order.h`): the save's load-in frame read some 100 programs on the draws' threads before. Not with the RAM low. `0`: neither read nor written, as before. The stall line says why the draws read what they still read themselves (named by no order, read ahead and let go before, not read ahead yet) and how many they took in this order; the log says once when the whole order has been gone through, and how many programs it keeps |
|
||||
| `RX_VK_PROGRAM_ORDER_KEEP=0` | A program read ahead in the last sessions' order stays until a draw takes it, whatever the draws meet meanwhile, and nothing in the order is passed by: the workers read all of it (4096 at most) while the title and the file select stand still, and a program a draw's ask read that the order names is kept with these, not on the shelf, which lets its oldest go. Only a RAM gone short lets them go (all of them). The pipelines' warm keeps its window: the first 512 kept and not taken are read with it, those beyond without, each warmed as the draws take warmed ones (the nearest first). Some 40 to 50 KB a program packed: the owner's order of 2411 programs some 100 to 120 MB, 4096 at most some 200 MB. The owner's test37 session (SDR, where the order's sessions were HDR Native) took another path through the order and its load-in frame read 115 programs itself, 64 of them read ahead and let go as behind where the draws were, 51 passed by. `0`: the rules before: 512 at most, each until the draws are 256 places past it (where the draws are: the middle place of the last 16 programs they needed first), what lies behind not read |
|
||||
| `RX_VK_PROGRAM_READAHEAD_NOTES=0` | The read-ahead's next programs are those after it in the pipeline notes' order (the order earlier sessions first built their pipelines: an area's programs lie together), 4 workers, the compute programs too (in the order play first dispatched them, `triforce_computes.bin`), and a full shelf lets its oldest go; the program cache file's order only for a program neither names. `0`: the file's order (since the shaders are prepared before the title, the preparation's), one worker, no compute programs, and nothing more read once 512 wait, as before (test16's file select load took 7 of some 600 programs read ahead that way) |
|
||||
| `RX_VK_PIPELINE_WARM=0` or `1` | Where the precompile keeps no pipelines (the RAM spared) and the fast pipeline link is on, the read-ahead's worker that has just read a program makes, before any draw can have it, the stage libraries of the pipelines the pipeline notes name for it (asked of the caches only: one the driver would have to compile is left to its draw) and links them; the draw takes the link made in place of linking its own, and the whole build follows as before (`gpu/vk/pipe_warm.h`). At most 32 notes a program variant, 512 links held for draws that have not come (the oldest let go beyond) and 1024 libraries no draw has had; nothing made while the RAM is low or short, and what is held goes back to the driver while it is short. `0`: none, as before; `1`: the libraries only. The log's `vk: pipelines made behind the draws` lines count them |
|
||||
| `RX_VK_PIPELINE_WARM_FOLLOW=1` | Off by default (test19: the lab's tour showed no gain). `1`: the warm follows the draws: a program read within 64 places (in the notes' order) after one the draws consumed in the last 2 s has its libraries made and linked, one within 192 its libraries only, one farther (or once the draws stopped taking) nothing; and the libraries no draw has had are kept up to 2048, the oldest let go beyond (with the links made from it), so the bound never stops the warm (`gpu/vk/pipe_warm.h`, test18: 2266 links made, 276 taken, the libraries at their bound from the load on); it spends at most 250 ms of the workers' time a second (500 saved up), and never makes again a library it let go. Off: every program read is warmed whole and the libraries stop at 1024, as before |
|
||||
| `RX_VK_COMPUTE_WARM=1` | Off by default (test19: the lab's tour showed no gain). `1`: where the precompile keeps no pipelines, the read-ahead's worker that has just read a compute program play dispatched (`triforce_computes.bin`) makes its compute pipeline, asked of the caches only, as its first dispatch would; the dispatch that takes the program has it. One a program while it waits on the read-ahead's shelf, destroyed when the shelf lets the program go and given back while the RAM is short; once a program variant a session. Off: each compute pipeline is made at its first dispatch, as before |
|
||||
| `RX_VK_PIPELINE_ORDER=0` | Each session writes the order its draws first needed their pipelines in (a new one, or one the precompile kept that no draw had), the first 4096, into `triforce_pipeline_order.bin` beside the pipeline notes now and then, from the presents on a thread of its own (this session's first, then those of the sessions before it has not met). Where the pipeline warm runs (`RX_VK_PIPELINE_WARM`), the next session finds each one's note before play and the warm makes, for a program variant the read-ahead has read, the pipelines that order names for it and no other (32 at most a variant; the notes' other pipelines are not made then, and the warm does not follow the draws): their stage libraries asked of the caches and their links, held for the draws by their place in the order, 1024 at most (beyond, the farthest ahead goes, and none is made farther), and 2048 libraries no draw has had at most. Where the draws are: the middle place of the last 16 pipelines they needed first; a link held 512 places behind it goes with the libraries only it held, and none is made there (`gpu/vk/pipe_order.h`): the save's load-in frame made 86 to 92 ms of pipelines on the draws' threads, the warm's links from the notes let go before the load came (lab round 37: 1964 made, 54 taken). The first session after this writes the file; the next one has it. `0`: neither read nor written, the warm from the notes as before. The stall line counts the pipelines new to the draws, those linked ahead (of them in this order), and the links and stage libraries the draws made themselves; the log's `vk: pipelines made ahead in the order` lines the rest |
|
||||
| `RX_VK_PIPE_AHEAD=<places>` | (Triforce, 11-22, the performance audit's PIPE-81, after Source B's: its first step; `gpu/vk/pipe_ahead.h`) A count, nothing built: with the last sessions' pipeline order in use (light mode), each pipeline the draws first need is counted by its place against where the draws are: within this many places ahead (1024 unset, at most 4096), further ahead, behind, not in the order, and those within the window that the warm had linked. A line every 600 presents while it grows (`vk: pipelines first drawn, by their place in the order: ...`) says whether a builder of whole pipelines just ahead of the draws would meet a third of them or more, the entry's rule for building it at all. `0`: not counted (`gpu/vk/vk_draw.cpp` `order_met`) |
|
||||
| `RX_VK_VRAM_MB=<MB>` | The texture cache's budget as a fixed size, whatever the card. Unset: on a card the driver allows more than 8.5 GB, no size (`RX_VK_TEX_PRESSURE`); on a smaller one 45% of what the driver allows; with `RX_VK_TEX_PRESSURE=0` 70% of it above 8.5 GB, at most 6 GB (the rule before) |
|
||||
| `RX_VK_TEX_PRESSURE=0` or `=<pct>` | (Triforce, after NX2navi's `TEX_VRAM_PRESSURE`) On a card the driver allows more than 8.5 GB, with VK_EXT_memory_budget, not an integrated GPU, without `RX_VK_VRAM_MB` and without a quality preset whose budget applies, the texture budget follows the live use instead of the fixed 6 GB: textures go once the driver's count, less the free space in the blocks they can take back, is past 80% of what the driver allows now, the least recently used first, each unused 2 s or more, 64 a pass, down to 2% under that line (`gpu/vk/tex_pressure.h`); the 90% and 97% rules and the compaction's lines stay. While the system's commit is low the fixed size comes back (`RX_VK_TEX_COMMIT_FREE_MB`). `0`: the fixed size, as before; `<pct>`: that line (50 to 95). The log's `vk: texture budget` line at the start says which, and the `vk: VRAM:` line every 600 presents the budget (by live use: what the textures may hold before the line) and the line |
|
||||
| `RX_VK_TEX_KEEP_TAKEN=0` | (Triforce, after NX2navi's `TEX_KEEP_TAKEN`) While VRAM is short, a texture uploaded from memory that the game let go 30 presents ago goes first (`RX_VK_LET_GO_PCT`); one the game has taken back before waits 8 times the longest it stayed let go before a take-back, when that is longer (`gpu/vk/tex_pressure.h`): made again, the frame that wants it draws the dummy (NX2navi: Kakariko's haze gone every few seconds). The eviction by age still takes it. `0`: the 30 presents alone, as before |
|
||||
| `RX_VK_TEX_COME_BACK=0` | (Triforce, after NX2navi's pause for drops that come back) Where the budget follows the live use: when in 240 presents 64 or more textures the line let go went and half or more of them were made again within 600, the line stands aside for 600 presents, doubled at each such period up to 7200; a period with fewer back ends the doubling (`gpu/vk/tex_pressure.h` `ComeBack`). The 90% and 97% rules and the compaction's lines still hold meanwhile. The log says when a pause begins. `0`: no pause |
|
||||
| `RX_VK_TEX_STEP=<n>` | The live textures one pass of `texture_sweep` over the budget lets go of at most (64 where the budget follows the live use; the render targets drawn over since do not count), the rest at the next look two presents on, not ten. Set, on any card, the fixed size's too; `0`: no step, one cut as before |
|
||||
| `RX_VK_TEX_COMMIT_FREE_MB=<MB>` | Where the budget follows the live use: while less of the system's commit than that is free (the default: the larger of 4 GB and a tenth of the limit, NX2navi's), the fixed size is the budget again (Windows charges a discrete card's memory to the commit, and an allocation that fails there ends the game). `0`: never. Not on Linux |
|
||||
| `RX_VK_TEX_COMMIT_RESERVE_MB=<MB>` | (Triforce, 11-22, the performance audit's OS-05, after Source A's; off unless set, not measured here) On any card, while less of the system's commit than that is free (the page files' room to grow counted in, below), the textures are held to what they hold less the shortfall, never under 1 GB, as the 97% rule takes its excess: the cache's passes give blocks back before an allocation is refused at Windows' commit limit (`HOS_COMMIT_WAIT_MS` waits such a refusal out). Source A's is 2560, on by default there; one `vk: commit limit near` line when it begins and ends and every 600 presents while it holds. Not set or `0`: no cut, as before. Not on Linux (`gpu/vk/commit_room.h`, `gpu/vk/vk_texture.cpp` `texture_frame`) |
|
||||
| `RX_VK_TEX_COMMIT_GROWTH=0` or `=1` | (Triforce, 11-22, OS-05) Whether the room the page files still have to grow (each one's largest size less its size now, a drive's page files capped together by its free space, asked of Windows at most once a second) counts as free commit, for the reserve above and for `RX_VK_TEX_COMMIT_FREE_MB`'s fall-back alike: Source A's first rule, without it, retired textures for 90 minutes while a page file Windows manages could still grow. Read only while less is free than the higher of the two lines. Not set: counted where `RX_VK_TEX_COMMIT_RESERVE_MB` is set, else not (as before); `0`: never; `1`: also without the reserve (`vk_texture.cpp` `commit_now`, `page_file_growth`) |
|
||||
| `RX_VK_SWEEP_ALL=1` | Every texture of the cache ordered by its last use for every pass of `vk_texture.cpp` `texture_sweep`, as before. By default a pass under the budget (every 10 presents, and the dead render targets' passes between them) orders only the textures it can take: those the GPU drew and those the game let go; the others it passes over all the same. Over the budget every texture is ordered. The log's "passes for the dead render targets between the looks" line says what a pass costs |
|
||||
| `RX_VK_TEX_RATE_HIGH=0` | The texture eviction's waits (`vk_texture.cpp` `texture_sweep`: 60 presents unused over the budget, 30 after the game let a texture go, 600 or 3600 for one the GPU drew, the compaction's 600 and 1800) count presents above 60 a second, as before. By default they count their seconds at any rate up to 480 a second (`gpu/vk/present_rate.h`, after NX2navi's M19): at the Market's 94 a second "60 presents" were 0.64 s. Below 60 they count seconds unless `RX_VK_TEX_RATE_LOW=0`. The log says the first time a run goes above 68 a second |
|
||||
| `RX_VK_TEX_DETAIL=full`, `reduced` or `low` | The texture detail for one run, over the Insert menu's Texture detail row: whole textures, or the large ones without their largest level, or without two (`gpu/vk/tex_detail.h`). Unset, the row decides, and its Automatic goes by what the driver allows: full from 10,240 MB, low under 6,144 MB, never full on an integrated GPU |
|
||||
| `RX_VK_TEX_BC1_PARTS=0` | (Triforce, 11-21, the performance audit's TEX-50) An ASTC block of two to four partitions with a colour endpoint mode each is refused unread by the opaque-texture scan, as before, so a texture with one stays BC3. The default reads those modes as the decoder does (their bits below the weights) and takes the block as opaque when every partition's mode has no alpha and is not HDR; the blocks the decoder takes for an error are still refused. The `kept as BC1` log lines count the blocks read this way. And a first start, with no options file, now keeps opaque textures as BC1 as the row's default says: it started at off (`gpu/vk/astc_opaque.h`, `vk_menu.cpp` `tex_bc1_at_start`) |
|
||||
| `RX_VK_TEX_BC1=0/1` | Opaque ASTC textures kept as BC1 instead of BC3 for one run, or not, over the Debug tab's Opaque textures as BC1 row (on by default since options_version 5): a texture the renderer keeps as BC3 whose every block the GPU decoder writes opaque is encoded as BC1, half the video memory, the same texels (`gpu/vk/tex_bc1.h`, `gpu/vk/astc_opaque.h`). One whose memory later holds blocks with alpha, or that a copy on the GPU writes, becomes BC3 for good; the log counts both |
|
||||
| `RX_VK_ASTC_BC_ALL=1` or `0` | Large decoded ASTC textures are kept as BC3, a quarter of RGBA8: the coarse blocks (6 x 6 texels and up) from 2 MB as RGBA8 (`RX_VK_ASTC_BC_MB`), the finer ones (UI, characters) only from 64 MB. Where the driver allows under 8.5 GB the finer ones go from 2 MB too (`gpu/vk/astc_bc_policy.h`). (Triforce, 11-22, the performance audit's TEX-12, after Source A's) And on every integrated GPU, whatever it is allowed: its allowance is a share of the system's RAM (a 16 GB handheld reported 10,361 MB), so it kept those textures as RGBA8 in RAM the game needs too. The log says `all large ASTC textures kept as BC3`, with `an integrated GPU` there. `1`: on any card; `0`: the finer ones as RGBA8 on any card, which is an integrated GPU allowed 8.5 GB or more as it was before |
|
||||
| `RX_VK_BC_SMALL_LAYERS=<n>` | (Triforce, 11-21, the performance audit's TEX-09) The layers one run of the ASTC-to-BC3 re-encode takes for a level of 1024 x 1024 or less, 1 to 8, through a small scratch of its own (1024 x 1024 x 8, 32 MB, and 4 MB more of output buffer). Unset: 8, and 1 on AMD's cards until the picture is checked there (one layer a run was for dark lines between Hyrule Field's ground tiles on an RX 6600). A terrain array of 90 layers and 11 levels was about 990 runs of a decode, an encode, a copy and three full barriers; its levels from the second down are now about 210. Each layer of a run is copied by a region of its own. No small scratch at reduced or low texture detail. `1`: none, every run one layer, as before (`gpu/vk/bc_scratch.h`, `vk_texture.cpp` `bc_small_layers`, `bc_encode_into`) |
|
||||
| `RX_VK_BC_LAYERS=<n>`, `RX_VK_BC_SCRATCH_FULL=1` | The layers one run of the GPU's ASTC-to-BC3 re-encode decodes and encodes, 1 to 8 (1: with more, an array's layers past a run's first came out wrong in the smaller levels, dark lines between Hyrule Field's ground tiles). (Triforce, 11-21, the performance audit's TEX-10) The 2048 x 2048 scratch, its output buffer and the second scratch are made with that many layers, 20 MB a layer (36 with the second), not with 8 whatever a run used: 140 of 160 MB were held and never written. `RX_VK_BC_SCRATCH_FULL=1`: 8 layers, as before (`gpu/vk/bc_scratch.h`, `vk_texture.cpp` `bc_scratch_layers`) |
|
||||
| `RX_VK_TARGET_POOL_MB=<MB>`, `RX_VK_POOL_IDLE=<presents>`, `RX_VK_TARGET_POOL_TRIM=0` | (Triforce, 11-21, the performance audit's TEX-29) The render targets' pool gives a dead target's image to the next target of its shape. Its bound is a twelfth of what the driver allows, 256 to 768 MB (`<MB>` sets it; 0: no pool); it was a fixed 256, which starved at 2560 x 1440. It is trimmed oldest first, whatever the shape. A shape nothing has taken from the pool goes after 8 presents (`RX_VK_POOL_IDLE`), one that was taken before after 120. Above 90% of the allowance it holds half its bound, above 97% nothing, unless a floor of the targets' blocks is kept. `RX_VK_TARGET_POOL_TRIM=0`: emptied above 90%, as before; with `RX_VK_TARGET_POOL_MB=256 RX_VK_POOL_IDLE=120` the pool as it was. The 600-present line `vk: render target images` has the reuse, the bound and the tier (`gpu/vk/target_pool.h`, `vk_texture.cpp` `pool_trim`) |
|
||||
| `RX_VK_STAGING_BURST_S=<seconds>` | (Triforce, 11-22, the performance audit's TEX-28, after Source A's) The uploads stage their bytes in host-visible arenas of 64 MB, and each present gives the idle ones back beyond what is kept: 256 MB, 64 MB where the driver allows 6.5 GB or less (`RX_VK_STAGING_KEEP_MB=<MB>`), and any nothing staged into for 30 s (`RX_VK_STAGING_IDLE_S=<n>`, 0: kept). An arena is taken again only once the GPU is done with the submission that read it, so frames that stage in a row use two in turn, and with 64 MB kept the one that had its turn was freed at every second present of a load and made again the present after. An idle arena staged into in the last 2 seconds now stays while the arenas add up to 256 MB (`gpu/vk/staging_keep.h`), unless VRAM is over 90% of the allowance. Nothing changes where 256 MB are kept anyway. The log's `vk: memory:` line counts the arenas made so far and their time. `0`: as before |
|
||||
| `RX_VK_WROTE_REGIONS=0`, `=2` | (Triforce, 11-21, the performance audit's TEX-02) Every sampled, uploaded texture is asked once an epoch whether the GPU wrote into its storage since its upload, which took a lock and walked the map of written ranges; nearly every answer is no. One bit a 2 MB region of GPU addresses, set where a GPU write ever reached, answers no without the lock when the storage touches no set region; a set bit asks the map as before. `0`: the lock and the walk each time; `2`: a check, both asked and compared, the count of disagreements (which must be 0) logged (`gpu/vk/wrote_regions.h`, `vk_texture.cpp` `texture_gpu_wrote`, `gpu_wrote_into`) |
|
||||
| `RX_VK_PREDICT_BG_STREAM=0` | (Triforce, 11-21, the performance audit's PIPE-83) The builder of the predicted pipelines behind the game (`predict_background`: the starts after a translator or driver update, resting while the game is in play) keeps its plan as record numbers and reads a pipeline's two notes from `triforce_pipelines.bin` when it builds it, each worker through its own handle; it held every note (~27 KB) and a full note of every predicted pipeline until it ended, about 1 GB. The same programs and pipelines are built. `0`: all of them held, as before (`gpu/vk/predict_stream.h`, `vk_draw.cpp` `predict_background`) |
|
||||
| `RX_VK_PREDICT_CHECK=0`, `=2` | (Triforce, 11-22, the performance audit's PIPE-68, after Source B's and Source A's; `gpu/vk/predict_check.h`) On unless `0`. The predicted pipelines are built once for the driver's own cache and a record says which programs are done; a driver can lose them all the same (NVIDIA's cache cap, a cleared cache folder, AMD's Windows driver keeping a third). At a start whose record says all are built, the builder's idle thread, after its delay and only while the frame shown is a light one, samples up to 12 of them spread over the plan: each built for the driver's cache alone and timed, then once more as the warm reference; cold is a first build over 4 times the second and 3 ms more, and the sample ends once its majority is decided. Cold: the record is removed, so they are all built again behind the game at light frames, and `<notes>.predicted.check` takes a strike for these shaders and this driver; cold again after that, the log says this driver does not keep them and no later start checks. Warm: the strike is forgotten. `2`: checked at every start, a recorded stop too. Also counted: the builds in play of programs the plan predicts that took over 20 ms (`vk: predicted pipelines: N builds in play of predicted programs took over 20 ms`, at most every 10 s while it grows). `0`: neither the check nor the counter, as before (`gpu/vk/vk_draw.cpp` `predict_check_cold`, `predict_background`, `predicted_play_build`) |
|
||||
| `RX_VK_PREDICT_TYPES=1` or `0` | (Triforce, 11-22, the performance audit's PIPE-44, after Source A's; `gpu/vk/predict_types.h`) Unset: each predicted pipeline is built once for its program, by its key as play keys it (two predicted layouts that differ only in what is dynamic, a stride, gave the same pipeline twice): the same pipelines, fewer builds. `1`: that, and the plan drops a predicted vertex layout whose attributes definitely differ from the program's vertex inputs (an integer input against a float attribute, or the other signedness, read from the vertex module's SPIR-V; scaled formats count as integers): such a pipeline is invalid. Off by default because it changes which pipelines are built, and it applies to a plan made after it is set (a plan is made once for the shaders). `0`: every predicted pipeline built, as before. The predicted lines say how many were left out as built already, how many programs the programs' file did not lend, the notes of variants it has no shaders for and the layouts that did not fit (`gpu/vk/vk_predict.cpp` `predict_pipelines`, `gpu/vk/vk_draw.cpp` `predict_first`) |
|
||||
| `RX_VK_DRAWN_PAGE_MAP=0` | (Triforce, 11-21, the performance audit's TEX-21) A texture in a virtual pool is checked through its mapping, 64 KB a page, and its page map was recorded at its first check: one the GPU drew first had none, so once the game finalized it the first check took its whole storage as mapped since and uploaded it from memory over what the GPU drew (HUD labels and render targets blank after an area change, up to 15 MB an upload). The map is recorded at the first draw that says who drew it, and the later check gives only the pages mapped anew or written by the CPU. Not for a write that gives no reason (a clear, the movie upscaler). The log counts them. `0`: at the first check, as before (`gpu/vk/drawn_page_map.h`, `vk_texture.cpp` `texture_written`) |
|
||||
| `RX_VK_DEAD_SUM=0`, `RX_VK_DEAD_POOL_NOW=0`, `RX_VK_DEAD_MIN_UNUSED=3` | (Triforce, 11-21, the performance audit's TEX-30) The frame's transient render targets, drawn over from another start: the other targets' parts over one are added up, each byte once, so one covered at its head by one target and at its tail by another, each under an eighth, is let go too (it was kept for good); it is dead after 1 unused present, not 3; and its image goes to the render targets' pool in the pass that lets it go, once the GPU is done with it, not at the next present. `RX_VK_DEAD_SUM=0`: judged one by one; `RX_VK_DEAD_POOL_NOW=0`: pooled at the next present; `RX_VK_DEAD_MIN_UNUSED=3`: three presents: each as before (`gpu/vk/target_overlap.h` `ShiftedOver`, `vk_texture.cpp` `texture_sweep`) |
|
||||
| `RX_VK_ASTC_REUSE=2` | (Triforce, 11-21, the performance audit's TEX-14, its first step; off when not set) When the game comes back to an area it makes its textures again, mostly at other addresses, and the same ASTC blocks are decoded and encoded again while the image made of them the first time may still be in VRAM. With `2` a new ASTC texture's blocks are hashed at its first upload (128 bits), the textures the game finalizes that were uploaded from memory and never drawn into are listed, and each new texture with a listed one's blocks and shape is counted: the log's `decoded ASTC used again` line says how many MB of images a copy would have stood in for, and what the hashing cost. Every texture is decoded as before. The copy itself is not built: the walk it was measured on would have reused 0 to 11 MB, and the audit asks for 500 MB on a tour before it is (`gpu/vk/astc_reuse.h`, `vk_texture.cpp` `reuse_upload`) |
|
||||
| `RX_VK_KEEP_OBSOLETE=1` | (Triforce, 11-21, the performance audit's TEX-43) A texture the GPU drew in one frame is kept after the game finalizes it, for being made again over that storage and sampled undrawn (the terrain's tiles; `RX_VK_KEEP_FINALIZED_MB`). One of 32 KB or more, in no virtual pool, whose storage the CPU has written three quarters of since the release holds nothing that memory holds any more: the sweep lets it go whatever the budget says, where only a draw over it did, and the game's own writes kept each for good (about 200 render target images an hour in players' logs). The log counts them. `1`: kept, as before (`gpu/vk/finalized_obsolete.h`, `vk_texture.cpp` `texture_sweep`) |
|
||||
| `RX_VK_SLOT_RETIRE=<presents>` | (Triforce, 11-21, the performance audit's TEX-76) The game draws each character's shadow map at a fixed address, sized from the camera every frame, and each size is a texture of its own: the ones before it, drawn over, waited 120 presents unused before the sweep took them. A depth target of one level and layer whose start a newer target of the same format and another size was drawn over since its last use goes once unused for `<presents>` presents, 16 when not set; nothing is written back, and its image goes to the render targets' pool. The log counts them. `0`: after the sweep's own wait, as before (`gpu/vk/slot_retire.h`, `vk_texture.cpp` `texture_sweep`) |
|
||||
| `RX_VK_COUNT_BY_TEXELS=1` | (Triforce, 11-22, the performance audit's TEX-86, after Source A's) A texture the GPU samples in the game's own compressed blocks (BC, and ASTC where the log says `vk: ASTC sampled natively`) is counted by its whole blocks (`gpu/vk/tex_count.h`): a level is its width and height rounded up to blocks, times the block's bytes and the layers. It was counted by texels: a byte a texel for BC (twice BC1's and BC4's size) and the 16 bytes of an ASTC block for every texel (16, 64 and 144 times the size at 4 x 4, 8 x 8 and 12 x 12), so the texture budget let textures go for memory nothing held, and the first size estimate left out the top levels of arrays that fit. The count is still lifted to what the allocator gave. Decoded ASTC (NVIDIA's and AMD's desktop drivers) and the renderer's own BC3 and BC1 are counted as before. The log says once which count is on. `1`: by texels, as before |
|
||||
| `RX_VK_EVICT_STALE_FLUSH=1` | (Triforce, 11-22, the performance audit's TEX-R404, after Source A's) Over the budget, the eviction writes each chosen texture the GPU drew back into the game's memory before it goes. The game draws its HUD labels inside the storage of a 1920 x 1080 target it no longer uses; when that target went, its old texels landed over the labels, which were then uploaded from them as noise. A write-back is now left out when another image the GPU drew overlaps the texture's storage and was drawn in a later frame than the texture's own last drawing (`gpu/vk/evict_reason.h`; one drawn in the same frame keeps it); the texture goes all the same. The log names the first 20 (`vk: texture #... goes without its write-back`) and counts them on the `written back into their memory before eviction` line. Nothing changes while no pass writes anything back (a card under its budget). `1`: every chosen texture is written back, as before |
|
||||
| `RX_VK_COMPACT=1`, `RX_VK_COMPACT_PCT=<50..90>`, `RX_VK_COMPACT_MB=<MB>` | VRAM compaction (off by default): the textures that are not render targets in a pool of their own, moved out of sparsely used blocks a few a present so that the blocks go back to the driver, and the texture budget holding the driver's count at that share of what it allows (85); the free space that starts a run (512; 192 while the count is over 90%) |
|
||||
| `RX_VK_EVICT_EXCESS=0` | (Triforce, on by default; `=0`, or the lab's knob `noexcess`, as before) With the compaction on, what the 90% and 97% rules ask of the hold's pass is their excess (`gpu/vk/vram_compact.h` `rule_target`, `gpu/vk/vk_texture.cpp` `texture_frame`): their budget less 64 MB, where the pass went down to 7/8 of it. Those budgets are the textures less the excess, not sizes: 62 MB over the 90% line on an 8 GB card took about 400 MB of textures, all on screen again within the camera's turn, made again, and over the line again. Over the budget's size (the fixed one, or the live line's) the pass still goes down to 7/8 of that size; with both, the lower |
|
||||
| `RX_VK_COMPACT=1`, `RX_VK_COMPACT_PCT=<50..90>`, `RX_VK_COMPACT_MB=<MB>` | VRAM compaction (off by default): the textures that are not render targets in a pool of their own, moved out of sparsely used blocks a few a present so that the blocks go back to the driver, and the texture budget holding the driver's count at that share of what it allows (85); the free space that starts a run (192 while the count is over 90%). `RX_VK_COMPACT_MB` unset: (Triforce, 11-22, the performance audit's TEX-38, after Source A's) a sixteenth of what the driver allows, 64 to 512 MB (`vram_compact.h` `default_min_free`): 512 on a card that allows 8 GB or more, as before; 327 on a 6 GB laptop that allows 5,233 MB; 512 before the first look. `RX_VK_COMPACT_MB=512`: 512 on every card, as before. `tests/test_compact_start.py` |
|
||||
| `RX_VK_MEMORY_PRIORITY=1` or `<targets>,<textures>` | (Triforce, 11-22, the performance audit's TEX-47, after Source A's; off when not set) When a card is short of video memory, Windows moves some of the game's device memory to system memory, read over the bus from then on, and picks what goes. With the switch and a driver that has `VK_EXT_memory_priority`, the game ranks its memory: the render targets' blocks 1.0 (kept first), the sampled textures' 0.25 (moved out first), everything else the driver's own 0.5 (`gpu/vk/memory_priority.h`); `<targets>,<textures>` sets the first two, each from 0 to 1. The textures then come from their own pool also where it is not compacted (`RX_VK_COMPACT`), and a sparse texture's chunks name the textures' rank. The log says `vk: memory priorities on` with the three numbers, or that the driver has no such extension. Source A measured it on an RTX 4090 with the video memory taken by other programs: with 1.3 GB moved out, a median GPU frame of 24 to 29 ms against 109 to 179; with room on the card, no difference. Not run on a small card, and a driver may ignore it: off by default, and without the switch the extension is not asked for. `0`, `off`, `no` or a value that cannot be read: off |
|
||||
| `RX_VK_SKIP_COMPRESSED_OVER_RT=1` | A compressed (ASTC, BC, ETC2/EAC) texture made over storage render targets used is not uploaded, as before: bosses and new enemies all blue the first time |
|
||||
| `RX_VK_NO_WRITEBACK` (set) | GPU writes to mirrored pools are not copied back to guest memory (an experiment, for A/B) |
|
||||
| `RX_VK_MSYNC_PASS=0`, `=1` | (Triforce, 11-21, the performance audit's MEM-48) A mirror sync's pages that come from a capture's buffer go to the pool's VRAM mirror by the write-back diff's compute pass, one dispatch a sync, where they were copy regions of a page or a few each (700 to 1,300 a frame, each read over PCIe after the one before). A run not in whole words, a GPU without subgroup votes in compute and a sync whose chunk table finds no room keep the copies; a mirror written twice in one upload command buffer, once by the pass, gets a full barrier between. On for NVIDIA when not set; `1`: on every GPU; `0`: copy regions, as before. The log's `mirror syncs' pass (600 submits)` line counts dispatches, chunks, fallbacks and barriers; `RX_VK_LAB_AB=nomsyncpass` is the in-run comparison (`gpu/vk/msync_pass.h`, `vk_memory.cpp` `msync_pass_add`, `msync_pass_flush`) |
|
||||
| `RX_VK_WB_FORGET_SORT=1` | (Triforce, 11-21, the performance audit's MEM-61) At the end of each part with write-backs the write-back diff clears, for every page written since, the bit that says its pool's shadow knows the page. The three lists of such pages (the game's writes, the collection's, the replay's) were sorted first, thousands of pages a part, only so that the pool found for one page serves the next; the bits and the counts do not depend on the order, and the lists hold ascending runs as they come. They are no longer sorted. The log's `forgetting the written pages took` line gives the time. `1`: sorted, as before (`vk_memory.cpp` `shadows_forget_written`) |
|
||||
| `RX_VK_WB_DIFF=0` | Write-backs copy every byte again instead of only what changed against a VRAM shadow (the diff is on by default; v2.0.9 and v2.0.10 had it off, issue 1161) |
|
||||
| `RX_VK_LAZY_WB=1`, `=2` or `=0` | The lazy write-back for one run, over the Debug tab's Lazy write-back row (`lazy_write-back=1`; on by default since options_version 5; `docs/lazy-writeback.md`, `gpu/vk/wb_lazy_state.h`): a whole page of what shaders write into a mirrored pool that the CPU leaves alone is no longer copied back at every submission; its page has no access, and whoever touches it (the game's code or the host's, by a fault; a host call that hands it to the kernel; the GPU through the import) has it copied down first. Windows only. On AMD's driver the row alone leaves it off, as NX2navi did after games ended in play there; `=1` turns it on anyway. `=2` is the self-check: every copy is still made, the pages are protected and paid as if not, and each payment compares what it would have copied with guest memory: a `MISMATCH at guest` line is a page it would have got wrong. The log says `vk: lazy write-back on` (or why off) at the start, and every 600 submits `vk: lazy write-back (600 submits)`: the MB not copied, the pages owed, the pages paid by a fault, a host call, a GPU read, a mirror given up and a recording's submit, the payments and their time, and what failed |
|
||||
| `RX_VK_LAZY_PAYERS=<1..4>` | (Triforce, 4 by default; `=1` as before) How many of the lazy write-back's payments copy at once (`gpu/vk/wb_lazy_payers.h`, NX2navi's `wb_lazy_unlocked`). A payment's copy goes onto the queue after everything submitted so far and its thread waits for it (0.5 to 3 ms each in the owner's sessions, up to 21 ms at a load); with one at a time, a second thread that touched another owed page, or the replay giving a mirror up, waited for the first one's copy to come back before its own went onto the queue. Each payment holds a slot of its own (a command buffer and a fence); the pages of two payments are never the same ones (`gpu/vk/wb_lazy_state.h`), and each copy still goes onto the one queue after everything before it. The 600-submit line `vk: lazy write-back (600 submits)` ends with the payments that waited for a free slot and the most slots in use at once |
|
||||
| `RX_VK_GEN_SEED=<n>` | The page write generations (`gpu/vk/write_gen.h`) start at `n` and not at 1 (decimal, or hex with `0x`; 2^62 at most). They are 64 bits; the 32-bit ones of earlier builds ran out after days of play, and textures and mirrors then kept old contents. `RX_VK_GEN_SEED=0xFFFF0000` takes a play test across 2^32 within seconds: the log says `the write generations start at` and, at the crossing, `this run has made 4294967296 write generations`, and nothing in the picture should change there |
|
||||
| `RX_VK_UPSCALE_OFF=1` | (Triforce) The upscaler Off for one run: the game's gameplay upscale networks are not dispatched, their input is stretched linearly into their output (`gpu/vk/upscale_off.h`). Softer than the networks below 100%; only while the game's own upscaler draws (DLSS and FSR take the dispatch first). Not set: as before, the networks run, and on cards that run them on software multiplies the first one is stood in for. The lab's `upoff` phases switch it within a run |
|
||||
| `RX_VK_STUB_SAME_SIZE=0` | (Triforce, 11-22, the performance audit's GPUWORK-21, after Source B's and Source A's; `gpu/vk/stub_same_size.h`) On unless `0`. A gameplay upscale network that cannot run (left out or not translated: `RX_VK_NO_F16`, a card or driver that cannot build it) is stood in for by its input drawn into its output (`gpu/vk/vk_draw.cpp` `upscale_stub`) also where the two are one size: `9dc98871eb7fc101`, the network of Render resolution 100%, was declined as "no larger than its input" and skipped whole, a black 3D under the HUD. Both gameplay networks now take the typed upscale first (it reads the scene's shared-exponent picture and undoes the network's square-root encoding; it was the 3/4 network's alone), then the blit, which warns of wrong colours for both, then nothing. The picture is the upscaler Off's: softer, none of the network's anti-aliasing. Only where the network cannot run. `0`: the decline, and the 3/4 network alone in the typed upscale, as before. `RX_VK_NO_UPSCALE_STUB=1` still turns the whole stub off; `RX_VK_FORCE_UPSCALE_STUB=2` sends every network through it, for a look |
|
||||
| `RX_VK_DLSS=0/1/2`, `RX_VK_DLSS_PRESET=<letter>`, `RX_VK_DLSS_LOG=1`, `RX_VK_REFLEX`, `RX_VK_REFLEX_LOG=1`, `RX_VK_ANTILAG=0/1` | NVIDIA DLSS and Reflex, and AMD Anti-Lag 2, for one run (`=-` on Reflex or Anti-Lag 2: its extensions not even enabled): the readme's NVIDIA DLSS section. `RX_VK_DLSS_TRACE=<frames>` logs one frame's large draws every that many frames, `RX_VK_DLSS_DEBUG=motion` shows the motion vectors, `=alternate` the game's own upscale every other frame |
|
||||
| `RX_VK_FSR4=0/1/2`, `RX_VK_FSR_DIR=<folder>` | AMD FSR 4 for one run (0 never, 1 FSR 4, 2 its Native AA), and another folder for AMD's DLLs (not from `hookshot_env.txt`): the readme's AMD FSR 4 section. `RX_VK_FSR4=1/2` and `RX_VK_DLSS=1/2` go over the Upscaler row, a saved FSR 3.1 too where the engine they ask for runs; `RX_VK_FSR=1` goes over them (one upscaler at a time). The DLSS trace and debug switches above apply to FSR 4 too, on the frame chain they share |
|
||||
| `RX_VK_ASYNC=1` | The replay on a thread of its own instead of the game's submitting thread |
|
||||
| `RX_VK_FSR4=0/1/2`, `RX_VK_FSR_DIR=<folder>` | AMD FSR 4 for one run (0 never, 1 FSR 4, 2 its Native AA), and another folder for AMD's DLLs (not from `hookshot_env.txt`): the readme's AMD FSR 4 section. `RX_VK_FSR4=1/2` and `RX_VK_DLSS=1/2` go over the Upscaler row, a saved FSR 3.1 too where the engine they ask for runs; `RX_VK_FSR=1` goes over them (one upscaler at a time; `=2`: FSR 3.1 as its Native AA, the 3D at full size). The DLSS trace and debug switches above apply to FSR 4 too, on the frame chain they share |
|
||||
| `RX_VK_FSR4_DRS=1`, `RX_VK_FSR4_AUTO_EXPOSURE=0` | (Triforce) AMD FSR 4's context as before, for comparing (`gpu/vk/vk_fsr4.cpp` `make_context`, `gpu/vk/temporal_rules.h` `plan_context`): `RX_VK_FSR4_DRS=1` one context with FSR's dynamic-resolution flag for every render size from the start, where a fixed Render resolution otherwise gets a context of its own size without it (and the game's dynamic resolution one with it once the size first changes); `RX_VK_FSR4_AUTO_EXPOSURE=0` its HDR contexts (Upscale before bloom, or an RGBA16F picture with RenoDX's HDR Native) without FSR's auto exposure. A creation or evaluation of DLSS or FSR 4 that fails leaves the game's upscale for 2 s and is tried again; the fourth failure in a row turns that engine off (the Upscaler row then shows Game and says so) until the row is set again |
|
||||
| `RX_VK_SHARPEN_3D=0`, `RX_VK_FSR3_BOOST=0`, `RX_VK_UPSCALE_LOD_BIAS=0/<negative>` | (Triforce) For comparing with before (`gpu/vk/sharpen_rules.h`): the Sharpness row sharpens the 3D before the HUD with DLSS, DLAA, FSR 3.1 and FSR 4 (FSR 3.1's own RCAS, FSR 4's own sharpening, FSR 3.1's RCAS pass over DLSS's picture) and then nothing at the window; `RX_VK_SHARPEN_3D=0` the finished picture at the window for every upscaler, the HUD with it, as before. `RX_VK_FSR3_BOOST=0` FSR 3.1's sharpness the row's alone, without the share towards the most below three quarters of the output's width. `RX_VK_UPSCALE_LOD_BIAS=0` no texture LOD bias while DLSS or FSR scales the 3D with jitter (unset: log2(render / output) - 0.5 in quarters of a level, so -1.5 at Performance, -1 at Quality and -0.5 at the Native AAs, on the wrapping, mipmapped textures of the draws into the scene at the upscaler's render size (or, in the frame the dynamic resolution steps, at the size the scene's camera is drawn at), never the shadow maps, captures, other passes or dispatches; the log says `vk: texture LOD bias` at each change); a negative number, such as `RX_VK_UPSCALE_LOD_BIAS=-1` or `=-1.5`, that bias (-3 at most) in the same draws and only while an upscaler with jitter runs, for trying other values (the log says it once at the start) |
|
||||
| `RX_VK_FSR4_FG=0/1/4` | AMD's frame generation, the Frame generation row's AMD FSR (`gpu/vk/vk_fsr4.cpp` `fsr4fg`): 0 never in this run (`RX_VK_FSR4=0` too, which keeps AMD's library unloaded); 1, as without it, the newest frame generation AMD's library lists for the card, FSR 4's where listed (RX 9000), else FSR 3.1's (3.1.6 on RX 6000 and 7000), the row naming it ("AMD FSR (3.1.6)") and the log the version in use; 4 FSR 4's alone, as before (FG-1) |
|
||||
| `TRIFORCE_MENU_AHEAD=0` | (Triforce) What the options' rows ask of NVIDIA's and AMD's libraries (whether DLSS, FSR 4 and FSR 4's frame generation run here: NGX's start, AMD's loader) asked at the options' open on the presenting thread, as before, and not once on a thread of its own as soon as the game runs (`gpu/vk/vk_menu.cpp` `menu_ahead`; its log line says how long each took). For comparing: the first open waited about 2 s for them in test11 |
|
||||
| `RX_VK_AA=<0..3>`, `RX_VK_AA_KEEP=1` | The anti-aliasing for one run, over the Insert menu's Anti-aliasing row: 0 the game's own SMAA, 1 none, 2 FXAA, 3 SMAA at its Ultra preset in its place (`gpu/vk/aa_rows.h`, `gpu/vk/vk_aa.cpp`). Unset, the row decides; with DLSS, DLAA or FSR drawing from the tone mapped picture (Upscale before bloom off) the game's SMAA is left out (none: they want jittered frames no other anti-aliasing has been over), and before the bloom it runs as the game's; the switch goes over both, for an A/B (`RX_VK_AA=0`: the game's SMAA under the upscaler). The game's SMAA is three compute programs known by their keys (game build `e0f3ccfa`): with another choice in effect its composition is replaced and its edge and weight passes are left out, and `RX_VK_AA_KEEP=1` keeps those two running. The log says `vk: aa:` when a composition is first replaced, why one was left to the game, and when none was found for a minute |
|
||||
| `RX_VK_ASYNC=1` | The replay on a thread of its own instead of the game's submitting thread (off by default). The game still presents only once the GPU has run its frame, so nothing is released early. (Triforce, `gpu/vk/async_rules.h`; not yet run in the game) With it on, each with a switch of its own to put it back: a signal of the CPU's takes its capture on the game's thread only when a submission handed over before it may still wait for that event, which the host tells at the submit (`RX_VK_ASYNC_EVENT_NOTES=0`: at every signal while the thread has work, as before; the `vk: the replay thread: ... signals of the CPU's` line counts both); the submit's capture shares its copies with the copy helpers and has them all made before the hand-over (`RX_VK_ASYNC_CAPTURE_BEHIND=0`: every copy on the game's thread); the game's submit keeps its virtual core for the hand-over (`HOS_ASYNC_KEEP_CORE=0`: given up and taken again); the game's present poll waits for the replay thread asleep and looks at its event between yields only once the GPU has the frame (`HOS_PRESENT_POLL=0`: the game's own sleeps); a present runs the jobs behind it until the frame's stopped work has gone on and then makes go only what was already stopped at the present before (`RX_VK_PRESENT_DRAIN=1`: every present forces every stopped queue past its wait, as before; `=0`: none does; `RX_VK_PRESENT_WAIT_MS`, 8: how long it runs the jobs behind); and the thread looks for its next job for 100 µs before it sleeps (`RX_VK_ASYNC_SPIN_US=<0..1000>`). Still open, and why it stays off: after a ten-area tour the player's character is drawn black for one frame, 0 to 3 times a minute at the Market (`gpu/vk/async_rules.h` has what the lab ruled out). Two diagnostics for it: `RX_VK_LOG_READBACK_ZERO=1` logs each readback of GPU results that is all 0 where it was not at its readback before, and `RX_VK_REGION_TRAP=1` (with `RX_VK_PASS_TRACE=1` and the frame log) logs the passes of the last frames and writes their render conditions to `region_cond_<n>.txt` when the picture's bottom middle goes dark for a frame, eight times a run. `RX_VK_ASYNC_FORCED=1`: the thread as it was before 11-14, every catch-up forcing |
|
||||
| `RX_VK_OCC_QUERY=0`, `RX_VK_OCC_QUERY_SKIP=<addr>[,...]`, `RX_VK_OCC_COND=0`, `RX_VK_OCC_LOG_COND=1`, `RX_VK_OCC_GATE=0` | (on by default since 11-17: +2.7 to +4.6% fps in every scene measured, 10-12% fewer draws, no flicker added; `RX_VK_OCC_QUERY=0` turns it off) The game's samples-passed reports (which objects show; a count of 0 is not drawn) get their real counts: the replay still writes each report as it passes it with 4096 (`RX_VK_OCC_FIXED`, everything visible), and the draws from its reset are counted by Vulkan occlusion queries (precise where the card has them); once the GPU is done with the recording that holds them the count is written into the report's record over the 4096, as NX2navi writes it at completion. The done recordings are found with one look at the recordings' timeline (nothing waits) as a game's fence wait ends (the game reads its counts after it), at each recording's begin and each present, at the latest when the recording's slot comes round. The game's report records rotate (a ring: an address is another object's at its next report), so a count is written only while its report is still the newest at its address, and only when it is whole (every query available, no draw left out, reset and report in one recording, read back within 8 presents); otherwise the record keeps its 4096. The game also predicates draws and clears on these words (render conditions), decided as NX2navi decides them: a word whose newest report is a samples-passed one whose count is still to come is unknown and the condition renders; a word whose count was written (or for which none comes) is read from memory as before, where a written 0 skips the draws and clears it gates. A count goes in only over the value the replay left in the word (a compare-and-write: never over the game's CPU re-arming the record), and a record rewritten in the stream after its report (a small buffer clear, a report of another counter type) gets no count. A parallel replay's segments signal the recordings' timeline and are sealed as they are submitted, so their counts are written too. The count word is written as the replay writes a report (`guest_cpu_write`, and the word kept out of the mirrors' write-backs). With hostQueryReset (Vulkan 1.2) a recording's queries are reset on the CPU once read, only those it used. `RX_VK_OCC_QUERY_SKIP`: report addresses that keep the fixed answer (the log's `occlusion query sites` lines list the first 64 addresses with their counts). The log's `occlusion queries` line every 600 presents: the reports, how many were counted whole and why not, the counts written at completion (and of them 0), those not written and why (lost, superseded by a later report at the address, late, a render condition's), the mean latency in presents and how many were written after the next present, where the done recordings were found, and the render conditions on samples-passed words (rendered because a count was pending, rendered or skipped from a written count, from memory with no count to come, as before in the knob's off phases) with the draws and clears they skipped. `RX_VK_OCC_COND=0`: a condition on a samples-passed word always renders (the counts are still written), to tell the counts from the skips. `RX_VK_OCC_LOG_COND=1`: one such condition in 499 logged (300 lines), with both words, their sources (pending, written, memory) and the present. (11-22) `RX_VK_OCC_GATE=0` turns off (on by default since its lab A/B: no fps lost at the Market): a count over a draw that a written count's condition left out (between the reset and its report) is the skip's 0, not the object's, and a condition on it renders; the count is still written for the game's CPU. Why: the game's CPU reads a count after its frame wait and draws a hidden object's proxy instead of the object, but most counts land after it has read the 4096, so it goes on drawing the object under the condition, the condition reads the skip's 0 and the object stays hidden (tickets 11-21: items and trees in the Forest Temple, black interiors; `RX_VK_OCC_QUERY=0` showed them). The line's second half counts these counts (counted whole, written, of them 0), the conditions that read one and those the switch rendered, on or off. Cost with it on: a hidden object under such a condition is drawn every other frame instead of never. Lab: the knob `occq` in `RX_VK_LAB_AB` runs the queries the whole run and writes the counts of the reports made in its phases (`gpu/vk/occ_answer.h`, `vk_replay.cpp` "occlusion queries") |
|
||||
| `RX_VK_OCC_DEPTH_ONLY=1` | (Triforce, 11-22, the performance audit's GPUWORK-01, after Source A's; off unless set) Only the occlusion windows whose draws write no colour are counted. The game's occlusion tests draw depth only, but a query here opens at a window's first draw in the pass and stays open to the report or the pass's end, so it can enclose draws that write colour, and a query's end waits for the work it encloses. With the switch, a window's first draw that writes a colour target gives its count up: the open query ends before that draw, no query begins again until the next reset, and the report is incomplete, so its record keeps the 4096 placeholder and what it gates is drawn. Fewer objects are culled with it (Source A: 4% fewer for 0.25 ms of GPU time), never one missing; whether that pays here is for a measurement to say, which is why it is off. The lab's knob `occdepth` (`RX_VK_LAB_AB`) is the switch in its phases. Counted either way, in a third `vk: occlusion queries` line every 600 presents: the counting windows that met a colour draw, the colour draws recorded under an open query, the counts given up and the reports kept 4096 for it (`gpu/vk/occ_answer.h` `Counter::colour_draw`; `gpu/vk/vk_replay.cpp` `occ_draw_colour`; `gpu/vk/vk_draw.cpp` `draw`) |
|
||||
| `RX_VK_OCC_LEAN=3` | (Triforce, 11-22, the performance audit's DESC-82, after Source B's: its first step, the census; off unless set) The game's occlusion proxies write no colour, yet each runs its whole fragment shader and fills its fragment descriptors. With `3`, each draw an open occlusion query counts that writes no colour is classed (`gpu/vk/occ_lean.h`): lean, or the first refusal of a pipeline without the fragment stage (no fragment stage, depth out, sample mask, demote, stores, image writes, atomics, global memory, alpha to coverage, alpha test), and its fragment descriptors added to its class; every 600 presents `vk: occlusion proxies (RX_VK_OCC_LEAN=3` gives each class's draws a frame and descriptors a draw. Nothing else changes: no pipeline, key or descriptor. The entry's lean pipelines (`1`) and their check (`2`) are built only if the census finds more than one fragment descriptor a proxy to save; they are not built, and read as off. `tests/test_occ_lean.py` |
|
||||
| `RX_VK_GATE_WAIT_MS=<1..100>` | (Triforce, off unless set, after NX2navi's `gate_wait_ms`, whose default is 10; not yet run in the game) A submission whose first command buffer begins (after state only) with a wait for an event the game signals from the CPU, unmet in the replay's order, waits at its submit for that signal, at most this long, before its capture and its replay (`gpu/vk/gate_wait.h`, `gpu/vk/vk_replay.cpp` `gate_wait_signal`): the capture then holds what the game wrote up to its signal, and the submission is replayed in the submit call on the submitting thread, instead of stopping at the wait and going on from the order thread (`sync_watch`) with a second capture once the signal came. The wait holds no lock and looks at the event's word in the game's memory. One that runs out goes as before; after 4 that ran out, once they are more than a tenth, it is off for the run (logged). A `vk: RX_VK_GATE_WAIT_MS` line every 600 waits gives their mean and how many ran out; the order line's `captures at a CPU event's wait` and `made to go by presents` say how often the old way was taken |
|
||||
| `RX_VK_WAIT_HOLD_PAUSE_US=<us>` | (Triforce, 11-22, the performance audit's THREAD-11, after Source B's; off unless set, not measured here) A wait of the game's for a fence whose work is stopped at an event's wait holds up to `RX_VK_WAIT_HOLD_US` (3 ms) for another queue to let it on before it forces it, looking again and again: a flush, the timeline's value, the order lock to ask whether the stopped work is still needed, and a yield of the processor. With `<us>` (2 is Source B's; 1000 at most) a gap of PAUSEs that long stands between two looks instead of the yield, and the order lock is taken only when one of its sections has left since the waiter last asked, which is the only time its answer can change; the fence's own value is still read at every look, and the bound and the forced drain stay. Source B measured 59.2 to 59.9 fps and a 1% low of 40.0 to 42.1 with such a gap on a loop that runs far more often than this one (about 15 holds a second here). `0` or not set: as before. Either way the `vk: queues ordered by their events` report is followed by a `vk: fence waits held for stopped work` line: the holds, their looks, the time held, the takes of the order lock and how the holds ended (`gpu/vk/wait_hold.h`, `gpu/vk/vk_replay.cpp` `wait`) |
|
||||
| `RX_VK_WAIT_HOLD_PAUSE_US=<us>` | (Triforce, 11-22, the performance audit's THREAD-11, after Source B's; off unless set, not measured here) A wait of the game's for a fence whose work is stopped at an event's wait holds up to `RX_VK_WAIT_HOLD_US` (3 ms) for another queue to let it on before it forces it, looking again and again: a flush, the timeline's value, the order lock to ask whether the stopped work is still needed, and a yield of the processor. With `<us>` (2 is Source B's; 1000 at most) a gap of PAUSEs that long stands between two looks instead of the yield, and the order lock is taken only when one of its sections has left since the waiter last asked, which is the only time its answer can change, or while a signal of the CPU's (`nvnEventSignal` without the replay thread) is queued and no section has acted on it yet; the fence's own value is still read at every look, and the bound and the forced drain stay. Source B measured 59.2 to 59.9 fps and a 1% low of 40.0 to 42.1 with such a gap on a loop that runs far more often than this one (about 15 holds a second here). `0` or not set: as before. Either way the `vk: queues ordered by their events` report is followed by a `vk: fence waits held for stopped work` line: the holds, their looks, the time held, the takes of the order lock and how the holds ended (`gpu/vk/wait_hold.h`, `gpu/vk/vk_replay.cpp` `wait`) |
|
||||
| `RX_VK_INDEX_REBASE=0` | (Triforce, 11-21, the performance audit's DESC-12) A direct indexed draw binds its index buffer at offset 0 and gives its offset as the draw's first index, so draws at different offsets of one buffer bind it once: bound at each draw's own offset, nearly every indexed draw called `vkCmdBindIndexBuffer`. Only where the offset is a whole number of indices; indirect indexed draws are as before. `RX_VK_PROF_FINE=1` counts the index binds looked up and those made. `0`: every draw binds at its own offset (`gpu/vk/index_rebase.h`, `vk_draw.cpp` `draw`, `bind_index`) |
|
||||
| `RX_VK_RECORDER_OWN=0` | (Triforce, 11-21, the performance audit's REPLAY-04) With the recorder thread, a recording slot's descriptor pools are reset, its extra pools destroyed and its grown streams given back in `draw_begin_recording`, after the recording's first packets, as before: each is a boundary the replay waits at until the recorder has run what came before. The default makes that reset in `begin_recording` right after the slot's command pools are reset, where the recorder has nothing held. The 120-present line `a frame's recorder thread` counts the replay's waits (`gpu/vk/vk_draw.cpp` `draw_slot_reset`, `vk_replay.cpp` `begin_recording`) |
|
||||
| `RX_VK_DESC_POOL_GROW=0`, `RX_VK_DESC_POOL_DIV=<1..32>`, `RX_VK_STREAM_MB=<MB>` | (Triforce, 11-22, the performance audit's DESC-33, after Source A's; `gpu/vk/desc_pool_grow.h`) Each recording slot (48 at most) keeps a descriptor pool for the programs over the push limit, and the driver may keep host memory for all of it 48 times. A slot's pool is made at half size (4096 sets, 16384 uniform buffers and sampled images, and so on) and made whole (8192 sets) at the slot's first reset after a recording ran out of it; a recording that runs out gets a whole extra pool on the spot, as it always did. `RX_VK_DESC_POOL_GROW=0`: whole pools from the start, as before. `RX_VK_DESC_POOL_DIV=<n>` divides both sizes by n. `RX_VK_STREAM_MB=<n>` is the size of each slot's first upload stream (2 by default; a recording that needs more is moved to a bigger one, kept for the slot's next 120 recordings; 24 with `RX_VK_STREAM_GROW_OFF=1`, which skips draws instead). Every 3000 recordings the log's two "the recording streams" lines say the most one recording held of its stream, how many slots are still at half, how many were made whole, the extra pools made, and the most sets and descriptors of each type one recording took: what a smaller default is judged by (the stream's stays 2 MB until those numbers are read in play) (`gpu/vk/vk_draw.cpp` `make_pool`, `slot_reset_locked`, `draw_begin_recording`) |
|
||||
| `RX_VK_CAP_SPARE_FIT=0`, `RX_VK_CAP_STEP_MB=<n>` | (Triforce, 11-22, the performance audit's MEM-33; the first on by default, the second off unless set; not yet run in the game) `RX_VK_CAP_SPARE_FIT=0`: a capture with no spare buffer makes its own with the spares' lock held, takes any spare that fits, and a spare made ahead is made once, as before. Without it (`gpu/vk/cap_spare.h`, `gpu/vk/vk_memory.cpp` `cap_buf_get`) the buffer is made with the lock let go, so the replay's take does not wait behind a submitting thread's allocation to give a buffer back; where spares are not capped at 32 MB a capture leaves alone a spare of 32 MB or more that is over twice its size class, for the load that wants it; the spare a capture of a class of 32 MB or more takes is made again on a worker; a spare is made ahead only if it fits what spares may hold together with the ones held and being made, and a buffer given back counts the ones being made; when the allocation fails, any spare that fits, then the exact size. `RX_VK_CAP_STEP_MB=<n>`: where spares are capped at 32 MB (a card of 6.5 GB or less), a capture over 16 and up to 32 MB gets a buffer in steps of n MB, not the whole 32 MB (Source A: 4; its worst frames were worse in two pairs, so off here). The 600-take line counts the buffers made on a submitting thread and their ms |
|
||||
| `RX_VK_MIRROR_PIECE_IDLE=<submissions>` | (Triforce, 11-22, the performance audit's MEM-26, its first step only; off unless set; not yet run in the game) A mirror in pieces keeps a piece's memory until the whole mirror goes, also where no game pool is left in its import. Giving such pieces back is NOT ported (the unbind and the free must wait for the GPU and for the bind timeline). What it would find is counted (`gpu/vk/piece_idle.h`): every 1200 submissions, after the `mirrors in pieces` line, the MB of pieces no game pool covers; with this switch each binding given a mirror in pieces also notes the submission in its pieces, and the line says how many of those MB no binding reached for that many submissions (600 is the wait the release would use). Under about 100 MB at the end of a tour on a 6 GB card (`RX_VK_TEST_CARD_MB=6144`) the release is not worth porting |
|
||||
| `RX_VK_RANGE_WALK=0` | (Triforce, 11-22, the performance audit's MEM-42; on by default; not yet run in the game) The capture and collection loops (`gpu/vk/vk_memory.cpp` `memory_collect_written`, `capture_make`'s staging and behind loops, `take_capture`'s loops over the texture storage's ranges) ask of every page whether it lies in a sorted range list with a binary search, as before. Without it each loop has a cursor that only moves forward (`gpu/vk/range_walk.h`), one pass over the ranges for the loop; an address below the last one asked searches again, so any order gives the search's answers (`tests/test_range_walk.py` compares them). The `RX_VK_LAB_AB` knob `norangewalk` searches in its phases |
|
||||
| `RX_VK_TABLE_STREAM=0` | (Triforce, 11-22, the performance audit's SYNC-04; on by default; not yet run in the game) The guest page table's changes of a recording (the table the shaders' pointers go through, `gpu/vk/vk_memory.cpp` `memory_table_flush`) are flushed as before: a (page, value) pair for every page changed, all of them sorted, the last of each page kept: 16 bytes a page and a sort on the recording thread, a few ms when a pool of hundreds of MB is imported ahead or merged. Without it the changes are streamed (`gpu/vk/table_stream.h`): runs in page order and apart are appended one after the other, any other set is swept by page with the newest change owning each page, through one reused batch of 8,192 entries. The transfers are the same ones in the same order (`tests/test_page_table_stream.py` compares the two, entry for entry). A flush over 1 ms is logged with its pages, changes and transfers (the first 16) |
|
||||
| `RX_VK_CAPTURE_BEHIND=0` | (Triforce, on by default; `=0` as before) The submit-time capture (`gpu/vk/vk_memory.cpp` `capture_make`) takes the write marks on the game's submitting thread and copies there only the pages read at once (the late pages and the texture storage); the rest of its copies, in parts of 32 pages, go to the capture's copy helpers while that thread replays the submission (`gpu/vk/capture_behind.h`). A read of a page as submitted copies its part itself when no helper has taken it yet, or waits for the helper; every part is done and flushed before a recording is submitted and before the thread goes back to the game. Only with the replay on the submitting thread (not `RX_VK_ASYNC=1`), with copy helpers (`RX_VK_CAPTURE_THREADS`, not `RX_VK_LOW_SPEC`), at least 128 pages behind and at most 512 ahead (a scene's load: copied whole as before). At the Market the old copy was ~0.25 ms of each submit's ~0.35 ms capture, six submits a frame. The 120-present line `a frame's captures at the game's submits` gives their time on the game's threads (the scan, the copies made there, and the copies behind made or waited for by the replay), the MB copied and behind, and who copied the parts; the stall line names the captures when they took 1 ms or more |
|
||||
| `RX_VK_CAP_FREE_ASYNC=0` | (Triforce, on by default; `=0` as before) A retired capture buffer the spares cannot keep (past 12 buffers or 768 MB, 32 MB on a card of 6.5 GB or less) is destroyed on a thread of its own (`gpu/vk/vk_memory.cpp` `cap_buf_free`, `destroy_buffer_later`), not in the capture's take on the replay's thread with its locks held: freeing a 64-256 MB host-cached buffer makes the driver unmap and release every page, the `retires` part of a `vk: slow capture take` line. Only buffers the GPU is done with get there; the memory comes back a moment later. The 600-take line counts the buffers and MB handed over. With no thread to be had they are destroyed where they were retired |
|
||||
| `RX_VK_PARALLEL=<0..8>`, `RX_VK_PARALLEL_AUTO=<2..8>` | The parallel replay: a frame's large command buffers recorded on that many threads, always, or wherever short trials as the game runs find it faster than the replay thread alone; `RX_VK_PARALLEL=0`: never. The Debug tab's Replay helper threads row is the second without an environment to set (`replay_helper_threads=` 1 to 7 for 2 to 8 threads); either switch goes over the row. The row's Automatic (8), which it is unless the options file says otherwise, is the count `gpu/vk/replay_threads.h` gives the PC: none with four cores or fewer or on AMD's driver for RDNA 1 and 2, 2 threads with five to seven cores, 3 with eight or more. The log says at the start what was chosen (`replay helper threads:`) and at each trial which way is in use (`parallel replay, trial`). Between trials the tuner's rule (`gpu/vk/replay_threads.h` `rule`) judges every 120 frames on the replay thread alone: where the replay is plainly the frame's limit (70% of the frame or more) and the helper threads would take at least 1 ms and a tenth of the frame off it, never counting below the GPU's time (the game waits for each frame's GPU work), they are turned on without a trial; where they cannot pay (no command buffer large enough to cut, the GPU's time about as long as the replay's, or the replay under 60% of the frame: the game's own threads are its limit) no trial is made while that holds; trials decide the rest, and a trial that takes the helpers back off holds the rule's "on" for 3,600 presents. The log says so (`parallel replay, the rule`) when its verdict changes. `RX_VK_PARALLEL_RULE=0`: trials only, as before. With DLSS or FSR on, for upscaling or frame generation, the dry pass tells their hooks in the draws in stream order, and each command buffer from the first draw or dispatch that takes what they noted (the upscale, or the engine's point before the bloom) on is recorded in order after the segments before it (`gpu/vk/stream_order.h`); `RX_VK_PARALLEL_TEMPORAL=0`: the replay thread records alone then, as before. The frame statistics' `a frame's parallel replay:` line says in how many frames the threads recorded, the command buffers in segments and alone (and why), and the time in them; the next one (`its parts`) where that time went: the setup before the threads begin, how soon each command buffer's first segment reached the GPU, the helper threads' wake-up, each command buffer's longest segment, the replay thread's wait, the submits, the recording lock's waits, and the segments' GPU time (the `GPU time` line counts the segments too; it left them out). After each trial's line a `trial N's parts` line gives its three parts' frame, replay thread's time, GPU time, the rest of the frame, full barriers and draws, and the helper threads' parts. That header lists what an audit of the parallel replay found and fixed, and what is left. `RX_VK_NO_DRY_COPIES=1`: the dry pass does not mark what copies and texture clears write, as before. The replay thread records the next segment to submit itself when no helper thread has begun it (it only waited before, on the core kept for it); `RX_VK_PARALLEL_OWN=0`: it waits for the helper threads, as before. `RX_VK_PARALLEL_SEG=<32..>`: the draws and dispatches a segment is cut for (160); `RX_VK_PARALLEL_SEGS=<2..12>`: the most segments a command buffer is cut into (6), without which a smaller `RX_VK_PARALLEL_SEG` cuts no finer once a command buffer has six segments' worth |
|
||||
| `RX_VK_PARALLEL_BACKOFF=0` | The tuner's trials as before: one every 1,800 presents after one that kept the way, each with 104 frames the other way. With the back-off (the default, `gpu/vk/replay_threads.h` `Tuner`) each trial that keeps the way, the other way longer than both of its own parts, puts the next one off twice as long (3,600, 7,200, then 14,400 presents at most; a margin of 10% or more counts twice), one too close to tell keeps the old interval (the first such one after a change of scene or way: 900 presents), and once the way has held so the other way is tried for 64 frames, not 104; a short trial that finds the other way shorter changes nothing itself, a full trial follows at once. The next trial comes early (900 presents after the last one at the soonest, never while the rule says no trial) when the scene clearly changes: two windows of 120 frames in a row whose draws moved by a quarter (200 at least) or whose replay or GPU time moved by 30% (1 ms at least) from the first window after the way was set, or a trial whose first and last parts did. In the lab's steady Market a trial came every ~2,090 presents, each ~100 frames 1.9 ms longer; the trial lines now end with when the next may come, and `parallel replay, the scene changed` says when one was brought forward |
|
||||
| `RX_VK_TUNER_SETTLED=0` | The tuner's rule judges every window and a trial the scene changed under decides, as before. With settled scenes only (the default, `gpu/vk/replay_threads.h` `Tuner`, after Kaefergeneral's #dev-fixes post: on his 5900X and RTX 3090 the rule turned the helper threads on in an area's load and the first trial after the arrival kept them, and Automatic was 6 to 10% slower than Off) the rule judges a window of 120 frames only if the one before it describes the same scene (its draws within a quarter, 200 at least, its replay and GPU times within 30%, 1 ms at least: the back-off's test), so a load's window and the first after it get no verdict and the rule's "on" comes a window after a scene settles; a trial whose first and last parts moved so changes nothing, either way, and the next comes 120 presents after a window like the one before it, not 900 or more later. The trial's line says so (`the scene changed under it`), and `parallel replay, the scene settled` when the next is set |
|
||||
| `RX_VK_RECORDER=thread`, `=defer`, `=sync`, `=off` | (unset: `thread` on an NVIDIA GPU with 12 logical processors or more, and on an AMD GPU with 6 physical cores or more in the process's affinity but for AMD's Windows driver for RX 5000 and RX 6000 (`gpu/vk/recorder_auto.h`); else off. The log's `vk: recorder thread on/off by default` line says which rule applied; `RX_VK_LAB_AB=norec` makes the calls directly in its phases, to measure it) `thread`: the replay thread's `vkCmd*` calls packed with copies of their arguments into a ring and made by a recorder thread of their own, in order (`gpu/vk/vk_recorder.h`); the replay thread waits for it at every command buffer begin, end and reset, submit, wait, readback and destruction. Beside the Replay helper threads too (no longer refused with `RX_VK_PARALLEL`): a parallel replay's segments make their calls themselves, after the ring has been run (`segments_begin`), and the recorder is idle while they record. With it on, the frame statistics' `a frame's recorder thread` line says how many of the replay's calls it made a frame and in how long, and how long the replay thread waited for it; what it saves is the replay time line's `commands` with it against without it. `defer` and `sync` are its test steps (held and run on the same thread). `RX_VK_RECORDER_RING_MB` (8), `RX_VK_RECORDER_SPIN_US`, `RX_VK_RECORDER_STATS=0`, `RX_VK_RECORDER_POISON=1`, `RX_VK_RECORDER_PIN` (`next`, `sibling` or a mask) |
|
||||
| `RX_VK_REPLAY_PRIO=0` or `1` | (Triforce, on by default, after NX2navi's `recorder_prio`; `=0` as before) A thread runs at the highest priority while it replays the game's frames (the submitting thread's replay and its capture, the present, the stopped work the order thread lets on: `gpu/vk/vk_replay.cpp` `ReplayPin`) and gets its own priority back after, so the game's code it runs between replays keeps the game's priority; the core kept for the replay is kept from the game's threads only, and the helper threads and the capture's copy helpers (above normal) outranked a replay at the game's priority there. A thread at that priority or above already is left alone. The replay thread of `RX_VK_ASYNC=1` and the recorder thread of `RX_VK_RECORDER=thread` take the level for good. `=1`: above normal. The log's `vk: a thread that replays runs at` line says which (`gpu/vk/replay_prio.h`) |
|
||||
| `RX_VK_REPLAY_PIN_SOFT=0`, `=2` | (Triforce, 11-22, the performance audit's THREAD-09, after Source A's; Windows) A thread is pinned while it replays to the core the host keeps for the replay, and that core is kept from the game's threads, not from other programs': where another program's high-priority thread holds it, the pinned replay waits for it, ready to run (Source A, a 6+8-core hybrid processor with an overlay service at priority 15: 14 to 37 frames over 40 ms in a 40 s run, none with this). Not set: on a processor with cores of more than one kind the pin is soft: the thread may run on the reserved core and on the fastest-kind cores among its own processors, the reserved core's last logical processor is its ideal one, and the core it is on when the pin starts is left out so that it moves at once; its affinity and ideal processor come back after the replay. The replay thread of `RX_VK_ASYNC=1` gets the same processors for good. A processor of one kind of core keeps the strict pin, and nothing changes there. `0`: the reserved core alone everywhere, as before; `2`: soft everywhere (the lab; the knob `pinsoft` of `RX_VK_LAB_AB` does it by phases). A thread of another program that wakes on the core every millisecond moves the replay again and again (Source A: 54.3 against 44.6 fps, one run each): `0` is for that. The log's `vk: a replay's pin is` line says which (`gpu/vk/replay_pin.h`, `gpu/vk/vk_replay.cpp` `ReplayPin`, `replay_pin_soft`) |
|
||||
| `HOS_VCORES=0`, `HOS_CORES=<1..8>`, `HOS_RESERVE_CORE=0` | No virtual cores; the cores the game sees (6); no core kept for the replay (one is, with 8 or more logical processors) |
|
||||
| `HOS_VCORES=0`, `HOS_CORES=<1..8>`, `HOS_RESERVE_CORE=0`, `=1` | No virtual cores; the cores the game sees (6); no core kept for the replay (one is, with 8 or more logical processors). (Triforce, 11-22, the performance audit's THREAD-49, after Source A's) On the cache die of a two-die X3D processor with SMT off none is kept unless `HOS_RESERVE_CORE=1` asks for it or the replay has a thread of its own (`RX_VK_ASYNC=1`): the game then has all 8 of the die's cores (Source A, a 9950X3D with SMT off: 59.8-59.9 fps with none kept against 56.9-57.4 with one). The log's `cpu: no core kept for the replay on the cache die` line says so. With SMT on and on every other processor nothing changes (`host/hos_x3d.h` `hos_x3d_keep_replay_core`) |
|
||||
| `HOS_AV1=mf`, `HOS_AV1_MISSING=1` | (Triforce, 11-22, the performance audit's IO-32; Windows) The game's AV1 movies on Media Foundation first (Microsoft's AV1 Video Extension), as before, and on dav1d only where the PC has no such decoder. Without it dav1d goes first where the FFmpeg the build links has libdav1d, as on Linux, and Media Foundation where it has not or where libdav1d does not open (`host/video_av1.h`, `host/nn_video.c` `dec_init_sized`): Media Foundation's path decodes in software on every GPU (`host/mf_av1.c` takes the synchronous decoder). `HOS_AV1=dav1d`, the old way to ask for dav1d first, is the default now. `HOS_AV1_MISSING=1`: Media Foundation as if the extension were missing. Each movie logs one line at its Stop, Flush or its decoder's end: its pictures, the decoder, the mean and the worst time to decode a submission, and the pictures skipped to catch up (`host/video_movie.h`). The two decoders were not timed against each other in the game |
|
||||
| `HOS_VIDEO_RGBA=1` | (Triforce, 11-22, the performance audit's IO-36) Every planar 4:2:0 movie picture (libdav1d's) is converted to RGBA when it is decoded, as before: a swscale pass, an 8 MB allocation and an 8 MB row copy for a 1080p picture, on the movie thread. Without it such a picture is kept as NV12 alone, as Media Foundation's NV12 pictures already were, and its pool makes the RGBA only if the game reads RGBA, which this game's movie library does not; planar 4:2:0 with alpha is converted at once either way. The NV12 buffer is zeroed only for an odd size or a format that is not copied into it, and U and V are interleaved 16 pairs a step with SSE2 (`host/video_nv12.h`, `host/nn_video.c` `queue_frame`). The movie's line in the log (IO-32) has the time a picture took with either. Not timed in the game |
|
||||
| `HOS_VIDEO_UNBOUNDED_CATCHUP=1`, `HOS_VIDEO_NO_CATCHUP=1` | (Triforce, 11-22, the performance audit's THREAD-82) The movie decoder's acquire skips the pictures behind the movie's clock and decodes the next ones to skip them too. `HOS_VIDEO_UNBOUNDED_CATCHUP=1`: with no bound but a guard of 600 steps, as before: with a decoder slower than the movie (19.5 ms a picture against 16.7) each picture decoded to catch up puts the clock further ahead, so an acquire decoded up to about 300 pictures with its lock held, threw them away, and the next started further behind. Without it each decode that gives a picture is timed, about the last 8 kept as one number, and the acquire stops decoding ahead when that number is more than the gap between two pictures (from the pictures' own times) or once it has decoded for one gap, and shows the late picture it has (`host/video_catchup.h`, `host/nn_video.c` `dec_acquire`): such a movie falls behind its sound instead of stalling, and its line in the log says "slower than it plays". A decoder faster than the movie catches up as before, a few pictures an acquire. `HOS_VIDEO_NO_CATCHUP=1`: no picture skipped at all (the movie in slow motion under 60 fps). Not run in the game |
|
||||
| `HOS_VIDEO_INPUT_LIMIT=<n>` | (Triforce, 11-22, the performance audit's IO-34) How many submissions the movie decoder holds, not decoded yet or decoded and not taken, before it refuses one: 16 when unset; `0`: no limit, every submission taken, as before. The game submits a whole movie at once on the thread that also takes the pictures, so at a movie's start the pictures that came due while it submitted were skipped. With the limit a submission is refused at once with the console's result for a full decoder (0x1c266f) and the game submits it again later; the stream's last submission is never refused, and a decoder with no codec behind it refuses nothing. Each decoder has its own "ready for a submission" event, clear from a refusal until an acquire, a Stop or a Flush makes room (it was one event for all, always set); the movie's line in the log counts the refusals (`host/video_input.h`, `host/nn_video.c` `dec_submit`, `input_room`). If a movie ever hangs on a black screen, try `=0` first. Not run in the game |
|
||||
| `HOS_MOVIE_PACE=0` | (Triforce, 11-22, the performance audit's PRESENT-30) The frame rate cap is kept while a movie plays, as before. Without it, at a cap of exactly 60 and with the Steady frame rate row off, the frames are paced by 240 while a movie player has updated in the last quarter second, and the game's own limiter and its time step follow as they do at a cap of 240 (`host/movie_pace.h`, `host/nn_vi.c` `movie_pace_decide`, `pace_fps`, `realtime_step`): the game's movie player takes one picture an update at most and its updates follow the frame loop, so at 60 every frame that came late was a picture the decoder skipped. Game time stays real time; the limiter is put back where it was when the movie is over, and the log has a line at the movie's start and one at its end. The cap itself (the Steady row, the vsync event, the hand-off to RivaTuner or Reflex) does not move. Movies use more CPU and power with it. Not run in the game |
|
||||
| `A64_NO_VBYTES=1`, `A64_NO_VMOVES=1` | (Triforce, 11-21, the performance audit's LIFT-02; read when the game is translated, and in the lift key) The translator writes TBL with a one-register table and EXT as loops of single bytes through the stack, a 128-bit vector load as two halves and DUP by element as a loop, each followed by a 16-byte read of the register that cannot be forwarded from those stores. `nx2/fastvec.py` rewrites each on the lifted text as one instruction's helper (PSHUFB; PALIGNR or two shifts; one 16-byte load; one broadcast), the same bytes in the same registers (`tests/test_vbytes.py` runs every shape both ways). `A64_NO_VBYTES=1` keeps the translator's text for TBL and EXT, `A64_NO_VMOVES=1` for the loads and DUP. A build with them needs a hot profile of its own (`RX_PGO=gen`) |
|
||||
| `A64_DMB_LIGHT=1` | (Triforce, 11-22, the performance audit's LIFT-47, after Source B's; read when the game is translated, and in the lift key; off by default) The translator writes every DMB and DSB as `recomp_barrier()`, an out-of-line full fence. x86-64 keeps load-load, load-store and store-store order in hardware, so with this set the eight DMBs that ask only for those (ISHLD, ISHST and the OSH, NSH and full-system forms) are lifted as `A64_DMB_LDST()` (`nx2/insn.py` `dmb_light`), which is nothing but a barrier for the compiler on x86-64 and the full fence on any other host (`nx2/rtgen.py`). A full DMB, every DSB, ISB and STLR's locked store keep theirs. Unset, empty or `0`: the kit's text. `tests/test_dmb_light.py` runs every such word both ways, alone and between a store and a load. Not counted in this game's code and not measured in the game: how many of its barriers are of these kinds is not known, and what other threads see cannot be shown by a test on one thread. With it set the hot functions that hold such a DMB no longer match the shipped hot profile: such a build needs one of its own (`RX_PGO=gen`) |
|
||||
| `HOS_FMA_FAST=0`, `A64_NO_FMA_SOFT=1` | (Triforce, 11-22, the performance audit's LIFT-06, after Source A's; only in a build without FMA3, x86-64-v2: with FMA3 none of it is compiled) Without a hardware FMA every double-precision FMADD, and each single-precision one its binary64 fast path cannot prove, ran the translator's exact body, integer code of many words (about 180 ns a call and 13.7% of a core on an i7-3770, Source A's figures). `nx2/fma_soft.py` puts a short way out at the head of `recomp_fpe_fma64` and `recomp_fpe_fma32` for the plain case: default FPCR (FZ or not), three normal operands and a normal result, by an exact product and sum in binary64, rounded to odd, then rounded once, FPSR.IXC set when inexact; everything else goes on into the body. `tests/test_fma_soft.py` compares both bodies with it and without, value and every byte of the context. `HOS_FMA_FAST=0` (read once when the game runs): the exact bodies only, as before. `A64_NO_FMA_SOFT=1` (read when the game is translated, and in the lift key): the translator's runtime text as it was. Not measured in a game on a CPU without FMA3 |
|
||||
| `A64_NO_FMA_STEPS=1` | (Triforce, 11-22, the performance audit's LIFT-05, after Source A's; read when the game is translated, and in the lift key; only in a build with FMA3, x86-64-v3) The single-precision FRECPS and FRSQRTS steps (2 - a*b and (3 - a*b) / 2, the Newton-Raphson steps of a reciprocal and a reciprocal square root) were computed in binary64 on x86-64 (two widenings, a multiply, a subtraction, a narrowing) with a test for a binary32 midpoint, which went to the exact body. `nx2/fma_steps.py` gives their fast paths the guard the fused multiply-add's already had: with FMA3 one fused instruction, the step rounded once as on the console; the tests that decide whether the result is kept are unchanged (`tests/test_fma_steps.py` runs every form both ways and against the exact body). `A64_NO_FMA_STEPS=1`: the translator's text for the two steps. The compile define `-DA64_FPX_NO_FMA32` returns all three (FMADD/FMLA and the two steps) to the binary64 form. Source A measured the steps alone at 23% to 29% fewer cycles; not measured in the game, and how often this game runs them is not counted |
|
||||
| `A64_VECFP=1` | (Triforce, 11-22, the performance audit's LIFT-03, after Source A's; read when the game is translated, and in the lift key; off by default) The translator's passes turn straight FADD, FSUB and FMUL .4S and the .4S by-element multiplies into one SSE op; other float forms stay a helper call a lane or exact integer code a lane. With this set `nx2/vecfp.py` lifts each of them as `if (!a64_vfpx_q_<form>(c, ...)) { the C as before }`: the helper does the op in SSE and writes the register only when the scalar fast paths' rules prove the result is the console's with no new FPSR bit. `1`: every group; or a list of `elem` (FMUL, FMLA, FMLS .2S by element), `vec` (FMUL .2S; FMLA, FMLS and FDIV .2S and .4S; the multiply-adds need FMA3 and decline without it), `minmax` (FMAX, FMIN, FMAXNM, FMINNM: scalar S, .2S, .4S; kept unless a NaN or, under FZ, a subnormal is involved) and `fcvt` (FCVT S to D and D to S, FCVTL, FCVTL2, FCVTN). Unset, empty or `0`: the translator's text. `tests/test_vecfp.py` runs 137 guest words three ways (with it, without it, exact). Source A measured +3.5% to +5.1% fps where the game's update bounds the frame, for 3.7% more compile time; not measured in this game, which is why it is off. A build with it needs a hot profile of its own (`RX_PGO=gen`) |
|
||||
| `HOS_HEAP_COMMIT=1` | (Triforce, 11-21, the performance audit's MEM-09) On Windows the game's heap (~6.9 GB asked for at its start) is left reserved and committed as it is touched, 64 KB at a first touch, by the handler that commits pages nothing handed out; committed whole, all of it counted against RAM plus the page file, touched or not. A fault another thread's commit answered is tried again, a commit refused at Windows' commit limit is tried again for 5 s, and a file read or write reads each page of its buffer once before the kernel's call, which faults nothing in. The log says `heap: left reserved` and counts the heap's first touches. `1`: all of it committed at the start, as before; so it is in large pages (`HOS_LARGE_PAGES`) and on Linux (`host/heap_commit.h`, `mkw_glue.c` `commit_on_touch`, `hos_mem.c` `SetMemoryHeapSize`) |
|
||||
| `RX_VK_DMABUF_NVIDIA=1`, `RX_VK_NO_DMABUF=1`, `RX_VK_DMABUF_PIECE_MB=<n>` | (Linux) Where the game's memory is a memfd, each pool of it is offered to the driver as a dma-buf made by udmabuf, larger pools as several of at most udmabuf's limit bound into one sparse buffer; the log's `vk: guest memory as dma-bufs:` line says yes or why not. `RX_VK_NO_DMABUF=1`: host-pointer imports only. `RX_VK_DMABUF_PIECE_MB=<n>`: pieces of at most n MB. (Triforce, 11-22, the performance audit's OS-08, after Source A's; not yet run on Linux) NVIDIA's proprietary driver refuses every such dma-buf and the pool ends as a host-pointer import anyway, after a refused import each (up to 198 ms for one during a load) and a kernel message at each start: with that driver none is offered, `/dev/udmabuf` is not opened, and the line says `no (NVIDIA's driver refuses them: host pointers)` (`gpu/vk/dmabuf_gate.h`). NVK and RADV keep their dma-bufs. `RX_VK_DMABUF_NVIDIA=1`: offered to NVIDIA's driver too, as before |
|
||||
| `HOS_COMMIT_WAIT_MS=<ms>` | (Triforce, 11-22, the performance audit's OS-04, after Source A's and Source B's; 5000 unless set) How long a refusal at Windows' commit limit (RAM and the page file full, error 1455) is waited out, tried again every 50 ms: such a refusal often ends in moments, as a page file set to grow grows or commit is given back. The game's own commits waited so already (the row above); now the host's own tables do too (a failed allocation there ended the game at once: `host/host_alloc.h`, `hos_xmalloc` in `host/mkw_glue.c`), and the renderer's failed `new`, which was an uncaught `std::bad_alloc`: its new handler sleeps 50 ms and lets the allocation be tried again, up to this long a thread, but only while the system has less than 4 GiB of commit left (a `new` refused with more free is not the limit's and throws at once), and not for the few allocations that have their own answer to a refusal (`gpu/vk/new_wait.h`, `gpu/vk/vk_main.cpp` `new_refused`). The log has `refused ... made after N ms` for the first rescues. `0`: no wait anywhere, a refusal fails at once; 60000 at most |
|
||||
| `HOS_TEXELS_KEEP_CORE=1` | (Triforce, 11-21, the performance audit's THREAD-20) `nvnTextureReadTexels` and `nvnTextureWriteTexels` reach the back end, which waits for all the GPU work submitted and then for the replay's lock; the game calls them from its loading threads at an area's arrival and at boot. The calling thread gives its virtual core up for that call, as it does around a readback, so the game threads allowed only on that core do not wait with it. A write kept for a renderer not started yet never waits and keeps the core. `1`: kept, as before (`host/rx_backend.c` `rx_nvn_read_texels`, `rx_nvn_write_texels`) |
|
||||
| `HOS_LOG_SYNC=1`, `HOS_LOG_BATCH_MS=<ms>` | The log's lines are written by a thread of their own, not by the thread that logs: a write that blocks (a console, a busy disk) holds no game thread. The writer waits a little after a line and writes the lines of that time in one write (`HOS_LOG_BATCH_MS`, 250 unless set, at most 2000; 0: the lines there are at each wake); a queue half full goes out at once, and every way out of the game writes the queue first. `HOS_LOG_SYNC=1`: each line written by the thread that logs it. (Triforce, 11-22, the performance audit's THREAD-56, after Source A's; not yet run on Linux) The writer is Linux's and the Steam Deck's too now: there every `hos_log` wrote stderr in the calling game thread, its virtual core held, so `HOS_LOG_SYNC=1` is Linux as it was. And `nvnCommandBufferEndRecording`'s two lines (the sweep's, the 50,000-recording line) are written after the recordings' lock is released, the virtual core given up, on every system (`host/hos_thread.c` `log_queue`, `host/nvn_native.c`) |
|
||||
| `RX_VK_INVALIDATE_QUICK=1` | (Triforce, 11-22, the performance audit's THREAD-25, after Source A's; off unless set. Source A has it on; here it waits for a run in the game, as a wrong answer skips the wait a readback of GPU results needs, the exposure flashes' kind of fault) The game invalidates each shape's 64-byte occlusion slot, per view, and the host gives the calling thread's virtual core up before every such call and queues for it again after, though nearly none has anything to wait for. With `1` the back end is asked first, with no lock and no wait: a range with nothing to invalidate, or one below 512 GB with no fast fence poll's order ticket still to wait for, no live texture's storage in it, no texture the GPU wrote ever listed there and no write-back ever made into its pages, is one `invalidate` returns from at once, so the call is left out and the core kept. The two "ever" answers are bits a page that are only set, before their tables change (`gpu/vk/page_marks.h`); a page written back once has the call made for good. Source A measured Hyrule Field at 50.5 to 51.4 fps and its 1% low up by two to four frames a second, its busiest worker keeping the core 31,444 times against 2,604. Every 10 s a `vk: invalidates:` line counts the calls left out and made. With `RX_VK_LOG_INVALIDATE`, `RX_VK_LOG_READBACK` or `RX_VK_INVALIDATE_WAIT` set every call is made; the lab's knob `noinvq` makes them all in its phases (`gpu/vk/vk_replay.cpp` `invalidate_quick`, `host/rx_backend.c` `rx_nvn_invalidate`) |
|
||||
| `HOS_OPEN_KEEP_CORE=1` | (Triforce, 11-21, the performance audit's IO-24) nn::fs::OpenFile gives its virtual core up only around the host's own open, as before. The default gives it up before the path is worked out too: that reaches the button prompts' builder, which builds a file at its first open (a quarter of a second and more) while the other openers wait for it, and none of them holds a core meanwhile. Every return takes the core back before it answers the game (`host/nn_fs.c` OpenFile, `host/prompts.c` `hos_prompts_font_file`) |
|
||||
| `HOS_SHORT_SLEEP_TIMER=1`, `=0` | (Triforce, 11-21, the performance audit's OS-31) How sleeps under a millisecond wait. Not set: with 6 or fewer logical processors in the process's mask (Windows, one processor group) the game's own short sleeps wait on the high-resolution timer, which wakes about 0.5 ms late, and the host's own yield until their time; with more, all yield, as before. The game's idle job workers sleep 20 us and look again about 7,600 times a second, and a yielding sleeper keeps a processor busy. `1` (any value but 0, off, no, false): every short sleep on the timer, as the switch did before; `0`: every short sleep yields. (Triforce, 11-22, the performance audit's OS-32, after Source A's; not yet run on Linux) Off Windows the processors are not counted and, not set, every short sleep yields, as before; `HOS_SHORT_SLEEP_TIMER=game` puts the game's own short sleeps on the timer and leaves the host's yielding at any processor count, on any system: Source A's default off Windows (on Linux the timer wakes tens of microseconds late, not half a millisecond), which was never measured in the game by itself and is to be tried on a Steam Deck before `SHORT_SLEEP_ELSEWHERE` in `host/short_sleep.h` makes it what Linux does when the switch is not set. The log's `sleep:` line says which (`host/short_sleep.h`, `host/hos_thread.c` `sleep_for`) |
|
||||
| `HOS_DISCARD_KEEP_CORE=0`, `=1` | (Triforce, 11-22, the performance audit's OS-40, after Source A's; not set: as before, which is NOT Source A's default. It reverses a choice made here on purpose, changes when a game thread waits, and was not measured here) A discard of memory the game let go is milliseconds in the kernel for a region of some MB, and before it the lazy write-back's pages there are paid and pages a GPU import holds are zeroed; the game threads allowed only on the freeing thread's virtual core wait all the while. Not set: the core is given up for the discard's own kernel call, and only with the memory limit on (`HOS_MEM_LIMIT`), as before. `0`: a freed region's thread gives its core up before the whole of the discard and takes it back once, after the region is unlisted and never under the regions' lock; the discard's own call gives it up without the limit too (`FreeMemoryPages`). Source A at its Market: core waits of 50 ms and more 287 to 19, the longest 177 to 45 ms, frames unchanged; the freeing thread once waited 124 ms for its core back, in a load. `1`: the core kept through every discard, with the limit too. With `0` the log counts such frees and the longest take-back every 1,000 (`host/hos_mem.c` `discard_core`, `hos_region_free`, `discard_pages`) |
|
||||
| `HOS_SHORT_SLEEP_TIMER=1`, `=0` | (Triforce, 11-21, the performance audit's OS-31) How sleeps under a millisecond wait. Not set: with 6 or fewer logical processors in the process's mask (Windows, one processor group) the game's own short sleeps wait on the high-resolution timer, which wakes about 0.5 ms late, and the host's own yield until their time; with more, all yield, as before. The game's idle job workers sleep 20 us and look again about 7,600 times a second, and a yielding sleeper keeps a processor busy. `1` (any value but 0, off, no, false): every short sleep on the timer, as the switch did before; `0`: every short sleep yields. The log's `sleep:` line says which (`host/short_sleep.h`, `host/hos_thread.c` `sleep_for`) |
|
||||
| `HOS_SHORT_SLEEP_POLLERS=<n>`, `=auto` | (Triforce, 11-22, the performance audit's OS-33, after Source A's; off unless set: it changes how the game's threads wait and was not measured here) A cap on the game threads that yield through a sleep under a millisecond at once. Where those sleeps yield (more than 6 logical processors, `HOS_SHORT_SLEEP_TIMER` not set) every idle job worker keeps a processor busy; with the cap, a game thread that would be one more than `<n>` yielding waits on the high-resolution timer for that sleep instead, about 0.5 ms late. `auto`: a quarter of the process's logical processors (2 with 8, 4 with 16, 6 with 24), Source A's rule, whose provisional numbers are 2.7% less CPU a frame at the same frame rate with 16 processors, and 1.7% of the frame rate lost in a 24-thread run held to 2. The game's MainThread and Presentation Thread and the host's own sleeps are never counted and always yield. Not set, `0`, `off`: no limit, as before. The start's `sleep: at most` line says the cap, and with `HOS_SLEEP_LOG=1` a `sleep:` line every 5 s counts the sleeps that took the timer for it (`host/short_sleep.h` `short_sleep_pollers`, `short_sleep_poll_enter`, `host/hos_thread.c` `sleep_for`) |
|
||||
| `HOS_OS_YIELD=0` | (Triforce, 11-22, the performance audit's THREAD-41, after Source A's; off unless set: it changes how the game's threads wait and was not measured here) The game's `YieldThread` hands its virtual core to a waiting game thread of equal or better priority and then also gives the processor up (`SwitchToThread`, a call into the kernel, thousands of times a second from the game's job system). `0`: only the virtual core's hand-over, the console's yield; Windows still takes the processor at the end of a time slice. Source A, the game held to 4 processors: half of its six executors' CPU time was inside that call, and without it the mean frame was 0.26 to 0.34 ms shorter; with 16 processors no difference. The risk is a thread yielding in a loop for one that waits on its processor, on PCs with 2 to 4 logical processors. Not set or `1`: both, as before. With `HOS_VCORES=0` the processor's yield stays whatever this says; `SleepThread(0)` and the short sleeps keep theirs (`host/hos_thread.c` `os_yield`, `YieldThread`) |
|
||||
| `HOS_VCORE_SPIN_US=<us>`, `HOS_VCORE_SPIN_MAX=<n>` | (Triforce, on by default; `HOS_VCORE_SPIN_US=0` as before) A game thread that waits for a virtual core it asked a holder to give up spins for it, up to 250 us, before it sleeps (`host/hos_thread.c` `vc_spin`, after NX2navi's `core_take`): the holder gives the core up at its next checkpoint, mostly within microseconds, and the waiter then runs without an OS wake. Who gets a core does not change, only how soon. At most `HOS_VCORE_SPIN_MAX` threads spin at once and the rest sleep at once: by default a quarter of the logical processors beyond 8, at most 4 (none with 8 or fewer, 1 with 12, 2 with 16). The `vcores:` line (every 5 s) counts the waits that ended in the spin and those that slept, the time spun, the spins left out for the cap, and how soon a granted core ran after a spin and after a sleep; the start's `vcores: a waiter spins up to` line says what is in effect. A yield with no thread waiting for a core takes no lock (NX2navi's yield fast path) |
|
||||
| `HOS_VIDEO_ACQUIRE_TRIES=1`, `HOS_VIDEO_KEEP_CORE=1` | (Triforce, on by default; each `=1` as before) A movie's Acquire (`host/nn_video.c` `dec_acquire`) waits for a picture up to 100 ms by the clock, and not at all after a Stop or Flush with nothing submitted since: the old wait, 50 sleeps of 2 ms that each took 4-5 ms on Windows, ran out only after ~250 ms once a movie had stopped, a frame that long at every load from the title. It also decodes without the calling thread's virtual core, taken back only after the decoder's lock is let go, so other game threads run meanwhile. `HOS_VIDEO_ACQUIRE_TRIES=1`: the 50 sleeps; `HOS_VIDEO_KEEP_CORE=1`: the core held throughout |
|
||||
| `HOS_VCORE_WIDE=1` | (Triforce, 11-22, the performance audit's GAME-22, after Source B's; off unless set, never measured there or here) A game thread whose core mask names two or more cores takes any free virtual core, not only those. Source B's reason: the game's two main threads may run on two of three cores only, where per-core job workers of better priority preempt them while the other cores stand idle. A thread with a one-core mask, which guards the game's lock-free per-core data, or with no mask, stays where it was; no two threads ever run under one core number either way. The risk is the one `HOS_VCORE_ANY=1` (every thread on any core, which crashed) has, for fewer threads: a thread that reads its core number, blocks and goes on under another. The log's `vcores: ... asks for ideal core ... and core mask` lines name every thread whose mask has two or more cores (all masks when the switch is set to anything): with no such line the switch changes nothing (`host/hos_thread.c` `vc_allowed`, `core_mask_note`) |
|
||||
| `RX_VK_HELPER_SPIN_US=<us>` | (Triforce, on by default; `=0` as before) The parallel replay's spins (`gpu/vk/helper_spin.h`, after NX2navi's replay workers): a helper thread with no segment to record spins up to 100 us for the next before it sleeps, and the replay thread spins as long for a segment a helper records before it sleeps for it. The frame statistics' `its spins` line, after the parallel replay's parts, says how often each ended in the spin or slept and the time spun |
|
||||
| `HOS_X3D=0` | (Windows) The game's threads and the replay's core are not kept on the die with the largest L3 cache. By default they are (the replay's core only where one is kept there: `HOS_RESERVE_CORE`), on a processor whose L3 caches differ in size and whose largest serves 8 or more cores (AMD's two-die X3D parts: a 7950X3D's or 9950X3D's 3D V-Cache die; `host/hos_x3d.h`), and the log's `cpu: logical processors ... share the largest L3 cache` line says so. On every other processor nothing changes and the log has no such line |
|
||||
| `HOS_X3D=0` | (Windows) The game's threads and the replay's core are not kept on the die with the largest L3 cache. By default they are, on a processor whose L3 caches differ in size and whose largest serves 8 or more cores (AMD's two-die X3D parts: a 7950X3D's or 9950X3D's 3D V-Cache die; `host/hos_x3d.h`), and the log's `cpu: logical processors ... share the largest L3 cache` line says so. On every other processor nothing changes and the log has no such line |
|
||||
| `HOS_SIM_CPU=<cores>/<threads>` | (Triforce, 11-22, the performance audit's OS-49, after Source B's; a lab switch, off unless set; Windows) A PC with fewer cores, simulated: the process is held to `<cores>` of the PC's physical cores (the die with the largest L3 first, then the fastest class, then the last ones, so that the replay's kept core is among them; one core always stays out; `host/hos_sim_cpu.h`), set before any thread, and the renderer sizes its pools and thread counts by `<threads>` (`gpu/vk/cpu_count.h`: an affinity mask alone does not change `hardware_concurrency`). `6/12` and `4/8` play as those PCs' pools do; `6` alone is 6/6. The log says `cpu: HOS_SIM_CPU=` with the mask, or that it was not applied (then the variable is cleared and nothing changes). Not run in the game. `tests/test_sim_cores.py` |
|
||||
| `RX_VK_SIM_GPU_SLOW=<factor>` | (Triforce, 11-22, the performance audit's OS-49, after Source B's; a lab switch, off unless set) A slower GPU, simulated (`gpu/vk/sim_gpu.h`): for each ms of work the recordings measure, (factor - 1) ms more, from over 1 to 10 (`2.5` is read). A one-invocation compute loop (`shaders/sim_gpu_burn.comp`) ends the replay's own recordings between full barriers, inside the GPU time they are measured over, so the frame's GPU time and Resolution Auto see the slower card; a third timestamp a recording measures what the burn took and the difference is carried (owed time held between -50 and 100 ms, burns of 1 to 16 ms, the loop's time calibrated 0.8/0.2). A line every 10 s: work, added, the factor reached, dropped. It repeats the fast card's relative answer only: no bandwidth, driver or upscaler cost. Not run in the game. `tests/test_sim_cores.py` |
|
||||
| `HOS_POWER_THROTTLING=1` | (Windows) Windows' power throttling of a window not in front left as it is (for A/B runs). By default every run, the shader preparation's too, opts out of it at the start: full speed, not the efficiency speed (EcoQoS) Windows 11 gives such a process, and the 1 ms timer resolution kept, which Windows 11 may otherwise ignore for it (Windows 10: full speed only). The log's `power:` line says what Windows took |
|
||||
| `HOS_HOST_ZSTD=0` or `2`, `HOS_HOST_ZSTD_AWAY=0/1/2` | The game's files (`.zs`) are decoded by the host's zstd (`third_party/zstd`) instead of the game's own translated decoder: the host decodes what `host/host_zstd.h` takes (whole buffers, and the loading threads' streams with their content size and room for it), the game's code the rest, errors included. (Triforce, 11-21, the performance audit's IO-01) On when not set: the host is given the content's exact size to write into (libzstd may park literals in spare room past a block, where the game's decoder never writes), and check mode ran clean in the game. `0`: the game's own decoder for everything, as before. `2`: check, the game's code decodes everything and the host beside it, and the log's `zstd check: ... MISMATCH` and `wrote past the end` lines name every difference (its `zstd:` lines count what was compared). `HOS_HOST_ZSTD_AWAY`: which decodes give their virtual core up: `1` (not set) whole buffers with 128 KB of input or room, streams never; `0` none; `2` whole buffers and stream feeds from 64 KB, as it was. Game build `e0f3ccfa` only. The lift needs both of its calls (`nx2/oot_patches.py` `patch_host_zstd`): a build log that says `the host's zstd decoder's calls were found at 1 of 2 places` has one only, and the switch then covers only that function |
|
||||
|
||||
### Building
|
||||
|
||||
@@ -165,46 +440,193 @@ A build reads and writes tens of gigabytes. When the game folder or the build fi
|
||||
| Switch | What it does |
|
||||
|---|---|
|
||||
| `HOOKSHOT_JOBS=<n>` (or `--jobs`) | Compilers at once (default one per CPU thread, each started when the free memory holds it; `HOOKSHOT_MEM_GATE=0` turns that check off) |
|
||||
| `HOOKSHOT_LIFT_JOBS=<n>` | Processes that write the lifted C (default the cores, at most 16) |
|
||||
| `HOOKSHOT_LOW_PRIORITY=1` | (Triforce, 11-22, the performance audit's OS-57) (Windows) The builders start the build tools (`make_game.py` and everything it runs) below normal priority, as every build was before. The default is normal priority: Source A reports, without a measurement, that below normal Windows kept the steps that are one process (the read, the lift) on a hybrid Intel CPU's efficiency cores while the performance cores idled, and a first build took over an hour. The compilers then compete with the desktop like any program (`HOOKSHOT_JOBS` bounds them). Read by Triforce Setup (`setup/Setup.cs`), the launcher's build (`launcher/Logic/Build/BuildRunner.cs`) and the two older builders (`builder/HookshotBuilder/Runner.cs`, `builder/py/hookshot_builder/core.py` `build_priority`); Linux sets no nice value either way. The shaders' preparation at a build keeps its workers below normal in the normal-class process (`gpu/vk/vk_draw.cpp` `prepare_priority`) |
|
||||
| `HOOKSHOT_LIFT_JOBS=<n>` | Processes that write the lifted C. Forked on Linux and macOS: default the cores, at most 16, one per 2 GB of free memory. Spawned on Windows (`nx2/emit_pool.py`), where they also cut the big chunks and share the page marks: default half the processors, at most 8, one per GB of free memory. `1`: one process, on every system; the C is the same whatever the count |
|
||||
| `HOOKSHOT_KEEP_BUILD_FILES=1` | Keeps the lifted C and the objects after a build, so that the next one compiles only what changed |
|
||||
| `HOOKSHOT_LIFTED_C_REGEX=1` | (Triforce, 11-22, the performance audit's IO-40) The lifted functions' definitions are found by the patterns' own `finditer`, tried at every line of the lifted C, as before. By default `nx2/lifted_c.py`'s `definitions` and `noted_starts` jump from one `"\nvoid f_"` to the next and try the pattern only there, for `build.py`'s memory estimate, `chunks.py`'s costs and cuts and `split_hot.py`'s pieces: the same matches (`tests/test_lifted_c_scan.py`), so the same files. `lifted_c.py`, `chunks.py` and `split_hot.py` are in the lift key: the first build with this change translates the game once more. The other half of the entry, translating while the package's romfs is still being extracted, is not here |
|
||||
| `HOOKSHOT_KEEP_LIFT=0` | (Triforce, 11-22, the performance audit's LIFT-27) The translated game is kept as little as before: its key (what decides whether a build translates the game again) hashes every `A64_`, `OOT_` and `RX_` variable and the folders' paths, and `nx2/released_translation_stamp.json`, and any failed build forgets it. By default the key takes of the environment only what shapes the C, the lifter's `A64_` switches and `RX_TARGET` (`tests/test_lift_key_env.py` holds the lift's files to that), and leaves out the release's shader translation stamp, which `build.py` alone reads; and a build that failed in its compile, only in the program's own files (`host/`, the renderers) and in nothing the lift wrote, keeps the key once, so that the retry compiles the rest without translating the game again (a second such failure in a row forgets it, as a generated header cut short fails the program's own files too). The key changes once with this, and the two values of the switch have different keys. Not here yet: keeping the translated game after a player's finished build, which is the owner's decision (`nx2/make_game.py` `lift_key`, `keep_lift_after_failure`) |
|
||||
| `HOOKSHOT_EARLY_COMPILE=0` | (Windows) Not compiled beside the lift: libopus and the two renderers, which need no translated code. By default `make_game.py` starts `build.py` with `RX_EARLY=1` for them while the game is translated (the lift's own processes take at most half the processors there), and the build proper finds their objects made |
|
||||
| `HOOKSHOT_MEM_GIANT=1` or `0` | Whether the memory gate counts a function of 1.2 MB or more by the square of its size (a chunk that is one 5 MB function takes 6.2 GB, not the 1.9 GB it is counted at otherwise). By default only on a PC with less than 14 GB free when the compile starts: with more, counting the giants in full halves the compilers running at once for no need |
|
||||
| `HOOKSHOT_MEM_LOG=1` | For each file compiled, `@@mem <file> <peak MB> <counted MB> <seconds>`: the most memory its compiler was seen holding, beside what the memory gate counted it at (`mem_need` in `nx2/build.py`). For fitting the gate after a change to the lift or the compiler |
|
||||
| `HOOKSHOT_PROMPTS_FOLDER_KEY=0` | (Triforce, 11-21, the performance audit's IO-43) A game built from romfs and exefs folders has its button prompts drawn at every build and update, as before. The default keeps them when nothing they are made from changed: with no unpack stamp to tell the dump by, the key is the content of the game's files the tool reads (the font, the textures, the layouts, the camera's pack, the size table, the dictionaries and every language's messages). Every game's key also holds `host/prompts.c` and the tools' release (`nx2/make_game.py` `prompts_key`, `prompt_game_files`) |
|
||||
| `HOOKSHOT_PROMPTS_WIDEN_EACH=1` | (Triforce, 11-22, the performance audit's IO-42) The button prompts' tool widens the "lines that stay" image (a 9 by 9 maximum, what a leader line pays to cross them) at every route, as before. By default the widened array is kept for the last image asked, and every route of one line is given the same image: 4 or 5 widenings a controller face where there were 35 or 37, each a filter over a 2-megapixel image (the five faces' lines on blank stand-in images: 21 s where they took 42). The lines drawn are the same (`nx2/prompt_assets.py` `guide_widened`) |
|
||||
| `HOOKSHOT_DUMP_LINKS_LSTAT=1` | (Triforce, 11-21, the performance audit's IO-41) A build from romfs and exefs folders checks them for links with an `os.lstat` for every entry, as before. The default reads each entry's kind and attributes from the folder's own listing (`os.scandir`), 1.3 to 1.5 s where the one-by-one check took 36 to 49 s over the game's 186,583 files, and stops the build at a folder it cannot list, which the walk passed over. The copy that follows makes each folder once and lets `os.link` say that a file is there already (`nx2/make_game.py` `refuse_dump_links`, `link_or_copy`) |
|
||||
| `HOOKSHOT_NUP_IN_PLACE=0` | (Triforce, 11-22, the performance audit's IO-51) A `.nup` package's unpack writes each file to `<name>.part` and renames it, and resolves the output folder and looks at every folder above the file for every file, as before. The default writes each file straight to its name (one `os.lstat`, then a create that fails if anything is there) and looks at each folder once: under a virus scanner on NTFS the temporary's create and rename and the repeated looks were most of a small file's cost over the game's 186,583 files. A file that is there whole is kept, one of another size is unlinked and written again, one whose write fails is deleted; the files written are the same (`nx2/nup.py` `extract`, `Destination`) |
|
||||
| `HOOKSHOT_FINISH_COPY_ALL=1` | (Triforce, 11-22, the performance audit's IO-45) The end of a build copies every file into the game folder as `<name>.new` and swaps it in, as before, and keeps no note. By default the finish notes in the game folder (`.triforce_installed.json`), for each file it installed, the SHA-1 of the source and the installed copy's size, modification time, file number and mode, and the next finish passes over a file whose source hashes the same and whose installed copy is still as noted, without reading it: `image.bin` and the prebuilt DLLs at an update that did not change them. The exe and the renderer are linked again at every build with the link's time in them, so they are still copied. A file not noted, changed by hand, on a file system without file numbers (FAT, exFAT), or under a note that does not read is copied; deleting the note copies everything (`nx2/make_game.py` `INSTALLED_NOTE`, `installed_same`) |
|
||||
| `HOOKSHOT_LINK_REPRO=1` | (Triforce, 11-22, the performance audit's IO-45) (Windows) The game's exe and the renderer are linked with `/Brepro`: lld writes a hash of the file where it otherwise writes the time of the link, so two links of the same objects are the same bytes, and the finish above passes an unchanged exe over instead of copying its 460 MB. Off unless set: it changes the bytes of every linked file once (`nx2/build.py` `repro_link`) |
|
||||
| `HOOKSHOT_MINGW=<dir>`, `HOOKSHOT_MSVC=1` | The llvm-mingw to build with; the Visual Studio route instead |
|
||||
| `HOOKSHOT_WORK_LOCK=0` | (Triforce, 11-22, the performance audit's IO-47) `make_game.py` takes no lock on its work folder, as before. By default it holds an OS lock on `<work>/make_game.lock` for as long as it runs (`msvcrt.locking` on Windows, `flock` elsewhere), taken after an earlier build left running is stopped and before the stale objects go: a second build on the same folder stops at once with the first one's process id instead of compiling the same objects again and racing on the staged game folder, which nothing prevented on Linux (two builders, a builder beside the launcher, `--cli`). The lock ends with its process, so a build that crashed never blocks the next; an error other than "held by another" is said and the build runs without the lock (`nx2/make_game.py` `lock_work`) |
|
||||
| `HOOKSHOT_SINGLE_INSTANCE=0` | (Triforce) Hookshot Builder takes no lock on its build tools, as before. By default a second builder for the same tools folder (`builder/HookshotBuilder/Instance.cs`, named after `Tools.Root`), its `--cli` included, waits up to 5 s for one that is ending, then says so and ends: two of them built into one work folder and stopped each other's build. Triforce Setup has its own lock (`setup/Guard.cs` `Instance`) |
|
||||
| `TRIFORCE_LEGACY_PATHS=1` | (Triforce) Triforce Setup keeps .NET Framework's legacy path checks, which refuse paths of 260 characters and more whatever its manifest says (an exe csc builds is taken as written for .NET 4.0). By default `setup/Setup.cs` `Main` sets the two switches 4.6.2 defaults to (`Switch.System.IO.UseLegacyPathHandling`, `Switch.System.IO.BlockLongPaths`) before any path is used, so that with Windows' own LongPathsEnabled on Setup reads and writes deep paths |
|
||||
| `HOOKSHOT_SEGMENT_HEAP=0` | (Triforce, Windows; on by default since 11-22: the lab's play A/B, 81.4 against 80.6 fps) The exe's manifest asks for Windows' segment heap (`heapType SegmentHeap`, Windows 10 2004 and newer) for the whole process, its DLLs and the graphics driver too, in place of the NT heap, whose lock AMD's driver's compile threads waited on through the shader preparation (`nx2/win_manifest.py`, `tests/test_segment_heap.py`). The heap is fixed when the process starts: a build with it and one without are the two arms. `0`: the NT heap, as before. The log's `the shaders' preparation runs on` line names the process heap (`segment` or `NT`) |
|
||||
| `RX_CPM=<dir>`, `WGPU_NATIVE=<dir>` | The CPM cache (Vulkan and SPIR-V headers, VMA, FFmpeg) and wgpu-native |
|
||||
| `RX_TARGET=windows` | On Linux: the Windows game, with llvm-mingw's Linux build in `HOOKSHOT_MINGW` |
|
||||
| `RX_RELEASE=1` | A release build (`make_game.py` sets it): no console window, the shaders prepared before the title |
|
||||
| `RX_VK_PGO=gen`, `=<file>`, `=0`; `RX_VK_PGO_FILE=<path>` | (Triforce, 11-21, the performance audit's LIFT-10) A play profile for the renderer's own code (`gpu/vk/*.cpp`; not the vendored recompiler, ImGui or zstd). `gen`: a build whose renderer counts its branches; run it with `RX_VK_PGO_FILE=<path>`, make `<path>.reset` once the scene to record is loaded and `<path>.dump` at its end, then `llvm-profdata merge -o work\triforce_vk.profdata <path>`. `<file>`: built with that profile (its contents are in the object keys, its path is not). Not set: `make_game.py` passes `work\triforce_vk.profdata` when the tools have one and the compiler can read it; `0`: none. A profile from another version of the renderer is taken function by function, the changed ones built without it. Record it again for each release, and compare against `RX_VK_PGO=0` before shipping one (`nx2/build.py` `vk_pgo_flags`, `host/pgo_rt.c`) |
|
||||
| `RX_PGO=gen` or `RX_PGO=<file>` | The hot code counting its branches for a profile, or built with one (`make_game.py` passes the shipped profile) |
|
||||
| `A64_HOT_STICKY=0` | (Triforce, 11-22, the performance audit's LIFT-30, after Source A's; read when the game is translated, and in the lift key) `nx2/split_hot.py` deals the hot functions out afresh and gives every hot piece code_hot.c's whole preamble, as before. By default `<work>\image\hot_pieces.txt` keeps each hot function's piece: the next translation leaves the functions it knows where they were and puts a new one in the piece it shares the most calls with, else in the cheapest, unless that makes the costliest piece more than 1.15 times a fresh split's (then the split is the fresh one); and each piece keeps only the declarations of `f_` and `hos_` names it uses. An update that changes a few hot functions then compiles one or two hot pieces of about ten MB again, not up to eight. The functions' text is the same either way, and the hot profile goes by function name, so it fits both. The file is not in the lift key (`tests/test_sticky_hot_pieces.py`) |
|
||||
| `RX_COLD_OPT=O2` or `=<file>`, `RX_COLD_PART_MB=<n>` | (Triforce, 11-22, the performance audit's LIFT-17, after Source A's; opt-in) The hot pieces are compiled at -O2 with the play profile and every other lifted chunk at -O1. `RX_COLD_OPT=O2`: every other lifted chunk at -O2 too; `=<file>`: only the chunks the file lists, one file name a line (`code_012.c`). Unset, empty or `O1`: as before. Each -O2 chunk is compiled as parts of `RX_COLD_PART_MB` MB by the lift's own estimate (6 when unset; `0`: no cut), cut only where a function begins and written under `<build>/split` with a key file a chunk, so a second build writes nothing (`nx2/cold_parts.py`, `tests/test_chunk_parts.py`): -O2's time grows with the square of a function's size. The lifted files themselves are not touched. Source A's figures: +4.4% and +5.5% fps where the game's update bounds the frame, for about twice the lifted compile's processor time (23 minutes against 5 without the cut); not built or measured with this game. The names begin with `RX_`, so setting them translates the game again once |
|
||||
| `RX_MARCH=<level>` | The CPU level to build for (default this PC's: `x86-64-v3`, else `x86-64-v2`, else `x86-64`) |
|
||||
| `RX_VK_ONLY=1` | Only the Vulkan back end, with no lift |
|
||||
| `OOT_WORK`, `OOT_DUMP`, `OOT_EXEFS`, `OOT_ROMFS` | The work folder and the dump's folders, over `work/dump.txt` |
|
||||
| `HOS_MODS=<folder>` | The mods folder, in place of `mods` in the game folder: for the running game (RomFS mods, cheat files) and for a build (the code patches it applies) |
|
||||
|
||||
## Mods: code patches and cheats
|
||||
|
||||
The game loads RomFS mods at its start (`host/mods.c`). The two kinds of mod that change what the game *does* go other
|
||||
ways, because the game's code is C compiled ahead of time and cannot be patched while it runs.
|
||||
|
||||
**The mods folder**, for the game and for a build alike (`host/mods.c` `find_mods`, `nx2/code_mods.py` `find_mods`,
|
||||
which must stay the same walk): a mod is a folder under `mods` that holds a `romfs`, `exefs` or `cheats` folder, at
|
||||
most four levels down, and nothing below a mod is looked at. A folder whose name starts with `-` or `_` is off, one
|
||||
that starts with `.` is skipped. Mods are ordered by their path under `mods`, ASCII letters' case ignored, and the last
|
||||
one wins. Triforce Setup's mod manager can switch a mod by renaming its folder and order mods by a prefix, and needs
|
||||
no other file.
|
||||
|
||||
**Code patches** (`<mod>/exefs/*.ips`, `*.pchtxt`), `nx2/code_mods.py`:
|
||||
|
||||
- `make_game.py` calls `build_step` once its copy of the game's code (`<work>/exefs`) is there and checked, after the
|
||||
copy the game folder keeps for updates (`.triforce_exefs`) is made and before the lift key is taken. The patches go
|
||||
into that copy, a module with a patch written again as a plain NSO, and the lift reads it: `lift.py`, `chunks.py`
|
||||
and the rest of the translator know nothing of mods.
|
||||
- The lift key hashes the files of that copy, so the patched bytes are in it: a patch added, taken out, switched off
|
||||
or edited translates the game again, a patch that changes no byte does not, and neither does a change to
|
||||
`code_mods.py` itself (`LIFT_KEY_NOT`).
|
||||
- A patch names its module by build id (the `.ips` file's name, the `.pchtxt` file's `@nsobid-`, trailing zeros left
|
||||
out or not) and may be for any module of the exefs. Offsets count the NSO's 0x100-byte header, as Atmosphere and the
|
||||
emulators read them; `.pchtxt` knows `@flag offset_shift`, `@enabled`, `@disabled`, `@little-endian`, `@big-endian`
|
||||
and `@stop`. Mods in order, a mod's files by name, the later write wins and the build says which file wrote over
|
||||
which.
|
||||
- A file is refused whole, with a line that says why: another build's, cut short, a value that cannot be read, a write
|
||||
into the header, past the image or between the segments. The last leaves out code caves, which the translator has
|
||||
not been tried on.
|
||||
- The build writes `triforce_code_mods.txt` into the game folder when it is done: a line per patch file, tab
|
||||
separated, `applied` or `refused`, the file's FNV-1a 64 and size, its path under the mods folder, then the module
|
||||
and the bytes changed, or the reason. The next build (the game's update offer runs `--update`) reads the mods folder
|
||||
again, so the same patches are built in again, and says which of the record's are gone, changed or refused now. The
|
||||
game reads the record at its start (`host/mod_code.h`) and logs for each patch file whether it is built in or waits
|
||||
for a build, and which built-in patches are gone from the folder.
|
||||
|
||||
**Cheats** (`<mod>/cheats/<first 16 digits of main's build id>.txt`), `host/mod_cheats.h` and `host/mod_cheats.c`:
|
||||
|
||||
- Read at the start, run by an interpreter once a game frame, at the game's present and on its thread
|
||||
(`hos_mod_cheats_frame`). Nothing is compiled in, so a cheat file needs no build. NX2navi compiled in the cheats that
|
||||
write instructions; Triforce's translator has no switchable patches, so such a cheat (an opcode 0 into a module's
|
||||
`.text`) is refused and belongs in a code patch.
|
||||
- The opcodes it runs are listed at the top of `host/mod_cheats.h`: 0 to A and C0 to C3 of Atmosphere's cheat machine.
|
||||
Any other is refused when the file is read, the whole cheat, by name. Registers are shared by one file's cheats within
|
||||
a frame and cleared between frames; a master code runs first.
|
||||
- Memory is the program image and the heap, each access checked (`mod_cheat_readable`); the modules' code
|
||||
(`hos_code` in the generated `hos_image.c`) is never written. Region 0 is the module main's base (`hos_main_base`),
|
||||
1 and 2 `HOS_HEAP_BASE`, 3 `0x08000000`, as `host/hos_svc.c` tells the game.
|
||||
- Each cheat that is not a heading has a row on the System tab, 96 at most (`gpu/vk/mod_cheat_rows.h`; option ids 130
|
||||
to 133 in `gpu/vk/vk_api.h`). The rows are not in the options file: `triforce_cheats.txt` in the game folder has a
|
||||
line for each cheat that is on, its number (FNV-1a 32 of `<mod>/<name>`) and that text, and a cheat listed there is
|
||||
on at the start.
|
||||
|
||||
## The options file
|
||||
|
||||
`hookshot_options.ini`, in the game folder, holds the Insert menu's rows as `<key>=<number>`: the row's name in lower
|
||||
case with `_` for spaces (`frame_generation=2`, `invert_camera_x=1`), except Output resolution, which is `output_res`.
|
||||
`gpu/vk/vk_menu.cpp` writes the whole file at every change (`save`), with `options_version=4`,
|
||||
`gpu/vk/vk_menu.cpp` writes the whole file at every change (`save`; (Triforce) merged by key with a change made outside the game meanwhile, and written whole: `RX_VK_OPTIONS_MERGE` above), with `options_version=8`,
|
||||
`microphone_device=<name>` when a microphone is chosen by name, and the controls (`pad_<button>` and `key_<button>`:
|
||||
the README's [Controls](../readme.md#controls)). A value outside a row's range is clamped when it is read. Where the
|
||||
number is not the obvious one:
|
||||
|
||||
| Key | Values |
|
||||
|---|---|
|
||||
| `quality_preset` | 0 Custom, 1 Minimum (4GB), 2 Low (6GB), 3 Balanced (8GB), 4 High (12GB), 5 Ultra (16GB+) |
|
||||
| `gpu_device` | 0 Auto, n the n-th card with Vulkan 1.3 as the row lists them |
|
||||
| `frame_generation` | 0 Off, 2 to 8 the multiplier |
|
||||
| `resolution` | 0 Auto, 1 100%, 2 75%, 3 67%, 4 50% |
|
||||
| `output_res` | 0 2048x1152, 1 2560x1440, 2 3840x2160, 3 1920x1080, 4 5120x1440, 5 5120x2880 |
|
||||
| `upscaling` | 0 Off, 1 Catmull-Rom, 2 Lanczos |
|
||||
| `frame_generation` | 0 Off, 1 Auto, 2 to 8 the multiplier, 9 AMD FSR frame generation (the made frame by AMD's frame generation library, FSR 4's or FSR 3.1's by the card, with the picture without the HUD; paused with HDR Native or RenoDX's RGBA16F scene, where the renderer's own makes the frame as at 2x: `gpu/vk/vk_fsr4.cpp` `fsr4fg`; where that cannot run the value stays in the file and the row shows and runs 2x: `gpu/vk/framegen_rows.h`) |
|
||||
| `dlss_frame_generation` | 0 Off, 1 2x, 2 3x, 3 4x: NVIDIA DLSS frame generation, the Frame generation row's NVIDIA DLSS values (`gpu/vk/merged_rows.h`; read at the start, `gpu/vk/vk_dlssg.cpp`). The row sets `frame_generation` to 0 with it |
|
||||
| `frame_rate_cap` | 0 60, 1 90, 2 120, 3 144, 4 165, 5 240, 6 the monitor's refresh rate; -1 45, -2 40, -3 30 |
|
||||
| `frame_pacing` | 0 Off, 1 frames on an even line. With `steady_frame_rate`, the Frame pacing row (`gpu/vk/merged_rows.h`): Off is 0 and 0, Even 1 and 0, Steady (any rate) 1 and 1, Steady (refresh steps) 1 and 2 |
|
||||
| `steady_frame_rate` | 0 Off, 1 any rate (the cap, then 200, 180, 160, 144, 120, 110, 100, 90, 80, 72, 66, 60, 55, 50, 45, 40, 36, 30 under it), 2 whole divisions of the display's refresh rate under the cap, from 30 up (`host/pace_steady.h`) |
|
||||
| `resolution` | 0 Auto, 1 100%, 2 75%, 3 67%, 4 50%, 5 58%, 6 33% (with DLSS and FSR: 3 Quality, 5 Balanced, 4 Performance, 6 Ultra Performance; their Native (100%) is `dlss` 2, 5 or 6) |
|
||||
| `output_res` | 0 2048x1152, 1 2560x1440, 2 3840x2160, 3 1920x1080, 4 5120x1440, 5 5120x2880, 6 5120x2160, 7 1536x864 (the handheld profile; before `options_version` 8 it was `handheld_profile=1`, read as 7) |
|
||||
| `ultrawide` | 0 16:9, 1 16:10, 2 21:9, 3 32:9 (before `options_version` 6, 1 was the monitor's shape: read as the nearest of the four, `gpu/vk/ultrawide_rows.h`) |
|
||||
| `hdr`, `renodx_hdr` | the HDR row (`gpu/vk/merged_rows.h`): `hdr` 0 Off, 1 Expanded, 2 Native (the game's own); `renodx_hdr` 1 with `hdr` 2 is Native + RenoDX |
|
||||
| `nvidia_reflex`, `amd_anti-lag_2` | the Low latency row by the card's maker: 0 Off, 1 On, 2 On + Boost (NVIDIA alone) |
|
||||
| `hdr_paper_white`, `hdr_peak_brightness` | nits: 80 to 400, and 400 to 10,000 |
|
||||
| `renodx_tonemapper` | 0 Vanilla, 1 PsychoV35, 2 RenoDRT |
|
||||
| `renodx_gamma_correction` | 0 Off, 1 2.2, 2 BT.1886 (2.4) |
|
||||
| `renodx_match_tone_slope` | 0 Off, 1 On (as out of the box): PsychoV35's rows, with `renodx_tone_slope` and `renodx_hue_clip` |
|
||||
| `renodx_tone_slope`, `renodx_highlight_saturation` | 0 to 100, 50 unchanged: the first PsychoV35's, the second RenoDRT's |
|
||||
| `renodx_blowout`, `renodx_flare` | 0 to 100, 0 none: RenoDRT's |
|
||||
| `renodx_hue_clip`, `renodx_hue_correction`, `renodx_color_grading` | percent, 0 to 100: 50, 100 and 100 out of the box; the first PsychoV35's, the others RenoDRT's |
|
||||
| `dlss` | the Upscaler: 0 the game's own, 1 NVIDIA DLSS, 2 NVIDIA DLAA, 3 AMD FSR 3.1, 4 AMD FSR 4, 5 AMD FSR 4 Native AA, 6 AMD FSR 3.1 Native AA. The row names the engine; 2, 5 and 6 are its Native (100%) on the Render resolution row (`gpu/vk/upscaler_rows.h`) |
|
||||
| `anti-aliasing` | 0 the game's own SMAA, 1 Off, 2 FXAA, 3 SMAA Ultra |
|
||||
| `texture_detail` | 0 Automatic, 1 Full, 2 Reduced, 3 Low |
|
||||
| `fps_counter` | 0 Off, 1 On, 2 Full (NX2navi's full performance overlay in the counter's place: `vk_menu.cpp` `draw_perf`, the frame times and the figures per frame `gpu/vk/perf_stats.h`, the machine's numbers sampled twice a second on a thread of their own `gpu/vk/perf_machine.cpp`) |
|
||||
| `opaque_textures_as_bc1` | 0 Off, 1 On |
|
||||
| `sharpness`, `rumble_strength`, `motion_sensitivity_x`, `motion_sensitivity_y` | percent |
|
||||
| `mouse_sensitivity` | hundredths: 100 is 1.00 |
|
||||
| `left_stick_dead_zone`, `right_stick_dead_zone` | -1 Default (XInput's ~24%), 0 to 50 percent (before `options_version` 8 a `stick_dead_zone` of 1 to 5, 5, 10, 15, 20 or 30% for both sticks, is read into each stick's own where it had none) |
|
||||
| `vibration_curve` | hundredths, 20 to 150: the power on the motors' rumble, 100 is 1.00 (none) |
|
||||
| `walk_key` | 0 Off, 1 Caps Lock, 2 Shift, 3 Ctrl, 4 Tab |
|
||||
| `walk_scale` | hundredths, 20 to 90: 51 is 0.51 of the stick's travel |
|
||||
| `robust_buffers` | 0 Auto (on), 1 Off, 2 On |
|
||||
| `game_speed` | 0 Per frame, 1 Real time |
|
||||
| `text_speed` | 0 Normal, 1 2x, 2 4x, 3 Instant |
|
||||
| `menu_sounds` | percent, 0 to 100: 80 unless set, 0 silent. The settings' own sounds, the game's interface sounds from its RomFS (`host/menu_sound.c`, `gpu/vk/menu_sound_events.h`; option ids 135 and 136), under `master_volume` |
|
||||
| `microphone` | 0 Off, 1 Default, 2 and up a microphone (the name in `microphone_device` decides which) |
|
||||
| `mouse_camera` | 0 As the stick, 1 Native |
|
||||
| `layout_preset` | 0 Nintendo, 1 Xbox / PC, 2 Swap Face, 3 Custom |
|
||||
| `button_prompts` | 0 Auto (the controller there at the start, XInput's without a name the Xbox's; none: the device used last in the run before, `<user>/prompts_auto.txt`, else the keyboard and mouse), 1 Switch, 2 Xbox, 3 PlayStation 4, 4 PlayStation 5, 5 Steam Deck, 6 Keyboard & mouse (`host/prompts.c`, `nx2/prompt_keys.py`) |
|
||||
| `replay_helper_threads` | 0 Off, 1 to 7 the parallel replay tried on 2 to 8 threads, 8 Automatic (the default; the Debug tab). A file that says 0 stays Off: delete the line, or choose Automatic on the Debug tab, for the default |
|
||||
|
||||
The Game language row is not in this file: its choice is the code line of `language.txt` in the game folder, which the
|
||||
menu rewrites in place (`gpu/vk/game_language.h`: the code alone on that line, the rest of the file as it was, and only a
|
||||
code the file itself names) and the host reads at the start (`host/nn_applet.c`). Nor are the Mod cheats rows: which
|
||||
of a mod's cheats are on is the host's to keep, in `triforce_cheats.txt` (*Mods: code patches and cheats* above).
|
||||
|
||||
(Triforce) `options_version=8` grouped the rows and merged those that chose one thing twice (`gpu/vk/merged_rows.h`; the
|
||||
436f9ac-based builds wrote 7 for it, which v0.5.1 had already given to Frame pacing's default: its 7 still has the old rows'
|
||||
lines, so a file before 8 is moved, and one of those builds' 7, without them, keeps what it has). A file
|
||||
from before keeps its values; these lines are read once and not written again, their rows' jobs done otherwise:
|
||||
`quality_preset` (the textures' and the mirrors' memory budgets are the card's own, from its video memory), `upscaling`
|
||||
(the window's filter follows the sizes: `gpu/vk/upscaler_rows.h` `auto_filter`, AMD FSR 1 when the picture is enlarged, the
|
||||
plain copy otherwise and always with HDR Native), `reflex_frame_limit` (off, the old default, until Reflex holding the cap
|
||||
with DLSS frame generation has been played: `RX_VK_REFLEX_LIMIT=1` for a run), `handheld_profile` (an `output_res`) and `stick_dead_zone` (each
|
||||
stick's own).
|
||||
|
||||
The other rows are 0 Off and 1 On. Some are read at the start, before the menu exists, by the code that needs them:
|
||||
`handheld_profile` and `output_res` (`host/nn_applet.c`), `ultrawide` (`host/hos_main.c`), `robust_buffers`
|
||||
(`gpu/vk/vk_main.cpp`) and the controls (`host/nn_hid.c`).
|
||||
`output_res` (`host/nn_applet.c`; 7 is the handheld profile), `ultrawide` (`host/hos_main.c`), `frame_rate_cap`
|
||||
(`host/nn_vi.c`; the monitor's rate is the display's the window opens on, rounded, 30 to 240, and 60 when it cannot be read), `robust_buffers`
|
||||
(`gpu/vk/vk_main.cpp`), `texture_detail` (`gpu/vk/vk_texture.cpp`, at the first texture) and the controls (`host/nn_hid.c`).
|
||||
|
||||
The `renodx_` rows other than `renodx_hdr` are read at every frame (`gpu/vk/vk_renodx.cpp` `fill_settings`), and the
|
||||
menu shows the rows of one tonemapper while it is the one chosen. RenoDX's mod for this game has Vanilla and PsychoV35;
|
||||
RenoDRT and the gamma correction come from `gpu/vk/shaders/renodx_glue`, three files of Triforce's own that
|
||||
`scripts/renodx_shaders.py` compiles together with RenoDX's scene shaders and its output encoder as the build fetched them
|
||||
(nothing of RenoDX is in the repository or changed). RenoDRT calls RenoDX's library; the gamma correction is on the
|
||||
finished picture, HUD included, and the tone mappers are given a peak with it taken out, so that the brightest light
|
||||
still lands on the HDR peak brightness. A change to a glue file changes the stamp of `renodx_spv.h`, which the next build
|
||||
then compiles again.
|
||||
|
||||
RivaTuner Statistics Server has no row: where it is installed and running it is the game's frame limiter, and an options
|
||||
file that still has the test builds' `rivatuner_limit=` is read by nothing (the next save drops the line). A thread of
|
||||
`gpu/vk/rtss_limit.cpp` looks twice a second: it puts the limit (`gpu/vk/pace_rule.h` `pace_rtss_want`: the cap in
|
||||
effect, Monitor's rate and the caps under 60 included, times the frames shown for each of the game's with frame
|
||||
generation 2x to 8x or AMD FSR 4) into RTSS's profile for the executable through RTSS's own hooks library
|
||||
(`RTSSHooks64.dll` from RTSS's folder: `FramerateLimit` and `FramerateLimitDenominator`, nothing else, no other profile),
|
||||
reads it back, and once RTSS is seen holding the game to it (`gpu/vk/rtss_decide.h` `rtss_decide`) the host's wait steps
|
||||
aside (`gpu/vk/vk_pace.cpp`), the in-house frame generation's display thread leaves its own spacing of made frames to
|
||||
RTSS, and Reflex is given no frame limit. A limit lower than the profile's waits 5 s before it is written, a higher one
|
||||
1 s, so a multiplier that changes in play does not rewrite the profile each time. What the profile held is kept in
|
||||
`triforce_rtss_profile.txt` in the game folder before the first change (one line per executable, the others' lines
|
||||
kept, a file that does not read left alone) and goes back at the exit; the next start puts it back after a kill or a
|
||||
crash, unless the player changed the profile since. The game's own wait holds the frames instead with frame generation
|
||||
on Auto (its presents follow the screen's refresh, not the game, so no limit on them holds the game) and with DLSS frame
|
||||
generation (Streamline presents on a pace of its own: `HOS_RTSS_DLSSG=1` tries RTSS there), during movies (the profile
|
||||
keeps the play limit), while no picture went out for 200 ms (a minimised window), while RTSS is not installed, not
|
||||
running or has not hooked the game, and when the profile does not keep the limit (no administrator rights) or RTSS
|
||||
refuses it, unless a limit the player set in RTSS, at or under the game's, holds the game (then RTSS holds movies too).
|
||||
A hold needs a present since the frame start two frames back, so a window that stops showing gives the frames back
|
||||
within two frames. RTSS's count past 1.05 times the limit three
|
||||
periods running, 30 s hooked with the limit and no hold, or the game's frames coming sooner than the cap give the frames
|
||||
back; RTSS is tried again a minute later (three times), and at once when RTSS starts again or another limit is wanted.
|
||||
The game never starts RTSS, installs nothing and asks for no rights. `vk: pace:` lines in the log say which limiter is in
|
||||
force and why; the Full overlay's Frame limit line names RivaTuner or NVIDIA Reflex after the rate when either holds the frames.
|
||||
|
||||
## Layout
|
||||
|
||||
@@ -216,9 +638,10 @@ The other rows are 0 Off and 1 On. Some are read at the start, before the menu e
|
||||
| `gpu/vk/` | the Vulkan back end (`hookshot_vk.dll`, `hookshot_vk.so`), with Eden's shader recompiler and sirit |
|
||||
| `gpu/render/`, `gpu/third_party/` | the WebGPU renderer the Windows exe falls back to when it was built with wgpu-native, its op stream (`rx_nvn.h`), suyu's decoders |
|
||||
| `builder/` | Hookshot Builder: the Windows app (.NET 8 WPF), its setup and MSI, and the Linux builder and launcher scripts |
|
||||
| `setup/` | Triforce Setup, the program players run (.NET Framework 4.8, WPF): `Setup.cs` installs, updates and runs `make_game.py`, `Shell*.cs` and `Main.xaml` are its window, `Guard.cs` is what keeps a run from damaging an install, `Mods.cs` the mods folder as the Mods page lists and changes it, `Steam.cs` and `SteamArt.cs` the Add to Steam shortcut and its pictures, `Sound.cs` the game's own decide and cancel sounds as Setup plays them (decoded in memory from the player's game folder: no sound is shipped; `TRIFORCE_SOUND_LOG=<file>` writes down, in a test run, the sounds that would have played) ([CONTRIBUTING.md](../CONTRIBUTING.md#triforce-setup)) |
|
||||
| `scripts/` | test runs, the profile and log readers, the save copier (`oot_saves.py`), release tooling, `scripts/linux` (dependencies, AppImage) |
|
||||
| `tests/` | the game-free checks, all run by `tests/run_all.py` |
|
||||
| `third_party/` | a64-kit (Hookshot's few changes marked `(hookshot)`), libopus, zbic (Atmosphere's zstd with ZBIC, for unpacking only), SDL2's licence, Dear ImGui and its Vulkan backend, zstd's decoder, the Libertinus fonts |
|
||||
| `third_party/` | a64-kit (Hookshot's few changes marked `(hookshot)`), libopus, zbic (Atmosphere's zstd with ZBIC, for unpacking only), SDL2's licence, Dear ImGui and its Vulkan backend, zstd's decoder, the Libertinus fonts, SMAA's reference shader and lookup textures (the Anti-aliasing row's SMAA Ultra) |
|
||||
|
||||
License: GPL-3.0-or-later, as mkw-recomp: a64-kit's translator is suyu's (GPL-2.0-or-later) and Eden's code is
|
||||
GPL-3.0.
|
||||
|
||||
@@ -29,3 +29,24 @@ It needs Vulkan headers but no GPU, SDK vendor runtime or game assets.
|
||||
Injected failures in these tests do not prove actual driver OOM recovery or
|
||||
device-loss handling. A successful game smoke validates integration, not a fault
|
||||
injection into the player's GPU.
|
||||
|
||||
## AMD FSR 4's frame generation
|
||||
|
||||
The Frame generation row's AMD FSR 4 (`vk_fsr4.cpp`, `fsr4fg`) holds one frame generation context of AMD's library and
|
||||
one image the made frame is written into. They follow the same rules:
|
||||
|
||||
- The image is made first, then the context; both are published together, only when both exist. A failure destroys
|
||||
what was made, turns AMD's frame generation off for the session (the row then shows what runs, 2x by the renderer's
|
||||
own interpolation) and is logged once. Nothing is retried per frame.
|
||||
- A context is replaced (another picture size, a scene larger than it was made for, another kind of depth range) before
|
||||
anything of the new frame is recorded: the queue and the device idle, AMD's context destroyed, then the image.
|
||||
- A context that failed while a frame was being recorded stays until the second present after it, since that frame's
|
||||
commands name it and are not on the queue yet; a context the row has left stays 120 presents. Both go at the
|
||||
present's tick (`framegen_present_tick`), in the same order.
|
||||
- At the process's exit (`HOS_OPT_EXITING`) nothing of AMD's is destroyed and nothing of it is called again: there is no
|
||||
swapchain or thread of AMD's to stop, and the context goes with the process, as the upscaler's does.
|
||||
|
||||
The order is one function, `framegen_rows.h` `teardown`, which the real code and `tests/test_framegen_fsr4.py` both
|
||||
run; the test also runs the real section against a stand-in for AMD's library and checks each of the cases above. It
|
||||
does not prove what AMD's library or a driver does with these calls: nobody has run them on a card with FSR 4's frame
|
||||
generation.
|
||||
|
||||
67
docs/launcher-color-scheme.md
Normal file
67
docs/launcher-color-scheme.md
Normal file
@@ -0,0 +1,67 @@
|
||||
# The launcher's color scheme
|
||||
|
||||
Settings > Launcher has a **Color scheme** row under Music and Sounds. Its value is a diamond swatch and the hex
|
||||
(`#E0B23E`). Selecting it (mouse, Enter, A) drops a picker open under the row, with the launcher's own .25 s height
|
||||
transition; selecting the row again, or Escape / B, closes it. The picker has a saturation / brightness square, a hue bar
|
||||
(one of the launcher's sliders), **Hex**, **R**, **G** and **B** fields checked as typed (six hex digits with or without
|
||||
`#`, whole numbers 0-255; a wrong value gets a red border and a line saying what is allowed), eight ready colors with
|
||||
Triforce gold first, and **Reset to gold**.
|
||||
|
||||
The color takes the place of the gold everywhere the launcher uses it: text and links, tab underlines, section titles,
|
||||
sliders and their diamonds, borders, hairlines, outlines, glows, the focus ring, the build line and the wordmark. The
|
||||
polished bars (Play, "Find my best settings", "Build the game", Apply, ...) keep their gradient, inner hairline, sheen and
|
||||
emboss, all made from the new color, and their text takes dark or light ink, whichever has more contrast. Triforce gold
|
||||
gives the look exactly as it was: with it the page sets nothing, and the stylesheets' own values draw it.
|
||||
|
||||
Two limits keep any color usable, both far from gold:
|
||||
|
||||
- **Text and lines never go too dark.** What is drawn in the color on the dark ground is lightened until it reads at
|
||||
4.5:1 or better on `#0a0908` (black draws them in `#787878`).
|
||||
- **Polished bars never go too light.** They are shaded no lighter than 78% lightness, so a near-white color gives a
|
||||
silver bar, not a flat white one.
|
||||
|
||||
## How it works
|
||||
|
||||
- **theme.css** holds the gold as a family of `r, g, b` variables (`--gold-rgb`, `--gold-hi-rgb`, ... and the bars'
|
||||
own `--gold-bar-*-rgb`), each with today's value; `v3.css`, `proto.css` and `app.css` use them in place of every gold
|
||||
literal, alphas kept.
|
||||
- **app.js** ("settings: Color scheme") moves each of today's shades from gold to the chosen color: the hue turns by as
|
||||
much, the saturation scales by as much, the lightness keeps its share of the way to white or to black. At gold it
|
||||
gives back every literal exactly. `paintColor()` sets the variables on `<html>` with the class `tint`, and removes them
|
||||
all for gold. The page adds the row itself after `sfx`; a row the launcher sends with `id: 'color'` takes its place.
|
||||
- **Pictures.** The crest (`img/crest.png`) is redrawn under `html.tint` with `hue-rotate / saturate / brightness`,
|
||||
masked by `img/crest-gold.png` (its metal only), so the three stones stay. The title ornaments and glows get the same
|
||||
filter chain from `--tint-h/-s/-b`. The game's own logo and its glow are the game's art and are not tinted. CSS
|
||||
`hue-rotate` is an approximation: far turns (blue, violet) come close but not exact.
|
||||
|
||||
## How it is kept
|
||||
|
||||
1. The page sends `{t:'cmd', c:'color', hex:'#rrggbb'}`, 400 ms after the last change (as `sound`), and keeps the same
|
||||
value in `localStorage['triforce.color']`, which it reads at start before the first screen.
|
||||
2. The launcher takes only a `#rrggbb` (`Logic/LauncherColor.cs`, `FromCommand` / `Parse`), in lower case. It takes it
|
||||
as the message arrives (`App/Backend.Core.cs`, `ColorArrived`, called from `OnMessage` before the command goes to a
|
||||
worker), into `LauncherColorChoice`, which holds the player's last color while the launcher runs. The command's
|
||||
worker (`On("color")`, `Keep`) then keeps the newest one as `Settings.Color`: `HKCU\Software\Triforce\Color` on
|
||||
Windows, `Color=` in `builder.ini` on Linux, `registry.txt` in a test run. Anything else is left out, with one log
|
||||
line a run.
|
||||
3. The launcher sends back `LauncherColorChoice.Current` (the player's last color, else the kept one) as `state.color`
|
||||
at start (`OnReady`, beside `sound`) and with the list the Settings page asks for each time it opens
|
||||
(`settings.list`). The page then uses it, and it wins over the page's own storage, unless a choice of the player's is
|
||||
still waiting to be sent. Because the launcher sends what it holds and not what it has written, a `--snapshot` run
|
||||
(which writes nothing) and a `settings.list` whose worker runs before the `color` sent just ahead of it both send
|
||||
the new color, not the old one.
|
||||
4. A settings file from before the color has no `Color`: it reads as Triforce gold (`#e0b23e`), which is also what
|
||||
**Reset to gold** sends. An uninstall removes it with the other settings.
|
||||
|
||||
## Checks
|
||||
|
||||
- `dotnet run --project launcher/LogicTests color`: the parsing of the page's message, the keeping and reading back in
|
||||
`registry.txt` and `builder.ini`, old files, a bad kept value, `--snapshot`'s no writes, and the color sent back
|
||||
under `--snapshot` and when the commands' workers run out of order.
|
||||
- `python -I -B tests/test_launcher_color.py`: the page and the launcher agree on the message, the default and the image,
|
||||
the image rides in the exe (the `ui/**` embed), and every word the page shows spells "color": the page's own
|
||||
(`app.js`, `demo.js`, `index.html`) and every string in the launcher's C# (`App`, `Logic`, `Core`), the game options'
|
||||
help the Settings rows show as their tooltips among them. That help is the game menu's own words
|
||||
(`gpu/vk/vk_menu.cpp`, which "game options: schema matches vk_menu.cpp" compares); the menu's own table is spelled
|
||||
"color" by the in-game change set, not by this one.
|
||||
- `node --check launcher/ui/app.js`.
|
||||
43
docs/launcher-music.md
Normal file
43
docs/launcher-music.md
Normal file
@@ -0,0 +1,43 @@
|
||||
# The launcher's music
|
||||
|
||||
The launcher plays the player's own game's music: Saria's Song on the reveal, the Temple of Time while it builds, the
|
||||
item fanfare when the build is done, and the Fairy Fountain on the main screens (`MUSIC` in `launcher/ui/app.js`).
|
||||
Nothing of it is in the repository or the exe: `launcher/Core/GameFiles` decodes the game's BWAV files into WAV files
|
||||
in `%LOCALAPPDATA%\Triforce\launcher\media` (`<track>-<key>.wav`), and the page asks for them as `audio/<track>.wav`.
|
||||
|
||||
## The loop
|
||||
|
||||
Each of the three songs plays an intro once and then loops from a sample after it to its end, as in the game:
|
||||
|
||||
| Track | File | Loop |
|
||||
|---|---|---|
|
||||
| Saria's Song | the eight `BGM_LostWood_0` .. `_7` layers, mixed | after a short intro, to the end (every layer the same) |
|
||||
| Fairy Fountain | `BGM_GreatFairyFountain` | after a long intro, to the end |
|
||||
| Temple of Time | `BGM_ShrineOfTime_Child` | from a fraction of a second in, to the end |
|
||||
| Item get | `BGM_ItemGet_Large` | none: it plays once |
|
||||
|
||||
The BWAV header gives the loop (each channel's end at 0x3C, 0xFFFFFFFF for none, and start at 0x40).
|
||||
`Bwav.Decode` reads it, `Bwav.Mix` keeps it when every layer has the same one, and `Bwav.ToWav` writes it as a `smpl`
|
||||
chunk after the samples: one forward loop, its last sample counted in, as the format has it. The samples stay at byte
|
||||
44, so a reader that knows only the plain header still reads the file.
|
||||
|
||||
The page plays music through Web Audio: a track is fetched and decoded once into an `AudioBuffer`, and an
|
||||
`AudioBufferSourceNode` loops it between `loopStart` and `loopEnd`, sample for sample (the context runs at the game's
|
||||
48 kHz, so nothing is resampled). Its volume goes through a `GainNode` with the same fades, volume and mute as before. A
|
||||
WAV without a `smpl` chunk (made by an older launcher) loops whole. Without Web Audio, or for a file it cannot decode,
|
||||
an `<audio>` element plays it as before. The item fanfare starts at its full level (a fade-in would swallow its first
|
||||
notes); a track faded out before it is read never starts, and a track that stops lets go of its decoded samples.
|
||||
|
||||
Before, an `<audio>` element looped the whole file: the intro came back on every loop, the end jumped to the start with
|
||||
a step in the wave, and the element paused between the end and the start.
|
||||
|
||||
A launcher with this change decodes the sounds again once (the WAV cache's key has `SoundVersion` in it, and the files
|
||||
made before are removed); the title art keeps its files.
|
||||
|
||||
## Tests
|
||||
|
||||
- `dotnet run --project launcher/LogicTests media`: the loop read from a BWAV the test writes, kept by the mix, written
|
||||
as `smpl` and read back, and the cache made again.
|
||||
- `python -I -B tests/test_launcher_music.py`: the page's music code run in node with stand-ins for Web Audio, fetch and
|
||||
`<audio>` (`tests/launcher_music_test.js`), and the page and the launcher checked to read and write the loop at the
|
||||
same place.
|
||||
185
docs/lazy-writeback.md
Normal file
185
docs/lazy-writeback.md
Normal file
@@ -0,0 +1,185 @@
|
||||
# Lazy write-back: built, on by default
|
||||
|
||||
**Since options_version 5 (2026-11-16): on by default, at the owner's word, with the row and `RX_VK_LAZY_WB` to
|
||||
turn it off; still never on AMD's driver by the row alone. The status below is from before.**
|
||||
|
||||
**Status (2026-11-15): built behind the hidden Debug tab's Lazy write-back row, off unless set, and `RX_VK_LAZY_WB`
|
||||
(`=1` on, `=0` off, `=2` the self-check) over it. Nobody has run it in the game.** The design below is what was built;
|
||||
"As built" at the end says how, and what only the game can show. The note was first written when it was studied for the
|
||||
port of NX2navi's features and left out, because in this code base it cannot be proven correct without running the
|
||||
game; the owner then asked for it to be built all the same, off by default, so that it can be turned on and compared.
|
||||
The reasons in "Why it is not built" still stand as the things to watch for.
|
||||
|
||||
## What it is
|
||||
|
||||
What a shader writes into a mirrored pool lives in the pool's VRAM mirror. At the end of every submission the renderer
|
||||
copies it back into guest memory, in case the CPU reads it (`gpu/vk/vk_memory.cpp`, the end of
|
||||
`memory_end_submission`). The game's CPU reads very little of it. NX2navi stopped copying a range once it had gone
|
||||
back one time, and copies it only when something touches it: "about a third more frames" in open areas, with 66 MiB a
|
||||
frame no longer crossing PCIe (7 to 9 ms of a frame's 13 to 15 ms of GPU time there).
|
||||
|
||||
How NX2navi knows about a touch: it takes the access away from the range's pages (no read, no write). The first access
|
||||
faults, the fault handler has the renderer copy the range down, gives the pages back and runs the access again
|
||||
(NX2navi's `runtime/memory.c` "Lazy write-back", `runtime/vk_memory.c` `lazy_fetch`). Around that sit a thaw before every
|
||||
kernel call that is handed guest memory, a list of ranges in flight, a window between the replay's last check and its
|
||||
hand-over, a hold after a touch, and a choice by driver (`wb_lazy_choice.h`: off on AMD's driver after two reports of
|
||||
the game ending in play, off outside Windows). It took several releases of play tests to get there, and one of its
|
||||
refinements still loses the device on some loads.
|
||||
|
||||
## What is different here
|
||||
|
||||
- **CPU writes are known late, CPU reads never.** Lifted stores mark their page in a table (`A64_MARK`, the soft
|
||||
write watch) and the renderer collects the marks at the next submission (`memory_collect_written`). A load is a
|
||||
plain pointer access (`A64_PTR`) with no hook. So "copy it when the CPU touches it" has only two possible sources
|
||||
here: page protection, as in NX2navi, or the calls the game makes before it reads.
|
||||
- **Most of the gain is already taken another way.** Since 2026-11-09 a write-back is a compute pass that compares
|
||||
the mirror with a VRAM shadow of guest memory and writes only the words that differ (`wb_diff`, on by default).
|
||||
Measured where NX2navi measured (Hyrule Field, about 69 MB a frame of written ranges): the whole copies took 4.4 ms
|
||||
of the GPU's 13, the pass takes about 0.7 ms. What a lazy write-back could still save here is that pass, the pages
|
||||
it reads into the shadows, and the waits of captures for write-backs still on the GPU (`wb_pages`, the "late"
|
||||
pages of `take_capture`). Nobody has measured how much that is.
|
||||
- **The order in which the GPU's words reach guest memory is tuned by play.** The exposure meter is a sum the GPU
|
||||
adds to and the CPU reads and clears; the readback waits in `vk_replay.cpp` `invalidate`, the late pages, the
|
||||
deferred syncs and the report words cut out of write-backs (`memory_cpu_owned`) each fixed a flicker that only
|
||||
showed in the game (white flashes, blinking cells, a dark frame every 15 to 30). A lazy write-back moves every one
|
||||
of those moments.
|
||||
|
||||
## Who reads what is written back
|
||||
|
||||
Everything below would see old bytes while a copy is owed, unless it is made to pay first.
|
||||
|
||||
| Reader | Where | Can it be caught? |
|
||||
|---|---|---|
|
||||
| The game's own code | lifted loads, any thread | Only by a fault. `nvnBufferInvalidateMappedRange` and `nvnMemoryPoolInvalidateMappedRange` reach the renderer (`host/nvn_native.c`, `rx_nvn_invalidate`), and it already waits there for the range's write-back. Whether the game calls one before every read of GPU results is not known; `HOS_NVN_LOG_FLUSH` logs the calls |
|
||||
| NVN event words | `nvnEventGetValue`, `hos_nvn_event_value` (the present poll, `host/hos_thread.c`) read guest memory with no call into the renderer | A hook in the host |
|
||||
| Host code given game buffers | `memcpy` and the like through `GP()`: audio voices on the mixer thread, the movie's packets, the codec, `nn_libc.c` | By a fault, on that thread |
|
||||
| The kernel | `ReadFile` and `WriteFile` straight on guest pointers (`host/nn_fs.c` `read_at` and the write beside it: every file read, every save); the same as `pread`/`pwrite` on Linux | **No.** A call into a page without access fails with an error and no fault (`host/mkw_glue.c` says so of reserved pages). Each needs a thaw before it, as NX2navi has |
|
||||
| The renderer's CPU reads | captures (`capture_make` on the game's thread and its helpers, `guarded_copy`, which drops a page that faults on Windows and is a plain `memcpy` on Linux), `memory_read_as_submitted`, texture uploads (`vk_texture.cpp` `guest_bytes`), shader code, the bank's descriptors | By a fault, but on threads that hold `pools_mu` or `replay_mu`, which a copy must not need |
|
||||
| The GPU through the import | every `vkCmdCopyBuffer` from `Pool::buffer` (`mirror_make`, `mirror_pieces_make`, `mirror_sync`'s live pages, `memory_rb_refresh`, `late_deferred_bound`, the shadows' fill) and every binding `gpu_buffer_bind` or `gpu_buffer` gives the import for | Never by a fault: each site has to pay first or bind the mirror |
|
||||
| Code that takes guest memory for complete | a mirror given back or remade (`mirror_free`, `mirrors_give_back`, `demote_trials`, `pieces_set_direct`, `retire_pool` with its inheritance), `texture_flush_to_guest` | Each has to pay first |
|
||||
|
||||
And three things undo a page's protection behind the renderer's back on Windows: `import_pool` commits its whole
|
||||
range again (`VirtualAlloc(MEM_COMMIT, PAGE_READWRITE)`, which its own comment says resets protection), `hos_commit`
|
||||
does the same for the game's mappings, and `hos_mem.c` discards and decommits pages.
|
||||
|
||||
## A design for Triforce
|
||||
|
||||
The states of a guest page of a mirrored pool, as a write-back sees them:
|
||||
|
||||
| State | Guest memory holds | How it is left |
|
||||
|---|---|---|
|
||||
| current | everything the GPU wrote | a written binding of a later submission: **written** |
|
||||
| written | all but this submission's writes | the submission's end: copied (the diff pass), then current; or, when the page went back once before and nothing forbids it, no copy: **owed** |
|
||||
| owed | an older version; the mirror alone has the GPU's | a reader: **paying**; a later written binding: stays owed, the newer write joins the same debt (one copy pays for all of them, mirror to guest, so they land in order by construction) |
|
||||
| paying | being copied by one thread | the copy done and waited for: current. Every other reader of the page waits for that copy, none starts a second |
|
||||
|
||||
Rules that keep it correct:
|
||||
|
||||
1. A page is owed only whole, only in a pool with a mirror in memory there (not a piece left in guest memory), only
|
||||
when the mirror holds every CPU write to it (`page_gen(page) <= synced_gen`: a copy from the mirror would
|
||||
otherwise put older bytes over the CPU's), and never where the replay writes words itself (the report words).
|
||||
2. Every reader in the table above pays before it reads. For CPU reads that means the page has no access while it is
|
||||
owed and the fault handler pays; for the kernel a thaw before the call; for the GPU's reads and for code that gives
|
||||
a mirror up, a call that pays the range first.
|
||||
3. A CPU write to an owed page pays first, like a read (the write faults). It does not drop the debt: the write
|
||||
covers part of the page and the rest is the GPU's. Only a write of the whole page, which nothing here can
|
||||
recognise, could drop it. The page then goes the usual way (marked, collected, synced into the mirror), and is not
|
||||
owed again for a while (NX2navi holds it 1 s, doubling to 32 s while it keeps being touched).
|
||||
4. A payment copies after everything submitted so far, in a command buffer of its own that the paying thread waits
|
||||
for; the part being recorded checks at its end whether a range it left owed was paid meanwhile and copies it then.
|
||||
5. The shadows of the write-back diff stay true without change: an owed page was not written by the pass, so the
|
||||
shadow still equals guest memory, and a payment is the same pass run later.
|
||||
6. A capture's wait (`wb_pages`) has nothing to wait for on an owed page: no submission writes it back. The wait
|
||||
moves into the payment.
|
||||
|
||||
The page-state logic belongs in a header of its own with tests for the rules above (each reader pays first; a CPU
|
||||
write pays and does not drop; debts merge; one payer at a time), as the write generations have theirs
|
||||
(`gpu/vk/write_gen_table.h`).
|
||||
|
||||
## Why it is not built
|
||||
|
||||
None of these can be settled by a test without the game:
|
||||
|
||||
- **The set of kernel accesses.** One missed `ReadFile` into an owed page is a short read handed to the game, with no
|
||||
fault and no log. The list above is what a search of `host/` found, not a proof; libraries and drivers that are
|
||||
handed guest pointers (the movie decoder, audio) were read, not run.
|
||||
- **The driver.** The pages are imported into Vulkan as host memory (`VK_EXT_external_memory_host`; a `udmabuf` on
|
||||
Linux). Taking their access away while the import lives is what NX2navi's AMD reports point at. Importing a pool
|
||||
whose pages have no access is likely to fail outright.
|
||||
- **Locks.** The fault can come on any thread, including the replay thread and the capture helpers while they hold
|
||||
the renderer's locks, and on Linux inside a signal handler. NX2navi's rules for this (ranges in flight, the
|
||||
window, fetches outside its freeze lock) were found by profiles and stalls in play.
|
||||
- **The game's timing.** See "tuned by play" above. Paying at the first touch instead of at the submission's end
|
||||
changes when the exposure meter, the event words and the culling's results appear to the CPU.
|
||||
- **Whether it is worth it.** With the diff pass in place the saving is unmeasured and may be small.
|
||||
|
||||
What was considered and rejected as a shortcut: paying only at `nvn*InvalidateMappedRange` and the host's own reads,
|
||||
with no page protection. It needs no fault handler and no thaw, but it is right only if the game announces every read,
|
||||
which nobody knows; a read that is not announced gets old data silently. `RX_VK_WB_MAX_KB` is the same bet made by
|
||||
size, and is off by default for that reason.
|
||||
|
||||
## What to do first, in the game
|
||||
|
||||
1. Measure what is left to gain: `RX_VK_PASS_TIMING` gives the GPU time of "wb diff" and "wb copies" a frame, and the
|
||||
log's "write-back diff (600 submits)" and "a capture taken" lines give the bytes compared and the takes that waited
|
||||
for late pages. `RX_VK_NO_WRITEBACK=1` is the upper bound on frame rate (its picture is wrong wherever the CPU reads
|
||||
results: do not judge the picture with it).
|
||||
2. If that is worth the work: log which written-back ranges the game announces before reading
|
||||
(`HOS_NVN_LOG_FLUSH`, `RX_VK_LOG_INVALIDATE`, `RX_VK_LOG_WRITEBACK`), and build the page probe NX2navi used before
|
||||
its lazy write-back (protect a written-back range, log the first access with its thread and address, give the page
|
||||
back): that turns "the CPU does not read most of it" from NX2navi's game build into a fact about this one, and
|
||||
finds the kernel accesses on the way.
|
||||
3. Then the host side (the fault hook, a thaw before `read_at` and the save writes, the commits that reset
|
||||
protection), the renderer side (rules 1 to 6), a Debug tab row that is off by default, and the same choice by
|
||||
driver NX2navi ended with.
|
||||
|
||||
## As built (2026-11-15)
|
||||
|
||||
Steps 1 and 2 above were not done first: the owner asked for it built behind a row, so that the first run with it on
|
||||
is the measurement. What is where:
|
||||
|
||||
- **The page states**, `gpu/vk/wb_lazy_state.h`, with the rules and a threaded stress test in
|
||||
`tests/test_wb_lazy_state.py`. A page is not owed at once: it is first WATCHED (read-only, still copied at every
|
||||
submission) until a collection of the write watch and a capture, both numbered after the watch began, have come
|
||||
and named no CPU write to it. That is rule 1 for the writes the renderer learns of late; the renderer adds the rest
|
||||
of it: the CPU's newest write generation for the page no newer than its pool's `synced_gen` (`lazy_page_takes`).
|
||||
That asks `cpu_gens`, the CPU's writes alone, and not the one table of generations: the first build asked that one,
|
||||
where every storage buffer a draw or a dispatch writes is stamped while it is recorded (`memory_stamp_gpu`) and
|
||||
`synced_gen` is raised only at the submission's end, so every page being written back looked newer than its sync
|
||||
and the owner's first session with the row on watched and owed nothing (`tests/test_lazy_wb_cpu_gens.py`). A CPU write to a watched page faults and ends the watch with no copy; to an owed
|
||||
page it pays first. A payment between a recording's end and its submit leaves the rest PENDING until that submit
|
||||
(`memory_lazy_submitted`, called from `submit_recording` under the queue's lock). A touch holds its 64 KB block
|
||||
for a second, doubling to 32 s, as NX2navi does. Whether it runs: `wb_lazy::choose`.
|
||||
- **The renderer**, `gpu/vk/vk_memory.cpp` "lazy write-back": `lazy_runs` at each write-back (after the CPU's words
|
||||
are cut and the runs clipped to a mirror's resident pieces; never a page with event words), the payment (`lazy_pay`:
|
||||
its own command buffer after everything submitted, the held submits flushed first, waited for; the shadows told of
|
||||
the pages, rule 5), `lazy_settle` before every copy out of an import and every binding given the import (rule 2 for
|
||||
the GPU), and before a mirror, a piece, a retired pool or an import is given up; `import_pool` holds the pages
|
||||
while it commits and imports. The fault handler is the process's first vectored handler, added when the renderer
|
||||
starts, ahead of host/mkw_glue.c's. Locks: see the section's comment (a payment needs neither `pools_mu` nor
|
||||
`replay_mu`, and logs nothing). `tests/test_lazy_wb_hooks.py` holds these hooks in place.
|
||||
- **Several payments at once** (2026-11-16, NX2navi's `wb_lazy_unlocked`): each payment holds one of four slots, a
|
||||
command buffer and a fence of its own (`gpu/vk/wb_lazy_payers.h`); it was one payer at a time, so a thread that
|
||||
touched another owed page, or the replay giving a mirror up, waited for the first one's copy before its own went onto
|
||||
the queue. Two payments never have the same pages (the table's PAYING has one payer), and each copy still goes onto
|
||||
the one queue after everything before it. `RX_VK_LAZY_PAYERS=1`: one at a time again. The threaded stress test runs
|
||||
with four slots and with one. What it does not change: a payment for a mirror given up still runs on the replay
|
||||
thread under `pools_mu` (NX2navi moved its own out of its pool lock as well); moving it would mean reworking every
|
||||
give-up path, for payments of which the owner's sessions have a few a minute.
|
||||
- **The host**: `hos_guest_io_begin` / `hos_guest_io_end` (`host/rx_backend.c`) around every kernel call on guest
|
||||
memory and every change of guest pages' state: `read_at` and the file writes (the saves too) in `host/nn_fs.c`, the
|
||||
image's read, `hos_commit`, the discards and decommits of `host/hos_mem.c` and `host/hos_svc.c`, the host zstd
|
||||
decoder's one-shot, and the debug dumps. `tests/test_lazy_wb_host.py` reads `host/*.c` for kernel calls that do
|
||||
neither, and every exception is in its list with a reason. Audio, the movie and the codec reach guest memory with
|
||||
the CPU only (`memcpy` and the like): their touches fault and pay.
|
||||
- **Linux**: off (`wb_lazy::choose`), the row not shown. **AMD's driver**: the row alone leaves it off;
|
||||
`RX_VK_LAZY_WB=1` turns it on there.
|
||||
- **The self-check**, `RX_VK_LAZY_WB=2`: guest memory is the eager copy (every copy still made); a payment reads the
|
||||
mirror's pages, what it would have copied, and compares them with guest memory. A difference is logged with its
|
||||
address at the next 600-submit line (`MISMATCH at guest`), not from inside the fault.
|
||||
|
||||
What only the game can show: whether the driver keeps an import whose pages have no access (NX2navi's AMD reports),
|
||||
whether the list of kernel calls is whole in practice (a missed one is a short read with no fault), how often the game
|
||||
touches what is owed (a touch costs a GPU round trip on the touching thread), and the timing of the exposure meter and
|
||||
the event words. The 600-submit line `vk: lazy write-back (600 submits)` says what it saved and who paid; with
|
||||
`RX_VK_LAZY_WB=2` the `check` line says whether it would have been right.
|
||||
@@ -25,3 +25,12 @@ These checks use no driver or real cache and do not reproduce issue #1022's repo
|
||||
its root cause, validate saving a multi-gigabyte driver cache, force an OS thread-resource failure, or change cache
|
||||
size/retry policy. The existing wait-then-store protocol for concurrent forced callers is unchanged; the current
|
||||
force call is in prepare-only startup. Driver/native faults and thread-detach failures are outside this change.
|
||||
|
||||
## The file's bound
|
||||
|
||||
The cache's size has a rule since: `gpu/vk/pcache_bound.h`. A file over the bound is not read (the cache starts
|
||||
empty), and data over the bound is neither fetched nor written, so the file keeps its last save
|
||||
(`RX_VK_PCACHE_MAX_MB`, in the development guide's switches). The fixture's `over-bound` case runs the save's side
|
||||
against the actual function: under and at the bound the data is written as ever, over it the old file stays, in a
|
||||
forced save and on the worker, and with the bound lifted the same data is written. The read's side is
|
||||
`tests/test_pcache_bound.py`. A read that ends early is refused whole by the header; no test makes a read end early.
|
||||
|
||||
@@ -109,6 +109,31 @@ that only exists on that side does not stay mixed in with the new save.
|
||||
- **Backups.** The script keeps one for every copy. Delete old ones when you no longer need them.
|
||||
- Start each program once and save in it before the first copy, so that its save folder exists.
|
||||
|
||||
## A save as one zip, from the game itself
|
||||
|
||||
The System tab of the game's Options has **Export save**, **Export save for a bug report** and **Import a save**
|
||||
(`host/save_share.c`; the readme's *Export and import a save*). They move one zip, not a folder:
|
||||
|
||||
```
|
||||
manifest.json "format": "triforce-save", "format_version": 1, the game's title id, the Triforce version, the export time (UTC)
|
||||
save/ account.sav, option.sav, slot_00/ ... as above
|
||||
bug-report/ in a bug report: triforce_log.txt, triforce_log.1.txt, triforce_options.ini, triforce_hang.txt
|
||||
```
|
||||
|
||||
- The export is the save the next start would load, read with no save in progress (`host/nn_fs.c` `hos_fs_save_read`):
|
||||
a finished commit still waiting in `user/save-tx/<user id>/ready-<n>`, else `user/save/<user id>`. It goes to the
|
||||
Downloads folder (`HOS_SAVE_EXPORT_DIR=<folder>`: there instead), or beside the game when that cannot be written.
|
||||
- The import takes the newest `.zip` of the game folder's `save-import` folder: this format or NX2navi's own
|
||||
(`"format": "nx2navi-save"`, the same layout). It is checked whole first (`host/save_pack.c`: the manifest, every path
|
||||
inside the save, no links, sizes, each file's CRC-32) and refused otherwise. Its save then waits in
|
||||
`user/save-tx/<user id>/import` and the zip moves to `save-import/imported`.
|
||||
- The next start copies the save about to be replaced to `user/save-backup/<user id>/import-<UTC date>-<time>` (kept until
|
||||
you delete it: the five rolling backups beside it are the saves' own), then makes the import the save
|
||||
(`host/save_tx_share.h`). Choosing the row again before that start cancels it.
|
||||
- Triforce Setup's Convert saves, used after an import was left waiting, drops the waiting import: what was converted last
|
||||
is the save.
|
||||
- With `HOS_SAVE_TX=0` the game writes its save in place, and nothing is exported or imported in that run.
|
||||
|
||||
## What was tested
|
||||
|
||||
- Linux (Arch-based), Hookshot v2.0.5 and Lichii built from source on the same day.
|
||||
|
||||
@@ -38,8 +38,8 @@ Agents may add and re-score rows; a person reviews the change like any other PR.
|
||||
| 0.29 | Amd RX 580 (Polaris, no mesh shaders): hair renders white and bright, and the game crashes when Link moves; a second RX 580 8 GB user reports the same | S1 | 0.145 | 0.4 | | #bug-reporting "Amd rx 580 crashes" (no log yet) |
|
||||
| 0.24 | Handheld with no keyboard (Asus ROG Xbox Ally X, controller only): fullscreen and the F1 options menu cannot be reached; no log yet | S1 | 0.12 | 0.4 | | #bug-reporting "Xinput devices cannot enter fullscreen or menu" |
|
||||
| 0.12 | AMD RX 9070 XT: texture glitch that appears after some play (since v2.0.5); same as the tracked AMD texture artifact | S2 | 0.10 | 0.4 | | #bug-reporting "texture glitch on 9070 XT aris"; #57, #1006 |
|
||||
| 0.11 | Intel GPU: black screen after loading a save (or on the Deku Tree), sound only, very low fps; three reports, no logs yet; tracked in the GitHub issue linked in #bug-reporting replies (#999) | S0 | 0.035 | 0.4 | | #bug-reporting "Intel G3E. Title and intro loads...", "intel arc b580 problems", "game gets stuck on a black screen when loading a save (intel gpu... gulp...)" |
|
||||
| 0.11 | Crash at the door after the owl intro (Navi wakes Link, 2.0.13): access violation in `igvk64.dll` at start-up; GPU not given yet, reach uses Intel's 3.5% until it is | S0 | 0.035 | 0.4 | | #bug-reporting "2.0.13 Crash after you see the owl during intro cutscene" (5 logs) |
|
||||
| 0.11 | Intel GPU: black screen after loading a save (or on the Deku Tree), sound only, very low fps; three reports, no logs yet; tracked in the GitHub issue linked in #bug-reporting replies (#999). What a log would settle: `vk: pool ... not imported (VkResult ...)` lines (the driver refusing host memory, whose bindings then draw nothing) against a `device lost` report. Not a case for running without the host-memory import: the renderer has no such path (`memory_init` in `gpu/vk/vk_memory.cpp` stops without it), and NX2navi's Intel branch, which tried one, found it drew the 3D black and went back | S0 | 0.035 | 0.4 | | #bug-reporting "Intel G3E. Title and intro loads...", "intel arc b580 problems", "game gets stuck on a black screen when loading a save (intel gpu... gulp...)" |
|
||||
| 0.11 | Crash at the door after the owl intro (Navi wakes Link, 2.0.13): access violation in `igvk64.dll` at start-up; GPU not given yet, reach uses Intel's 3.5% until it is. To check in the logs: the address. `igvk64.dll+0x79ddf3` (read of 0x190) is the driver's `vkGetPipelineCacheData` beside a pipeline build on the same cache, which `pipeline_cache_save` (`gpu/vk/vk_draw.cpp`) has kept apart for the saves in play since 11-12, on Intel; the saves asked for outright (the prepare, the stage libraries' pass, the exit) still fetch the cache whatever else builds. It is not the host-memory import: NX2navi's Intel branch took its crash for that first and found the same address | S0 | 0.035 | 0.4 | | #bug-reporting "2.0.13 Crash after you see the owl during intro cutscene" (5 logs) |
|
||||
|
||||
|
||||
## Fixed, waiting for a release
|
||||
|
||||
@@ -227,6 +227,11 @@ int nvn_depth_minus_w(void);
|
||||
void rx_nvn_submit(const uint64_t *handles, int n);
|
||||
/* (the same with the queue it came from: the game's queues have their own state and order each other with events) */
|
||||
void rx_nvn_submit_q(const uint64_t *handles, int n, uint64_t queue);
|
||||
/* (Triforce, gpu/vk/async_rules.h; host/rx_backend.c) the replay on a thread of the back end's own (RX_VK_ASYNC=1): whether it
|
||||
is on; the events the submit this thread makes next waits for; the present poll's wait for that thread (1: nothing left) */
|
||||
int rx_async_replay(void);
|
||||
void rx_nvn_submit_waits(const uint64_t *events, int n);
|
||||
int rx_nvn_replay_idle_wait(uint64_t timeout_ns);
|
||||
/* fence value of everything submitted so far; waits until the GPU reached a fence (1: reached, 0: timed out) */
|
||||
uint64_t rx_nvn_fence(void);
|
||||
uint64_t rx_nvn_fence_q(uint64_t queue); /* (hookshot) a fence of one queue */
|
||||
|
||||
@@ -15,6 +15,7 @@
|
||||
#include <span>
|
||||
|
||||
#include "common/alignment.h"
|
||||
#include "../vk/cpu_count.h" // ((Triforce, 11-22, OS-49) the threads to size by: HOS_SIM_CPU
|
||||
#include "common/div_ceil.h"
|
||||
#include "video_core/texture_cache/format_lookup_table.h"
|
||||
#include "video_core/textures/astc.h"
|
||||
@@ -766,7 +767,7 @@ std::unique_ptr<Image> decode(const TICEntry& tic, const TexInfo& info_in, uint6
|
||||
copy_layers >= 4 && level_size.depth == 1 && (size_t)level_size.width * level_size.height * 4 <= (8u << 20)) {
|
||||
const Extent3D block = AdjustMipBlockSize(tiles, li.block, level, li.num_levels);
|
||||
const u32 stride_alignment = StrideAlignment(tiles, info.block, gob, li.bpp_log2);
|
||||
const unsigned nthreads = std::max(1u, std::min<unsigned>(copy_layers, std::thread::hardware_concurrency() - 1));
|
||||
const unsigned nthreads = std::max(1u, std::min<unsigned>(copy_layers, cpu_count() - 1));
|
||||
std::atomic<u32> next_layer{0};
|
||||
static std::atomic<int64_t> t_swz, t_dec, t_up;
|
||||
auto work = [&] {
|
||||
|
||||
@@ -146,7 +146,9 @@ Designed after totk-pc-recomp's runtime/vk_*.c (MIT) and Eden's renderer_vulkan,
|
||||
- **Mesh shading**: `NOP_DRAW_MESH` records NVN2's DrawMeshTasks*; programs whose code writes the internal stage
|
||||
buffer are the task / mesh stages (marked in their headers as Eden's GraphicsEnvironment does). Unverified.
|
||||
- **Frame interpolation** (`vk_framegen.cpp`, `RX_VK_FRAMEGEN=1`): the WebGPU renderer's, ported to GLSL
|
||||
(`shaders/`, compiled with glslang): made frames between the game's, shown evenly by a display thread.
|
||||
(`shaders/`, compiled with glslang): made frames between the game's, shown evenly by a display thread. (Triforce) Its
|
||||
vectors are refined to one per 8x8 pixels, to a fraction of a pixel, and each made pixel takes one motion
|
||||
(`fg_half`, `fg_fine_flow`, `fg_fine_pick`, `fg_fine_synth`; `RX_VK_FRAMEGEN_FINE=0`: the 16x16 blocks as ported).
|
||||
- **Headless**: without a window presents render offscreen; `RX_VK_SHOTS=<dir>` writes every `RX_VK_SHOT_EVERY`-th
|
||||
as a PNG. `work/run-vk.ps1` runs it (refuses while Siege runs, stops past `-MaxGB` of working set);
|
||||
`work/vksum.py <run>` summarises the log.
|
||||
|
||||
149
gpu/vk/aa_rows.h
Normal file
149
gpu/vk/aa_rows.h
Normal file
@@ -0,0 +1,149 @@
|
||||
/* (Triforce) The Insert menu's Anti-aliasing row (vk_menu.cpp), and the rule that finds the game's own anti-aliasing passes
|
||||
among a frame's compute dispatches (vk_draw.cpp aa_stand_in, vk_aa.cpp), as plain functions. No Vulkan, no menu:
|
||||
tests/test_aa_row.py compiles and runs them. After NX2navi's aa setting (its runtime/vk_aa.c and vk_exec_draw.c
|
||||
aa_dispatch).
|
||||
|
||||
OoT anti-aliases every frame of the 3D with SMAA in three compute passes, right after the draw that adds the bloom and
|
||||
tone maps: the edges (4 x 4 pixel tiles), the blending weights from them (an indirect dispatch over the tiles with edges)
|
||||
and the composition, which samples the tone mapped picture and the weights and stores the smoothed picture into the
|
||||
target the gameplay upscale then takes. It stores each channel's square root, the three packed as one RGB9E5 word, through
|
||||
an integer view of that B10G11R11 target (shaders/aa_common.glsl). */
|
||||
#pragma once
|
||||
|
||||
#include <cstdint>
|
||||
|
||||
#include "upscaler_rows.h"
|
||||
|
||||
namespace mkwvk::aa_rows {
|
||||
|
||||
/// The Anti-aliasing row (opts.aa, saved as anti-aliasing; RX_VK_AA over it): the game's own SMAA as it is, none (the tone
|
||||
/// mapped picture stored as it is), FXAA over it (shaders/aa_fxaa.glsl), or SMAA at its Ultra preset in place of the game's
|
||||
/// (shaders/smaa_*.comp). The saved values keep their meaning: a new choice comes after these
|
||||
enum { GAME = 0, OFF = 1, FXAA = 2, SMAA = 3, kValues = 4 };
|
||||
inline const char* name(int v) {
|
||||
switch (v) {
|
||||
case OFF: return "Off";
|
||||
case FXAA: return "FXAA";
|
||||
case SMAA: return "SMAA Ultra";
|
||||
default: return "Game SMAA"; // (also for a value past the row; "Game", the owner's word, 11-16)
|
||||
}
|
||||
}
|
||||
/// The row's value as one of its choices, RX_VK_AA=<0..3> over it (env: its number, -1 unset). Anything else is the game's own
|
||||
inline int chosen(int row, int env) {
|
||||
const int v = env >= 0 ? env : row;
|
||||
return v >= GAME && v < kValues ? v : GAME;
|
||||
}
|
||||
/// DLSS, DLAA, FSR 3.1, FSR 4 and their native modes anti-alias the 3D themselves: while one of them draws (upscaler: what
|
||||
/// runs, upscaler_rows.h) the row does not apply
|
||||
inline bool upscaler_own(int upscaler) { return upscaler != upscaler_rows::GAME; }
|
||||
/// What is in effect: the choice with the game's own upscaler. ((Triforce, FSR3-06, FSR4-5) While such an upscaler draws from
|
||||
/// the tone mapped picture (Upscale before bloom off), none: the game's SMAA is left out, as AMD and NVIDIA ask for jittered
|
||||
/// frames no other anti-aliasing has been over (SMAA before them softened what they gather, and its edges, found anew in
|
||||
/// each jittered frame, flickered). It was the game's own SMAA before. Before the bloom (before_bloom) the upscaler takes
|
||||
/// the scene, which the SMAA has not been over: the game's own then, as before. env_set (RX_VK_AA for the run): its choice
|
||||
/// there too, for an A/B (RX_VK_AA=0: the game's SMAA under the upscaler)
|
||||
inline int effective(int choice, int upscaler, bool before_bloom = false, bool env_set = false) {
|
||||
if (!upscaler_own(upscaler) || env_set) return choice;
|
||||
return before_bloom ? (int)GAME : (int)OFF;
|
||||
}
|
||||
/// The row has nothing to choose then (shown without arrows; Left and Right leave the saved value alone), nor with
|
||||
/// RX_VK_AA set for the run
|
||||
inline bool fixed(int upscaler, bool env_set) { return upscaler_own(upscaler) || env_set; }
|
||||
|
||||
// ---- the game's passes ----
|
||||
|
||||
/// The game's three SMAA programs, by their keys (vk_shader.cpp code_key: a hash of the compute stage's control section and
|
||||
/// code as the game hands them over, the same on every PC and with every setting). Game build e0f3ccfa. NX2navi knows the
|
||||
/// same three by another hash of the same code (fcb21f2afa88378c, e44377cbfafc1101, 470d661afb1cb73a in its vk_aa.c): these
|
||||
/// are the programs of a build's shader list (triforce_shaders.bin) with those hashes. Another game version has other keys:
|
||||
/// nothing is then found, and the frame is drawn as before the row (vk_aa.cpp says so in the log)
|
||||
enum { NONE = 0, EDGES = 1, WEIGHTS = 2, COMPOSE = 3 };
|
||||
constexpr uint64_t kEdgesKey = 0x26fe0c70ab9fc547ull, kWeightsKey = 0x685adb41edc19de7ull, kComposeKey = 0x6637ac1d052d22d4ull;
|
||||
constexpr int role(uint64_t key) {
|
||||
return key == kEdgesKey ? EDGES : key == kWeightsKey ? WEIGHTS : key == kComposeKey ? COMPOSE : NONE;
|
||||
}
|
||||
|
||||
/// What a dispatch has at one of its program's texture or image bindings, as far as the rule looks
|
||||
enum { OTHER = 0, R11 = 1, E5 = 2 }; // (the image's own format: B10G11R11, the scene's and the picture's; E5B9G9R9; anything else)
|
||||
struct Bound {
|
||||
uint64_t image = 0; // (which image: any number that tells two apart; 0 nothing bound)
|
||||
int format = OTHER;
|
||||
uint32_t width = 0, height = 0;
|
||||
bool flat = false; // (a plain 2D image, one sample, seen as 2D at its first level)
|
||||
bool written = false; // (an image binding the program stores into)
|
||||
bool integer = false; // (the program reads or stores it as integer words, not as colours: the variant of the program
|
||||
// that runs is the one translated for the views bound, so this is the view's kind)
|
||||
};
|
||||
/// The composition's two ends among what its dispatch has bound (the program's own bindings, in its order): the target, the one
|
||||
/// flat B10G11R11 (or E5B9G9R9: the same words) image it stores integer words into, and the tone mapped picture, the one flat
|
||||
/// B10G11R11 texture of the target's size it samples as colours that is another image (the weights it samples too are read
|
||||
/// as integer words). src and dst: their places in the lists. why: nullptr when both are found, else what was not as expected;
|
||||
/// the game's own pass then runs, nothing is guessed
|
||||
struct Ends {
|
||||
int src = -1, dst = -1;
|
||||
const char* why = nullptr;
|
||||
};
|
||||
inline Ends compose_ends(const Bound* textures, int ntex, const Bound* images, int nimg) {
|
||||
Ends e;
|
||||
int found = 0;
|
||||
for (int i = 0; i < nimg; i++) {
|
||||
const Bound& b = images[i];
|
||||
if (!b.image || !b.written || !b.integer || !b.flat || (b.format != R11 && b.format != E5) || !b.width || !b.height) continue;
|
||||
if (!found++) e.dst = i;
|
||||
}
|
||||
if (found != 1) return e.dst = -1, e.why = found ? "it stores into more than one such target" : "it stores into no B10G11R11 target as integer words", e;
|
||||
const Bound& d = images[e.dst];
|
||||
found = 0;
|
||||
for (int i = 0; i < ntex; i++) {
|
||||
const Bound& b = textures[i];
|
||||
if (!b.image || b.image == d.image || b.integer || !b.flat || b.format != R11 || b.width != d.width || b.height != d.height) continue;
|
||||
if (!found++) e.src = i;
|
||||
}
|
||||
if (found != 1) return e.src = e.dst = -1, e.why = found ? "it samples more than one picture of its target's size" : "it samples no B10G11R11 picture of its target's size", e;
|
||||
return e;
|
||||
}
|
||||
|
||||
/// Whether a dispatch of the game's edge or weight pass is left out: only the composition reads what they make, so they go
|
||||
/// while another choice than the game's own is in effect and the composition before this one was replaced. One that could
|
||||
/// not be replaced runs as the game's, and these two run again from the next frame on, until one is replaced again.
|
||||
/// keep (RX_VK_AA_KEEP=1): never left out
|
||||
constexpr bool leave_out(int role, int mode, bool last_replaced, bool keep) {
|
||||
return (role == EDGES || role == WEIGHTS) && mode != GAME && last_replaced && !keep;
|
||||
}
|
||||
|
||||
/// What the row says under its help about the passes (vk_aa.cpp aa_state): nothing asked or nothing seen yet, the game's
|
||||
/// composition replaced, or none replaced for a while with another choice than the game's own in effect
|
||||
enum { QUIET = 0, REPLACED = 1, MISSING = 2 };
|
||||
/// The frames with a choice in effect and no composition replaced after which that is said (a minute at 60 a second: a
|
||||
/// loading screen or a menu without the 3D that lasts longer says it too)
|
||||
constexpr uint64_t kMissingFrames = 3600;
|
||||
|
||||
/// What vk_aa.cpp keeps about the game's passes from frame to frame (one for the run, under its lock)
|
||||
struct Track {
|
||||
int mode = GAME; // (the choice in effect in the frame before)
|
||||
bool last_replaced = false; // (the last composition met was replaced: what leave_out asks)
|
||||
int state = QUIET;
|
||||
uint64_t since = 0; // (the frame the choice came into effect, or a composition was last replaced)
|
||||
/// A frame begins with `want` in effect. true: the row's line became MISSING just now (said once in the log)
|
||||
constexpr bool begin(uint64_t frame, int want) {
|
||||
bool missing = false;
|
||||
if (want == GAME) {
|
||||
// (the game's own passes all run from here on: a composition replaced before says nothing about the next choice's
|
||||
// first frame, whose edge and weight passes must run until one is replaced again)
|
||||
last_replaced = false, state = QUIET;
|
||||
} else if (mode != want) {
|
||||
since = frame, state = QUIET; // (another choice: counted anew)
|
||||
} else if (state != MISSING && frame - since >= kMissingFrames) {
|
||||
state = MISSING, missing = true;
|
||||
}
|
||||
mode = want;
|
||||
return missing;
|
||||
}
|
||||
/// A composition met with a choice in effect: replaced, or left to the game
|
||||
constexpr void composed(uint64_t frame, bool replaced) {
|
||||
last_replaced = replaced;
|
||||
if (replaced) since = frame, state = REPLACED;
|
||||
}
|
||||
};
|
||||
|
||||
} // namespace mkwvk::aa_rows
|
||||
145
gpu/vk/achievement_rows.h
Normal file
145
gpu/vk/achievement_rows.h
Normal file
@@ -0,0 +1,145 @@
|
||||
// (Triforce) Achievements on the menu's side: the list the host names (host/achievements.c; vk_api.h HOS_OPT_ACH_*), which of
|
||||
// them are earned, the one a notice shows, and the Achievements page's choice and pages. vk_menu.cpp draws them.
|
||||
//
|
||||
// The achievements are the host's: it reads the lists, asks the save data and keeps what was earned. The menu is told
|
||||
// each one's name, the line about it, its icon and the day it was earned, a byte a call as the save rows' lines are, and
|
||||
// later each one earned as it is. An icon is drawn, not a picture file: a shape and a tone (host/achievements.h has the
|
||||
// same in the same order, ACH_SHAPES and ACH_TONES; tests/test_achievements.py holds the two together).
|
||||
//
|
||||
// The host's calls (text, add, done, earn) come on its threads, one at a time (its lock); the list is drawn on the
|
||||
// renderer's: names, lines and icons are written before count says they are there and never after, a day before its
|
||||
// achievement's earned says so.
|
||||
//
|
||||
// No ImGui and no Vulkan: tests/test_achievements.py compiles it alone.
|
||||
#pragma once
|
||||
#include <atomic>
|
||||
#include <cstdint>
|
||||
#include <cstring>
|
||||
|
||||
namespace achievement_rows {
|
||||
constexpr int kSlots = 128; // (host/achievements.h ACH_MAX)
|
||||
constexpr int kName = 41, kAbout = 161, kDay = 16;
|
||||
constexpr int kEarned = 256; // (vk_api.h HOS_ACH_EARNED)
|
||||
enum Shape { kStar, kSword, kGem, kMedal, kNote, kToken, kShield, kShapes };
|
||||
enum Tone { kGold, kGreen, kRed, kBlue, kPurple, kOrange, kYellow, kSilver, kTones };
|
||||
constexpr int shape_of(int icon) { return (icon & 15) < kShapes ? icon & 15 : kStar; }
|
||||
constexpr int tone_of(int icon) { return (icon >> 4 & 15) < kTones ? icon >> 4 & 15 : kGold; }
|
||||
|
||||
struct List {
|
||||
char names[kSlots][kName] = {};
|
||||
char abouts[kSlots][kAbout] = {};
|
||||
char days[kSlots][kDay] = {};
|
||||
unsigned char icons[kSlots] = {};
|
||||
std::atomic<int> earned[kSlots] = {};
|
||||
std::atomic<int> count{0}; // (the achievements shown: 0 until the host's list is whole)
|
||||
std::atomic<int> fresh{-1}; // (the one earned now, until the notice takes it)
|
||||
// the host's side
|
||||
char in[kName + kAbout + kDay + 8] = {};
|
||||
int in_n = 0, added = 0;
|
||||
|
||||
/// HOS_OPT_ACH_TEXT: the next byte of what is being named
|
||||
void text(int byte) {
|
||||
if (byte > 0 && byte < 256 && in_n + 1 < (int)sizeof in) in[in_n++] = (char)byte;
|
||||
}
|
||||
/// `from` copied into a field of `cap` bytes, cut at a whole letter
|
||||
static void put(char* to, int cap, const char* from) {
|
||||
int n = (int)strlen(from);
|
||||
if (n > cap - 1) {
|
||||
n = cap - 1;
|
||||
while (n && ((unsigned char)from[n] & 0xC0) == 0x80) n--;
|
||||
}
|
||||
memcpy(to, from, (size_t)n);
|
||||
to[n] = 0;
|
||||
}
|
||||
/// HOS_OPT_ACH_ADD: the bytes so far are the next achievement's name and, each after a line feed, the line about it and
|
||||
/// the day it was earned; bits: its icon, and kEarned
|
||||
void add(int bits) {
|
||||
in[in_n] = 0;
|
||||
in_n = 0;
|
||||
if (count.load(std::memory_order_acquire) || added >= kSlots) return; // (shown already, or no room left)
|
||||
const int slot = added++;
|
||||
char* about = strchr(in, '\n');
|
||||
char* day = about ? strchr(about + 1, '\n') : nullptr;
|
||||
if (about) *about++ = 0;
|
||||
if (day) *day++ = 0;
|
||||
put(names[slot], kName, in[0] ? in : "(no name)");
|
||||
put(abouts[slot], kAbout, about ? about : "");
|
||||
put(days[slot], kDay, day ? day : "");
|
||||
icons[slot] = (unsigned char)(bits & 255);
|
||||
earned[slot].store((bits & kEarned) != 0, std::memory_order_relaxed);
|
||||
}
|
||||
/// HOS_OPT_ACH_COUNT: the list is whole: shown from now on
|
||||
void done(int n) {
|
||||
in_n = 0;
|
||||
if (count.load(std::memory_order_acquire)) return;
|
||||
count.store(n < 0 ? 0 : n < added ? n : added, std::memory_order_release);
|
||||
}
|
||||
/// HOS_OPT_ACH_EARNED: the achievement at that place was earned now, the bytes so far the day. false: no such place
|
||||
bool earn(int slot) {
|
||||
in[in_n] = 0;
|
||||
in_n = 0;
|
||||
if (slot < 0 || slot >= shown()) return false;
|
||||
if (!earned[slot].load(std::memory_order_relaxed)) put(days[slot], kDay, in);
|
||||
earned[slot].store(1, std::memory_order_release);
|
||||
fresh.store(slot, std::memory_order_release);
|
||||
return true;
|
||||
}
|
||||
int shown() const { return count.load(std::memory_order_acquire); }
|
||||
bool is_earned(int slot) const { return slot >= 0 && slot < kSlots && earned[slot].load(std::memory_order_acquire) != 0; }
|
||||
int earned_count() const {
|
||||
int n = 0;
|
||||
for (int i = 0, end = shown(); i < end; i++) n += is_earned(i);
|
||||
return n;
|
||||
}
|
||||
/// The one earned since the last ask, for the notice; -1 none
|
||||
int take_fresh() { return fresh.load(std::memory_order_relaxed) < 0 ? -1 : fresh.exchange(-1, std::memory_order_acq_rel); }
|
||||
};
|
||||
|
||||
/// The page's cards: `columns` to a row, `capacity` to a page (both 1 at least), n of them
|
||||
struct View {
|
||||
int choice = 0, page = 0;
|
||||
|
||||
static int pages(int n, int capacity) { return n <= 0 ? 1 : (n + (capacity < 1 ? 1 : capacity) - 1) / (capacity < 1 ? 1 : capacity); }
|
||||
/// The choice kept to the list, and the page shown to the pages there are
|
||||
void keep(int n, int capacity) {
|
||||
choice = n <= 0 ? 0 : choice < 0 ? 0 : choice >= n ? n - 1 : choice;
|
||||
const int last = pages(n, capacity) - 1;
|
||||
page = page < 0 ? 0 : page > last ? last : page;
|
||||
}
|
||||
/// The page that shows the choice
|
||||
void reveal(int capacity) { page = choice / (capacity < 1 ? 1 : capacity); }
|
||||
/// The arrows (cards: Left and Right, rows: Up and Down), Page Up and Down (turns), Home and End. true: the choice moved
|
||||
bool move(int cards, int rows, int turns, bool home, bool end, int n, int columns, int capacity) {
|
||||
if (n <= 0) return false;
|
||||
const int was = choice;
|
||||
const int step = cards + rows * (columns < 1 ? 1 : columns) + turns * (capacity < 1 ? 1 : capacity);
|
||||
choice = home ? 0 : end ? n - 1 : choice + step;
|
||||
keep(n, capacity);
|
||||
if (choice == was) return false;
|
||||
reveal(capacity);
|
||||
return true;
|
||||
}
|
||||
/// The wheel, or a page arrow that leaves the choice where it is: the page shown turned. true: it changed
|
||||
bool turn(int by, int n, int capacity) {
|
||||
const int was = page;
|
||||
page += by;
|
||||
keep(n, capacity);
|
||||
return page != was;
|
||||
}
|
||||
/// A click on a card. true: the choice changed
|
||||
bool choose(int slot, int n) {
|
||||
if (slot < 0 || slot >= n || slot == choice) return false;
|
||||
choice = slot;
|
||||
return true;
|
||||
}
|
||||
};
|
||||
|
||||
/// How far a notice has come in, 0 (out of sight) to 1 (in place): in over the first 250 ms of its `shown_ms`, out over
|
||||
/// the last 300
|
||||
inline float notice_in(int64_t age_ms, int64_t shown_ms) {
|
||||
if (age_ms < 0 || age_ms >= shown_ms) return 0;
|
||||
const float in = (float)age_ms / 250.0f, out = (float)(shown_ms - age_ms) / 300.0f;
|
||||
const float k = in < out ? in : out;
|
||||
return k > 1 ? 1 : k;
|
||||
}
|
||||
} // namespace achievement_rows
|
||||
@@ -13,4 +13,12 @@ inline bool all_by_vram(uint64_t allowed) {
|
||||
return allowed && allowed < (8704ull << 20);
|
||||
}
|
||||
|
||||
// (Triforce, 11-22, the performance audit's TEX-12, after Source A's) And every integrated GPU, whatever it is allowed. Its
|
||||
// allowance is a share of the system's RAM (a 16 GB handheld reported 10,361 MB), so by the allowance alone it kept the
|
||||
// fine-block textures of 2 to 64 MB as RGBA8, in RAM the game needs too, inside a texture budget capped at 2,560 MB.
|
||||
// Source A measured the difference on a discrete card (RX_VK_ASTC_BC_ALL 0 against 1, Kokiri Forest: decoded ASTC
|
||||
// 2,496-2,532 MB against 1,579-1,587, the GPU's frame the same); it was not run on an integrated GPU there either.
|
||||
// The shape of tex_detail::by_vram(allowed, integrated). RX_VK_ASTC_BC_ALL=0: the finer ones as RGBA8, as before
|
||||
inline bool all(uint64_t allowed, bool integrated) { return integrated || all_by_vram(allowed); }
|
||||
|
||||
} // namespace astc_bc_policy
|
||||
|
||||
142
gpu/vk/astc_opaque.h
Normal file
142
gpu/vk/astc_opaque.h
Normal file
@@ -0,0 +1,142 @@
|
||||
// (Triforce) Which ASTC blocks the GPU decoder (host_shaders/astc_decoder_comp_spv.h, Eden's) writes with alpha 255 in every
|
||||
// texel, told from the block's 16 bytes alone: what lets a decoded ASTC texture be kept as BC1 instead of BC3 (vk_texture.cpp,
|
||||
// tex_bc1.h). tests/test_tex_bc1.py
|
||||
//
|
||||
// The answer has to be the decoder's, not the ASTC specification's: a block the decoder takes for an error is written as
|
||||
// (0, 0, 0, 0), alpha 0, and its endpoint modes with alpha may still decode to 255 everywhere. So this says yes only where the
|
||||
// decoder's own code leaves no other outcome, and no everywhere else (a no keeps the texture BC3, as it was):
|
||||
//
|
||||
// - a void extent: the decoder's IsError takes one for an error when it is the HDR kind (bit 9) or bits 10 and 11 are not both
|
||||
// set, and FillVoidExtentLDR writes the constant's alpha, bits 112 to 127, divided by 65535: 255 only for 0xFFFF. It does
|
||||
// not look at the extent's coordinates, so neither does this.
|
||||
// - any other block, the errors in DecompressBlock's order: the reserved block modes (bits 0 to 3 zero; bits 0 and 1 zero
|
||||
// with bits 6 to 8 set), a weight grid larger than the block, four partitions with two planes, more than 64 weights, the
|
||||
// weights' bits under 24 or over 96, and no room left for the colour data (its count of the remaining bits wraps).
|
||||
// - then the colour endpoint modes. Modes 0, 1, 6, 8 and 9 (luminance and RGB, low dynamic range) give both endpoints the
|
||||
// alpha 255 as a constant (ComputeEndpoints; BlueContract and ClampByte pass it through), and the decoder's interpolation
|
||||
// of two equal 16-bit values, (65535 * (64 - w) + 65535 * w + 32) / 64, is 65535 for every weight, which it writes as
|
||||
// 65535 / 65535 = 1. Modes 4, 5, 10, 12 and 13 carry alpha, and 2, 3, 7, 11, 14 and 15 are the HDR ones, which take the
|
||||
// decoder's other path: no for all of those, without decoding them (such a block may be opaque all the same).
|
||||
// One partition has its mode in bits 13 to 16. Several partitions share one mode when bits 23 and 24 are zero (the mode is
|
||||
// bits 25 to 28).
|
||||
// - (Triforce, 11-21, the performance audit's TEX-50, after Source B's) When bits 23 and 24 are not zero each partition has a
|
||||
// mode of its own, read as the decoder reads it: 2, 5 or 8 more bits for 2, 3 or 4 partitions sit right below the weights
|
||||
// (at 128 less the weights' bits less their own count); with bits 25 to 28 under them they are a class bit a partition,
|
||||
// then two mode bits a partition, and a partition's mode is (bits 23 and 24 as a number, less 1, plus its class bit) * 4
|
||||
// plus its two bits. Yes only when every partition's mode is one of 0, 1, 6, 8 and 9. Those bits count in the decoder's
|
||||
// last error too: 96 bits of weights, the 29 read and 5 or 8 more wrap, and the block is an error. Such blocks were
|
||||
// refused unread, and a texture with one of them stayed BC3. parts_too = false (RX_VK_TEX_BC1_PARTS=0): refused, as before.
|
||||
#pragma once
|
||||
#include <atomic>
|
||||
#include <cstdint>
|
||||
|
||||
namespace astc_opaque {
|
||||
|
||||
/// Whether a block with a mode a partition is looked at (above); false: refused unread, as before. Set once, at the start
|
||||
inline bool parts_too = true;
|
||||
/// How many such blocks were found opaque (the log's)
|
||||
inline std::atomic<uint64_t> parts_passed{0};
|
||||
|
||||
/// Whether the decoder gives both endpoints of a colour endpoint mode the alpha 255: the modes without alpha, low dynamic range
|
||||
constexpr bool opaque_mode(unsigned cem) { return cem == 0 || cem == 1 || cem == 6 || cem == 8 || cem == 9; }
|
||||
|
||||
/// `n` bits (16 at most) of a block from bit `at`, the lowest first, as the decoder's StreamBits reads them
|
||||
inline uint32_t bits(const uint8_t* b, unsigned at, unsigned n) {
|
||||
uint32_t v = 0;
|
||||
for (unsigned i = 0; i < 3 && at / 8 + i < 16; i++) v |= (uint32_t)b[at / 8 + i] << (8 * i);
|
||||
return (v >> (at & 7)) & ((1u << n) - 1);
|
||||
}
|
||||
|
||||
/// The bits `n` weights take in the weight range `index` (0 to 11: 2, 3, 4, 5, 6, 8, 10, 12, 16, 20, 24 and 32 levels), as
|
||||
/// the decoder's GetBitLength counts them: the plain bits, and a trit for every 8/5 of a bit or a quint for every 7/3
|
||||
inline unsigned weight_bits(unsigned n, unsigned index) {
|
||||
static const uint8_t plain[12] = {1, 0, 2, 0, 1, 3, 1, 2, 4, 2, 3, 5};
|
||||
static const uint8_t kind[12] = {0, 1, 0, 2, 1, 0, 2, 1, 0, 2, 1, 0}; // (1: with a trit, 2: with a quint)
|
||||
unsigned total = n * plain[index];
|
||||
if (kind[index] == 1) total += (n * 8 + 4) / 5;
|
||||
if (kind[index] == 2) total += (n * 7 + 2) / 3;
|
||||
return total;
|
||||
}
|
||||
|
||||
/// Whether the decoder writes alpha 255 into every texel of this block of bw x bh texels (see the top of the file). False
|
||||
/// for a block size the decoder is not given (under 4 or over 12 a side)
|
||||
inline bool block(const uint8_t b[16], unsigned bw, unsigned bh) {
|
||||
if (bw < 4 || bh < 4 || bw > 12 || bh > 12) return false;
|
||||
const uint32_t mode = bits(b, 0, 11);
|
||||
if ((mode & 0x1FF) == 0x1FC) { // (a void extent)
|
||||
if (mode & 0x200) return false;
|
||||
if (!(mode & 0x400) || !bits(b, 11, 1)) return false;
|
||||
return b[14] == 0xFF && b[15] == 0xFF;
|
||||
}
|
||||
if (!(mode & 0xF)) return false;
|
||||
if (!(mode & 3) && (mode & 0x1C0) == 0x1C0) return false;
|
||||
// (the weight grid, by the decoder's FindLayout and DecodeBlockSize)
|
||||
const unsigned a = (mode >> 5) & 3, b2 = (mode >> 7) & 3, b1 = (mode >> 7) & 1;
|
||||
unsigned layout, w, h;
|
||||
if (mode & 3) {
|
||||
if (mode & 8) {
|
||||
if (!(mode & 4)) layout = 2, w = a + 2, h = b2 + 8;
|
||||
else if (mode & 0x100) layout = 4, w = b1 + 2, h = a + 2;
|
||||
else layout = 3, w = a + 2, h = b1 + 6;
|
||||
} else if (mode & 4) {
|
||||
layout = 1, w = b2 + 8, h = a + 2;
|
||||
} else {
|
||||
layout = 0, w = b2 + 4, h = a + 2;
|
||||
}
|
||||
} else if (mode & 0x100) {
|
||||
if (!(mode & 0x80)) layout = 9, w = a + 6, h = ((mode >> 9) & 3) + 6;
|
||||
else if (mode & 0x20) layout = 8, w = 10, h = 6;
|
||||
else layout = 7, w = 6, h = 10;
|
||||
} else if (mode & 0x80) {
|
||||
layout = 6, w = a + 2, h = 12;
|
||||
} else {
|
||||
layout = 5, w = 12, h = a + 2;
|
||||
}
|
||||
if (w > bw || h > bh) return false;
|
||||
const bool dual = layout != 9 && (mode & 0x400);
|
||||
const unsigned parts = bits(b, 11, 2) + 1;
|
||||
if (parts == 4 && dual) return false;
|
||||
const unsigned weights = w * h * (dual ? 2 : 1);
|
||||
if (weights > 64) return false;
|
||||
// (the weights' range, by the decoder's DecodeMaxWeight: bit 4 and two more bits, less 2, six further with bit 9)
|
||||
unsigned range = ((mode >> 4) & 1) | (layout < 5 ? (mode & 3) << 1 : (mode & 0xC) >> 1);
|
||||
if (range < 2) return false; // (cannot be: the reserved modes above)
|
||||
range = range - 2 + (layout != 9 && (mode & 0x200) ? 6 : 0);
|
||||
const unsigned wbits = weight_bits(weights, range);
|
||||
if (wbits < 24 || wbits > 96) return false;
|
||||
unsigned read, cem = 0, extra = 0;
|
||||
uint32_t base = 0;
|
||||
if (parts == 1) {
|
||||
cem = bits(b, 13, 4), read = 17;
|
||||
} else {
|
||||
base = bits(b, 23, 6);
|
||||
read = 29;
|
||||
if (base & 3) { // (a mode a partition: its other bits are below the weights)
|
||||
if (!parts_too) return false;
|
||||
extra = parts == 2 ? 2 : parts == 3 ? 5 : 8;
|
||||
} else {
|
||||
cem = base >> 2;
|
||||
}
|
||||
}
|
||||
// (the decoder's last error, its count of the remaining bits wrapping. With one shared mode it cannot: 96 + 29 + 2. With a
|
||||
// mode a partition it can: 96 + 29 + 8)
|
||||
if (wbits + read + extra + (dual ? 2u : 0u) > 128) return false;
|
||||
if (!extra) return opaque_mode(cem);
|
||||
// (the decoder's own order: (extra << 6 | base) >> 2 is a class bit a partition, then two bits a partition)
|
||||
const uint32_t all = ((bits(b, 128 - wbits - extra, extra) << 6) | base) >> 2;
|
||||
for (unsigned i = 0; i < parts; i++) {
|
||||
const unsigned cls = (base & 3) - 1 + ((all >> i) & 1);
|
||||
if (!opaque_mode((cls << 2) | ((all >> (parts + 2 * i)) & 3))) return false;
|
||||
}
|
||||
parts_passed.fetch_add(1, std::memory_order_relaxed);
|
||||
return true;
|
||||
}
|
||||
|
||||
/// Whether every one of `n` blocks (16 bytes each, in any order) is one the decoder writes opaque
|
||||
inline bool all(const uint8_t* blocks, uint64_t n, unsigned bw, unsigned bh) {
|
||||
for (uint64_t i = 0; i < n; i++)
|
||||
if (!block(blocks + 16 * i, bw, bh)) return false;
|
||||
return true;
|
||||
}
|
||||
|
||||
} // namespace astc_opaque
|
||||
133
gpu/vk/astc_reuse.h
Normal file
133
gpu/vk/astc_reuse.h
Normal file
@@ -0,0 +1,133 @@
|
||||
// (Triforce, 11-21, the performance audit's TEX-14, after Source A's) Decoded ASTC used again.
|
||||
//
|
||||
// The game's textures are ASTC, decoded on the GPU and for the most part encoded again as BC3 or BC1: tens of ms of GPU
|
||||
// time and the decoder's scratch for each. When the game leaves an area it finalizes its textures, and when it comes back
|
||||
// it makes them again, mostly at other addresses: the same blocks are decoded and encoded again while the image made of
|
||||
// them the first time may still be in VRAM, kept until the cache needs its room. Source A hashes an ASTC texture's blocks
|
||||
// at its first whole upload, lists a texture the game finalizes under that hash, and gives a new texture with the same
|
||||
// blocks and the same shape that image by a GPU copy, the old one going after it. Its tours of two areas four times at an
|
||||
// 8 GB budget: peak decoded ASTC 3,603 to 2,562 MB, evictions 23,378 to 16,109, the slowest frame 637 to 487 ms.
|
||||
//
|
||||
// What keeps a hit right: 128 bits of hash over the blocks as the decoder is given them; every field of the image that
|
||||
// the decode made and of the storage it was decoded from (Shape); and a source that was uploaded from memory and never
|
||||
// drawn into, still listed (used again at its own address or gone, it is unlisted).
|
||||
//
|
||||
// RX_VK_ASTC_REUSE: 0 or not set, off (the audit: off until measured in play); 2: the blocks are hashed, the finalized
|
||||
// textures listed and each hit counted, and every texture decoded as before: what a reuse would save, for no risk;
|
||||
// 1: the hit's image is copied.
|
||||
//
|
||||
// No Vulkan here: tests/test_astc_reuse.py compiles it.
|
||||
#pragma once
|
||||
#include <cstdint>
|
||||
#include <cstdlib>
|
||||
#include <cstring>
|
||||
#include <unordered_map>
|
||||
|
||||
#include "upload_same.h" // (content_hash: the first of the two lanes)
|
||||
|
||||
namespace astc_reuse {
|
||||
|
||||
enum Mode { kOff = 0, kCopy = 1, kCount = 2 };
|
||||
/// RX_VK_ASTC_REUSE's mode: 1 and 2 as they are, anything else off
|
||||
inline Mode mode(const char* asked) {
|
||||
if (!asked || !*asked) return kOff;
|
||||
const int v = atoi(asked);
|
||||
return v == 1 ? kCopy : v == 2 ? kCount : kOff;
|
||||
}
|
||||
|
||||
/// A second hash of n bytes, of other constants and another mix than upload_same::content_hash's: eight chains over 64
|
||||
/// bytes a step, as there, so that both read at memory's speed
|
||||
inline uint64_t lane2(const void* p, uint64_t n) {
|
||||
const uint8_t* b = static_cast<const uint8_t*>(p);
|
||||
const auto rotl = [](uint64_t x, int r) { return (x << r) | (x >> (64 - r)); };
|
||||
uint64_t h = 0xD6E8FEB86659FD93ull + n * 0x9FB21C651E98DF25ull, i = 0;
|
||||
if (n >= 512) {
|
||||
uint64_t c[8];
|
||||
for (int k = 0; k < 8; k++) c[k] = h + 0xA0761D6478BD642Full * (uint64_t)(2 * k + 1);
|
||||
for (; i + 64 <= n; i += 64) {
|
||||
uint64_t w[8];
|
||||
std::memcpy(w, b + i, 64);
|
||||
for (int k = 0; k < 8; k++) c[k] = rotl(c[k] + w[k] * 0xE7037ED1A0B428DBull, 29) * 0x8EBC6AF09C88C6E3ull;
|
||||
}
|
||||
for (int k = 0; k < 8; k++) h = rotl(h ^ c[k], 23) * 0x589965CC75374CC3ull + 0x1D8E4E27C47D124Full;
|
||||
}
|
||||
for (; i + 8 <= n; i += 8) {
|
||||
uint64_t w;
|
||||
std::memcpy(&w, b + i, 8);
|
||||
h = rotl(h + w * 0xE7037ED1A0B428DBull, 29) * 0x8EBC6AF09C88C6E3ull;
|
||||
}
|
||||
for (; i < n; i++) h = rotl(h ^ b[i], 11) * 0x589965CC75374CC3ull;
|
||||
h ^= h >> 32, h *= 0xD6E8FEB86659FD93ull, h ^= h >> 32;
|
||||
return h;
|
||||
}
|
||||
|
||||
/// The blocks' hash, 128 bits, made of pieces given in order (a big texture's storage is read 64 MB at a time): each piece
|
||||
/// goes through both lanes kStep bytes at a time, while they are in the cache, and is folded in with its length
|
||||
struct Hash {
|
||||
uint64_t a = 0, b = 0;
|
||||
bool operator==(const Hash& o) const { return a == o.a && b == o.b; }
|
||||
explicit operator bool() const { return (a | b) != 0; }
|
||||
};
|
||||
inline constexpr uint64_t kStep = 256u << 10;
|
||||
struct Hasher {
|
||||
Hash h;
|
||||
uint64_t bytes = 0;
|
||||
void add(const uint8_t* p, uint64_t n) {
|
||||
for (uint64_t at = 0; at < n; at += kStep) {
|
||||
const uint64_t m = n - at < kStep ? n - at : kStep;
|
||||
h.a = (h.a ^ upload_same::content_hash(p + at, m)) * 0x9E3779B97F4A7C15ull + m;
|
||||
h.b = (h.b + lane2(p + at, m)) * 0xC2B2AE3D27D4EB4Full ^ (m << 1 | 1);
|
||||
}
|
||||
bytes += n;
|
||||
}
|
||||
/// The hash, never all zero (zero is "not hashed")
|
||||
Hash done() const { return (h.a | h.b) ? h : Hash{1, 0}; }
|
||||
};
|
||||
|
||||
/// What has to be the same besides the blocks for one image to stand for another: the image the decode made (its format,
|
||||
/// size, levels and layers, the top levels left out and whether by the card's texture detail, kept as BC3 or BC1 or
|
||||
/// RGBA8, BC1 given up for BC3, resized for AMD) and the storage it was decoded from (the NVN format, the key's layout)
|
||||
struct Shape {
|
||||
uint32_t format, width, height, depth, levels, layers, skip, nvn_format, layout;
|
||||
bool bc, bc_resized, bc1, bc1_no, detail_skip;
|
||||
bool operator==(const Shape& o) const {
|
||||
return format == o.format && width == o.width && height == o.height && depth == o.depth && levels == o.levels &&
|
||||
layers == o.layers && skip == o.skip && nvn_format == o.nvn_format && layout == o.layout && bc == o.bc &&
|
||||
bc_resized == o.bc_resized && bc1 == o.bc1 && bc1_no == o.bc1_no && detail_skip == o.detail_skip;
|
||||
}
|
||||
};
|
||||
|
||||
/// The finalized textures whose image may serve again, by their blocks' hash. T is the cache's texture; one is in the list
|
||||
/// at most once: listed and unlisted by its pointer
|
||||
template <class T>
|
||||
class Freed {
|
||||
public:
|
||||
void list(const Hash& hash, T* t) { by_hash.emplace(hash.a, Entry{hash.b, t}); }
|
||||
void unlist(const Hash& hash, T* t) {
|
||||
auto [a, b] = by_hash.equal_range(hash.a);
|
||||
for (auto it = a; it != b; ++it)
|
||||
if (it->second.t == t) {
|
||||
by_hash.erase(it);
|
||||
return;
|
||||
}
|
||||
}
|
||||
/// One listed with this hash, both lanes, whose shape is `shape` (shape_of(T*) gives each one's) and that is not
|
||||
/// `self`; null: none
|
||||
template <class ShapeOf>
|
||||
T* find(const Hash& hash, const Shape& shape, const T* self, ShapeOf shape_of) const {
|
||||
auto [a, b] = by_hash.equal_range(hash.a);
|
||||
for (auto it = a; it != b; ++it)
|
||||
if (it->second.b == hash.b && it->second.t != self && shape_of(it->second.t) == shape) return it->second.t;
|
||||
return nullptr;
|
||||
}
|
||||
size_t size() const { return by_hash.size(); }
|
||||
|
||||
private:
|
||||
struct Entry {
|
||||
uint64_t b;
|
||||
T* t;
|
||||
};
|
||||
std::unordered_multimap<uint64_t, Entry> by_hash;
|
||||
};
|
||||
|
||||
} // namespace astc_reuse
|
||||
137
gpu/vk/async_rules.h
Normal file
137
gpu/vk/async_rules.h
Normal file
@@ -0,0 +1,137 @@
|
||||
// (Triforce) The replay thread's rules (RX_VK_ASYNC=1, off by default): what it took to make the thread a place frames can be
|
||||
// replayed on without the faults the lab's runs of it showed. Nothing here releases the game early: the game still presents only
|
||||
// once the GPU has run its frame (its own event poll), so the picture and its order are the submitting thread's replay's.
|
||||
//
|
||||
// What the runs showed (rounds 22 and 23, vk_replay.cpp's own counters): the thread took the replay of the game's main thread's
|
||||
// two queues off the part of the frame both of the game's chains wait for (about 1.3 ms at the Market), and the frame got only 0.4
|
||||
// to 0.8 ms shorter, because
|
||||
// 1. a capture was taken at every signal of the CPU's while the thread had any work (26,654 in 26,400 frames, 0.25 ms each on
|
||||
// the game's thread), where the submitting thread's replay needs one only when a queue really goes on from a wait for that
|
||||
// event (126 to 1,028 a run);
|
||||
// 2. the submit's capture made every copy on the game's thread, where the submitting thread's replay leaves them to the copy
|
||||
// helpers;
|
||||
// 3. the game's submit gave its virtual core up and took it again around a hand-over of a quarter millisecond;
|
||||
// 4. the game's present poll slept in steps of the system's timer, about 0.6 ms late on average;
|
||||
// 5. every present made the stopped queues go past their waits (5,649 to 15,166 waits gone past a run against 0), which is
|
||||
// also the first suspect of the flicker after a tour.
|
||||
// The rules below are 1 and 5, apart from Vulkan and from the replay's state so that tests/test_async_rules.py can run them; 2 to
|
||||
// 4 are in vk_replay.cpp submit_q, host/nvn_native.c nvnQueueSubmitCommands and host/hos_thread.c present_poll. Each has a switch
|
||||
// of its own, and all of it is behind RX_VK_ASYNC=1.
|
||||
#pragma once
|
||||
|
||||
#include <algorithm>
|
||||
#include <cstdint>
|
||||
#include <cstdlib>
|
||||
#include <deque>
|
||||
#include <vector>
|
||||
|
||||
namespace mkwvk::async_rules {
|
||||
|
||||
/// A switch that is on unless set to 0 (RX_VK_ASYNC_EVENT_NOTES, RX_VK_ASYNC_CAPTURE_BEHIND)
|
||||
inline bool on_unless_zero(const char* env) {
|
||||
return !env || !*env || std::atoi(env) != 0;
|
||||
}
|
||||
|
||||
// ---- 1. the capture at a signal of the CPU's ----
|
||||
// The game writes what a queue's work reads after it submitted that work, up to its signal of the event the work waits for; the
|
||||
// console's GPU reads memory after its wait. With the replay on its own thread that capture has to be taken at the signal, on the
|
||||
// game's thread (later the game has written on). It is needed only when a submission handed over before the signal holds a wait
|
||||
// for that event that the replay has not passed: every submission handed over after the signal has its own capture, taken after
|
||||
// the game's writes. The host tells the events a submission waits for (its walk at the submit sees the wait ops:
|
||||
// host/nvn_native.c resolve_deferred); a submission is noted from its hand-over until it is replayed whole.
|
||||
|
||||
constexpr uint64_t kAnyEvent = ~0ull; // (a submission whose waits the host did not tell: it may wait for any event)
|
||||
constexpr size_t kMaxNotes = 256; // (more than this pending: every signal takes its capture, as before, until those past it are done)
|
||||
|
||||
struct Notes {
|
||||
struct Note {
|
||||
uint64_t value; // (the submission's timeline value: its hand-over's)
|
||||
std::vector<uint64_t> events; // (the events it waits for; kAnyEvent: not told)
|
||||
};
|
||||
std::deque<Note> pending; // (in hand-over order)
|
||||
bool over = false; // (a submission past kMaxNotes is not replayed whole yet: waits_for answers yes)
|
||||
// ((Triforce) the submissions handed over past the bound, by value: their events are not known, so every event counts until
|
||||
// each of them is done. The flag fell once `pending` was empty, which could be while one past the bound was still stopped
|
||||
// at its wait, and its signal then took no capture)
|
||||
std::vector<uint64_t> unnoted;
|
||||
|
||||
/// A submission handed over. told: the host gave its waits (events, n; n 0: it waits for none, and nothing is noted)
|
||||
void add(uint64_t value, bool told, const uint64_t* events, int n) {
|
||||
if (told && n <= 0) return;
|
||||
if (pending.size() >= kMaxNotes) {
|
||||
over = true;
|
||||
unnoted.push_back(value);
|
||||
return;
|
||||
}
|
||||
Note note{value, {}};
|
||||
if (!told) note.events.push_back(kAnyEvent);
|
||||
else
|
||||
for (int i = 0; i < n; i++)
|
||||
if (std::find(note.events.begin(), note.events.end(), events[i]) == note.events.end()) note.events.push_back(events[i]);
|
||||
pending.push_back(std::move(note));
|
||||
}
|
||||
/// The submission of that value replayed whole (none of it stopped at a wait any more)
|
||||
void done(uint64_t value) {
|
||||
for (auto it = pending.begin(); it != pending.end(); ++it)
|
||||
if (it->value == value) {
|
||||
pending.erase(it);
|
||||
break;
|
||||
}
|
||||
for (auto it = unnoted.begin(); it != unnoted.end(); ++it)
|
||||
if (*it == value) {
|
||||
unnoted.erase(it);
|
||||
break;
|
||||
}
|
||||
if (unnoted.empty()) over = false;
|
||||
}
|
||||
/// Whether a submission handed over and not replayed whole may hold a wait for the event
|
||||
bool waits_for(uint64_t event) const {
|
||||
if (over) return true;
|
||||
for (const Note& note : pending)
|
||||
for (const uint64_t e : note.events)
|
||||
if (e == event || e == kAnyEvent) return true;
|
||||
return false;
|
||||
}
|
||||
};
|
||||
|
||||
/// Whether a signal of the CPU's takes a capture on the game's thread. behind: the replay thread has work handed over and not
|
||||
/// recorded. notes_on: RX_VK_ASYNC_EVENT_NOTES (off: every signal with the thread behind, as before)
|
||||
inline bool capture_at_signal(bool behind, bool notes_on, const Notes& notes, uint64_t event) {
|
||||
if (!behind) return false;
|
||||
return !notes_on || notes.waits_for(event);
|
||||
}
|
||||
|
||||
// ---- 5. what a present on the replay thread does with the queues stopped at a wait ----
|
||||
// Force: every stopped queue past its wait (as it was by default: the waits gone past). Keep: nothing forced (a load stood for
|
||||
// good with it: only a present makes a load's stopped work go). Stale: the jobs behind the present are run first, a few
|
||||
// milliseconds at most, until the frame's stopped work has gone on, and then only what was already stopped at the present before
|
||||
// is made to go (vk_replay.cpp drain_stale). RX_VK_PRESENT_DRAIN: unset or 2 stale, 1 force, 0 keep
|
||||
enum class PresentDrain { Keep = 0, Force = 1, Stale = 2 };
|
||||
inline PresentDrain present_drain(const char* env) {
|
||||
if (!env || !*env) return PresentDrain::Stale;
|
||||
const int v = std::atoi(env);
|
||||
return v == 0 ? PresentDrain::Keep : v == 1 ? PresentDrain::Force : PresentDrain::Stale;
|
||||
}
|
||||
|
||||
// ---- what is still open (2026-11-19, lab rounds 52 to 60; the thread stays off by default) ----
|
||||
// Standing at the Market the thread gave 104.3 fps for 97.5 and a 1% low of 81.7 for 75.2, with no flicker (3 runs each). After
|
||||
// a ten-area tour the player's character is drawn black for one frame, 0 to 3 times a minute at the Market (the default: none in
|
||||
// 2 tours); with the presents forcing again (RX_VK_PRESENT_DRAIN=1) 37 to 111 times, so rule 5 is what keeps it rare. Every pass
|
||||
// and dispatch of such a frame is there with its draws: the character is lit from wrong inputs, in a frame after one whose replay
|
||||
// ran late. Ruled out: these rules (the thread without them does it too), texture uploads (none in three of four such frames),
|
||||
// a readback handing the game a word it had set to 0 (none at 43 such frames: RX_VK_LOG_READBACK_ZERO), the conditions on
|
||||
// occlusion counts and the occlusion answers (RX_VK_OCC_COND=0, RX_VK_OCC_QUERY=0: still there), and the game's request flags
|
||||
// (all read set around eight such frames). Not settled: whether the lab's frame log, whose readback and disk writes are on this
|
||||
// thread, is what makes the replay late there (one tour each: 3 such frames with its writes, none with RX_VK_FRAMELOG_DROP=1).
|
||||
// RX_VK_REGION_TRAP=1 with RX_VK_PASS_TRACE=1 and the frame log logs the passes and conditions around the first eight.
|
||||
|
||||
// ---- the replay thread's look for its next job ----
|
||||
/// RX_VK_ASYNC_SPIN_US: how long the thread looks for a job before it sleeps, in nanoseconds: 100 us unset, 0 none, 1000 at most.
|
||||
/// The thread is on the core kept from the game's threads, and a wake of a sleeping thread is tens of microseconds for each of a
|
||||
/// frame's hand-overs
|
||||
inline int64_t spin_ns(const char* env) {
|
||||
if (!env || !*env) return 100 * 1000;
|
||||
return (int64_t)std::clamp(std::atoi(env), 0, 1000) * 1000;
|
||||
}
|
||||
|
||||
} // namespace mkwvk::async_rules
|
||||
197
gpu/vk/barrier_rules.h
Normal file
197
gpu/vk/barrier_rules.h
Normal file
@@ -0,0 +1,197 @@
|
||||
// (Triforce) Where the replay records less than a full barrier between work the game left unordered: a part's split inside a
|
||||
// pass, an event's signal, a dispatch right after another; and when held submits go to the queue (RX_VK_BATCH). Apart from
|
||||
// Vulkan so that tests/test_barrier_rules.py can run the rules over recorded command streams and check every hazard the game's
|
||||
// ordering asks for against what the recorded barriers order (and, a rule dropped at a time, that a hazard is then left open).
|
||||
//
|
||||
// What the game's ordering asks for (NVN's, as the replay keeps it): work after one of the game's barriers after the work before
|
||||
// it; a draw after a dispatch and a dispatch after a draw (the console's 3D and compute work share the engine and drain at the
|
||||
// switch); copies, clears and uniform updates after and before what they touch; the colour, depth and stencil outputs of draws
|
||||
// in the order they were drawn, from pass to pass too; a draw that samples a target its pass wrote, or writes one its pass
|
||||
// sampled, after the draws before (the replay ends the pass there itself: vk_draw.cpp draw_hazards); an event's word written
|
||||
// after the work before its signal, and a wait's work after the work before the signal it waits for. A signal is no barrier on
|
||||
// the console: the work after it in the same queue is not ordered after the work before it.
|
||||
//
|
||||
// 1. A part's split (RX_VK_SPLIT_BARRIER; vk_replay.cpp maybe_split). A long submission goes to the GPU in parts of 120 draws and
|
||||
// dispatches (RX_VK_CHUNK) so that the GPU starts early. Each part ended its pass with a full barrier, and the next began with
|
||||
// another: the GPU drained between two draws of one of the game's passes, about 27 times a frame. With the rule the open pass
|
||||
// goes on in the next part: the part ends it, and the next begins it again with the same targets behind a barrier of the
|
||||
// attachments' outputs alone (colour output and the depth and stencil tests, their writes before their reads and writes), and
|
||||
// keeps what the pass's draws sampled and wrote, so that draw_hazards ends the pass where it would have without the split. That
|
||||
// orders all that the same draws in one pass ordered (RX_VK_CHUNK=0): within a pass Vulkan orders the attachments' accesses
|
||||
// alone. Only where nothing else was recorded at the split (the part's uniform updates, its write-backs, an event's write held:
|
||||
// then the full barriers as before), and a pass open. A part's uploads keep their own full barriers (vk_texture.cpp upload_cmd,
|
||||
// texture_uploads_recorded), and anything recorded or submitted between the part's submit and the pass's start again (another
|
||||
// submitter's work on the queue) gives the next part a full barrier first. Unset: on with NVIDIA's driver; =pass: on any card;
|
||||
// =full (or any other value): as before. ((Triforce, P4) A parallel replay's segments are never split: they are the parts there,
|
||||
// each begun and ended with a full barrier (vk_replay.cpp record_segment), so nothing goes on across a segment's edge; the rule
|
||||
// acts in the serial replay's parts, as before. "Recorded between" is the replaying thread's own commands and every submit to the
|
||||
// queue (cmd_work.h work_between); RX_VK_BARRIER_CONTEXT=0: every thread's commands, as before)
|
||||
// 2. An event's signal (RX_VK_EVENT_BARRIER; vk_replay.cpp NOP_SIGNAL_EVENT). Each signal ended the pass and drained the GPU before
|
||||
// its word's write. With the rule nothing is recorded at the signal: its word's buffer is bound there (so its page is among the
|
||||
// recording's write-backs, as before), and the write waits for the next full barrier on its command buffer (the game's next
|
||||
// barrier, the next dispatch's or pass start's) and goes right after it, so it still follows all the work before the signal;
|
||||
// a recording's end writes what is still held behind a full barrier first, before its write-backs copy the words to the game's
|
||||
// memory (that barrier is recorded under the pools' lock, so the writes take none: their buffers were bound at the signals). The
|
||||
// writes count as work for the next barrier's look, so the write-backs' barrier is never left out after them. The writes held
|
||||
// for one barrier go in the order of their signals, a barrier of the transfers between each two, so the game sees them done in
|
||||
// that order (two signals of one word too). A wait met in the replay's order whose event's write is still held gets a full
|
||||
// barrier first: the waiting queue's work is then after the signalling queue's work before the signal, as the drain at the
|
||||
// signal made it. The game sees the word later (at that barrier, not at the signal). NX2navi records nothing at a signal at all
|
||||
// (its CPU writes the word once the submission is done). Unset: with the pass barrier's game mode (NVIDIA's driver,
|
||||
// pass_barrier.h); =late: on any card; =full (or any other value): as before. Not with the parallel replay.
|
||||
// 3. A dispatch right after another (RX_VK_DISPATCH_BARRIER; vk_draw.cpp dispatch). Off unless asked for: =game leaves out the
|
||||
// full barrier of a dispatch whose command buffer's last command was the dispatch before it (no barrier of the game's, no copy,
|
||||
// no draw between; not an indirect one, whose grid another dispatch may have written), as NX2navi's barrier_game does. The game
|
||||
// orders its dependent dispatches with its own barriers on the console; whether it always does, only the game can show (a
|
||||
// one-frame band or flicker where it does not), so it is a switch for the lab. (Triforce, wp14, P4) Beside the parallel replay too:
|
||||
// "nothing between" is the command buffer's own count of its commands (cmd_work.h), and a segment is a command buffer of its own;
|
||||
// with the global count the rule stood aside while the helper threads recorded, which at the Market is most frames
|
||||
// (RX_VK_BARRIER_CONTEXT=0: so again). The event rule above still does (event_hold).
|
||||
// 4. Held submits (RX_VK_BATCH; vk_replay.cpp submit_recording, submit_job). =1 (on by default with Mesa's RADV) holds every submit
|
||||
// until something needs the GPU's results or the queue's order. =parts holds a whole submission's recording only while the game's
|
||||
// submit is being replayed: a part (split off to start the GPU early) goes to the queue at once with what is held, so does the
|
||||
// held when there are RX_VK_BATCH_MAX (4 here) of them, and what is held goes when the replay of the game's submit returns.
|
||||
// Holding changes neither the command buffers nor their order on the queue, only how many vkQueueSubmit2 calls carry them.
|
||||
#pragma once
|
||||
|
||||
#include <cstddef>
|
||||
#include <cstdint>
|
||||
#include <cstdlib>
|
||||
#include <cstring>
|
||||
|
||||
namespace mkwvk::barrier_rules {
|
||||
|
||||
/// RX_VK_SPLIT_BARRIER: "pass" the rule on any card; "full" (or any other value) a full barrier at each split, as before; unset or
|
||||
/// empty: on with NVIDIA's driver (vendor 0x10de). every_draw: RX_VK_BARRIER_EVERY_DRAW (an experiment that ends the pass at each
|
||||
/// draw) keeps the old way
|
||||
inline bool split_in_pass(const char* env, uint32_t vendor, bool every_draw) {
|
||||
if (every_draw) return false;
|
||||
if (env && *env) return std::strcmp(env, "pass") == 0;
|
||||
(void)vendor; // ((Triforce, 11-16) off unless asked: the lab's Market had the same ~500 full barriers a frame with it on (the
|
||||
return false; // helper threads record most frames there, and the rule stands aside for them) and no gain (90.6 / 89.2 fps
|
||||
// against 92.8 / 90.3 with the old barriers)
|
||||
}
|
||||
/// ((Triforce, 11-22, the performance audit's REPLAY-14, after Source A's) Two workarounds for Intel's drivers at a submission's
|
||||
/// splits (vk_replay.cpp maybe_split and the dry pass's points). No other vendor has either unless its switch asks.
|
||||
/// RX_VK_SPLIT_BEFORE_DRAW: a part ends only right before a plain draw. On an Arc B580 a command buffer that begins with mesh
|
||||
/// draws drew its later plain draws as exploded triangles (Source A: 40 shard frames of 1,121 before, 0 of 1,101 after). Unset:
|
||||
/// on for Intel's GPUs (vendor 0x8086); 0 off; anything else on, on any GPU
|
||||
inline bool split_before_draw(const char* env, uint32_t vendor) {
|
||||
if (env && *env) return std::strcmp(env, "0") != 0;
|
||||
return vendor == 0x8086;
|
||||
}
|
||||
/// RX_VK_SPLIT_UBFLUSH: whether a part's end writes the submission's uniform updates so far into the game's memory. Without
|
||||
/// mirrors they land in the memory later parts' other reads and the game's CPU see, mid-submission, where an unsplit submission
|
||||
/// writes them once at its end: white blocks and flashes on Intel's integrated GPUs (Source A: 1 run of 3 before, 0 of 3
|
||||
/// after). Unset: on, as before, but off on Intel's GPUs without mirrors; 0 off; anything else on
|
||||
inline bool split_ub_flush(const char* env, uint32_t vendor, bool mirrors) {
|
||||
if (env && *env) return std::strcmp(env, "0") != 0;
|
||||
return !(vendor == 0x8086 && !mirrors);
|
||||
}
|
||||
/// A part may end right before this op: a plain draw (NOP_DRAW, _IDX, _INDIRECT: not a mesh draw, not a dispatch) that will
|
||||
/// draw (rendering on, no rasterizer discard)
|
||||
constexpr bool cut_ok(bool plain_draw, bool render_enabled, bool raster_discard) {
|
||||
return plain_draw && render_enabled && !raster_discard;
|
||||
}
|
||||
/// Whether a part's split goes on in the same pass: the rule on, a pass open at the split, nothing recorded at the part's end (quiet:
|
||||
/// the work count as it was after the pass's end) and no event's write held then (written behind a full barrier before the part's
|
||||
/// write-backs, which copy its word to the game's memory: the GPU drained there all the same). ((Triforce, P4) no parallel replay's
|
||||
/// argument: its segments are never split, maybe_split runs in the serial replay only)
|
||||
inline bool split_carry(bool on, bool pass_open, bool quiet, bool writes_held) {
|
||||
return on && pass_open && quiet && !writes_held;
|
||||
}
|
||||
/// Whether the next part starts with a full barrier: always when the pass is not carried, and when it is but something was recorded
|
||||
/// between the part's submit and the pass's start again (into its command buffer, by this thread into another, or submitted to the
|
||||
/// queue by any thread: cmd_work.h work_between)
|
||||
inline bool split_start_full(bool carry, bool recorded_between) {
|
||||
return !carry || recorded_between;
|
||||
}
|
||||
|
||||
/// ((Triforce, 11-22, the performance audit's SYNC-66, after Source B's) RX_VK_CHUNK: how many draws and dispatches a part of a
|
||||
/// long submission has (vk_replay.cpp maybe_split). 120 on every card since the parts were made: the GPU starts early, and each
|
||||
/// split costs a queue submit, two full barriers and the part's bookkeeping, about 27 times a frame in a serial replay. Source B
|
||||
/// cuts at 128 where the GPU idles while the CPU records and at 4096 on desktop drivers; what pays here is not measured on any
|
||||
/// driver, so the default by driver is 120 for each until the lab picks one (S1 and S2 with RX_VK_PARALLEL=0 and the lab knob
|
||||
/// chunk:<n>). vendor: the card's PCI vendor; radv: Mesa's RADV
|
||||
constexpr int64_t kPartDraws = 120;
|
||||
inline int64_t part_draws_default(uint32_t vendor, bool radv) {
|
||||
(void)vendor, (void)radv;
|
||||
return kPartDraws;
|
||||
}
|
||||
/// The size in force: RX_VK_CHUNK's number when the variable is set (0 or less: whole submissions, never split); else
|
||||
/// knob_or_default, the lab knob chunk:<n>'s number in its phases and the driver's default outside them
|
||||
inline int64_t part_draws(bool env_set, int64_t env_value, int64_t knob_or_default) {
|
||||
return env_set ? env_value : knob_or_default;
|
||||
}
|
||||
/// As the size was read before: the variable's number, 120 unset
|
||||
inline int64_t part_draws_before(const char* env) { return env ? std::atoll(env) : 120; }
|
||||
|
||||
/// RX_VK_EVENT_BARRIER: "late" the rule on any card; "full" (or any other value) as before; unset or empty: with the pass barrier's
|
||||
/// game mode (pass_barrier::game: NVIDIA's driver unless RX_VK_PASS_BARRIER says otherwise)
|
||||
inline bool event_late(const char* env, bool pass_game) {
|
||||
if (env && *env) return std::strcmp(env, "late") == 0;
|
||||
(void)pass_game; // ((Triforce, 11-16) off unless asked, as the split rule above: no gain measured)
|
||||
return false;
|
||||
}
|
||||
/// Whether a signal's write is held for the next full barrier: the rule on, no parallel replay (its segments' threads share the
|
||||
/// command buffers), and room for it (most: the held writes a command buffer may have)
|
||||
inline bool event_hold(bool on, bool parallel, int held, int most) {
|
||||
return on && !parallel && held < most;
|
||||
}
|
||||
/// A wait the replay finds met: a full barrier first when the write of its event's signal is still held
|
||||
inline bool wait_barrier(bool its_write_held) {
|
||||
return its_write_held;
|
||||
}
|
||||
|
||||
/// RX_VK_DISPATCH_BARRIER: "game" the barrier of a dispatch right after another left out; anything else, or unset: as before
|
||||
inline bool dispatch_game(const char* env) {
|
||||
return env && std::strcmp(env, "game") == 0;
|
||||
}
|
||||
/// Whether a dispatch records its full barrier. work_since: GPU work recorded on the command buffer since its start (vk_draw.cpp
|
||||
/// work_since_barrier); no_order: RX_VK_NO_DISPATCH_ORDER; right_after: the command buffer's last command was a dispatch's (by the
|
||||
/// command buffer's own count, cmd_work.h: with the parallel replay too)
|
||||
inline bool dispatch_barrier(bool work_since, bool no_order, bool game, bool right_after, bool indirect) {
|
||||
if (!work_since || no_order) return false;
|
||||
return !(game && right_after && !indirect);
|
||||
}
|
||||
|
||||
/// RX_VK_BATCH
|
||||
enum class Batch { Off, Hold, Parts };
|
||||
/// "parts": Parts; otherwise set: Hold where its number is not 0, Off where it is (as RX_VK_BATCH was read); unset: Hold with RADV,
|
||||
/// Off elsewhere
|
||||
inline Batch batch_mode(const char* env, bool radv) {
|
||||
if (env) {
|
||||
if (std::strcmp(env, "parts") == 0) return Batch::Parts;
|
||||
return std::atoi(env) != 0 ? Batch::Hold : Batch::Off;
|
||||
}
|
||||
return radv ? Batch::Hold : Batch::Off;
|
||||
}
|
||||
/// Whether a recording is held rather than handed to the queue: Hold always; Parts only a whole submission's (not a part's), and
|
||||
/// only inside the replay of a game's submit (in_submit: submit_job's)
|
||||
inline bool batch_hold(Batch m, bool part, bool in_submit) {
|
||||
if (m == Batch::Hold) return true;
|
||||
return m == Batch::Parts && !part && in_submit;
|
||||
}
|
||||
/// Whether what is held goes to the queue now, after a submit was added to it (held: recordings held now, the added one too): Hold at
|
||||
/// cap (RX_VK_BATCH_MAX, 32) as before; Parts at most (RX_VK_BATCH_MAX, 4) or when the added one was not held (a part: it goes with
|
||||
/// those before it, in order)
|
||||
inline bool batch_flush(Batch m, bool added_held, size_t held, size_t cap) {
|
||||
if (m == Batch::Off) return true;
|
||||
if (m == Batch::Parts && !added_held) return true;
|
||||
return held >= cap;
|
||||
}
|
||||
/// ((Triforce, 11-22, the performance audit's SYNC-09, after Source A's) RX_VK_BATCH_IDLE: a submit that joins the held batch
|
||||
/// sends it at once when the GPU has finished every batch sent. A held batch otherwise waits for a wait, a present, the 2 ms
|
||||
/// poll or the cap of 32 while the GPU sits idle beside it, until the game waits mid-frame at a capture's late pages; while the
|
||||
/// GPU is busy, submits still group. Source A has it on wherever batching is (RADV), off where guest memory stayed host-pointer
|
||||
/// imports (each submit there pays amdgpu's page check, 15-45 ms). HERE: on only with a number other than 0; unset it is OFF,
|
||||
/// on every driver, until a Linux RADV machine has run it (the audit's lab proof; with NVIDIA's driver and AMD's and Intel's on
|
||||
/// Windows nothing is held). No guard for host-pointer imports is built: with the switch off unless asked there is nothing for it to guard
|
||||
inline bool batch_idle_on(const char* env) { return env && std::atoi(env) != 0; }
|
||||
/// Whether a held batch is looked at for an idle GPU after a submit joined it: the switch on, Hold mode only (Parts sends by
|
||||
/// its own rule, when the game's submit returns), and something held (the cap's flush may just have sent it all)
|
||||
inline bool batch_idle(Batch m, bool held, bool on) {
|
||||
return on && m == Batch::Hold && held;
|
||||
}
|
||||
|
||||
} // namespace mkwvk::barrier_rules
|
||||
66
gpu/vk/bc_scratch.h
Normal file
66
gpu/vk/bc_scratch.h
Normal file
@@ -0,0 +1,66 @@
|
||||
// (Triforce, 11-21, the performance audit's TEX-10, after Source A's) How many layers the BC3 scratch has.
|
||||
//
|
||||
// Big ASTC textures are decoded into a 2048 x 2048 RGBA8 scratch and encoded again as BC3 on the GPU, into an output
|
||||
// buffer. Both were made with 8 layers, 128 + 32 = 160 MB (and the second scratch of the textures encoded at another
|
||||
// size, 128 MB more), whatever a run used; and a run decodes one layer (RX_VK_BC_LAYERS, the dark-line fix), so 140 MB
|
||||
// were held and never written, for good once made (RX_VK_BC_IDLE_S is 0). Sized by the layers one run uses they are
|
||||
// 16 + 4 = 20 MB. Source A measured 139 MiB less GPU memory on a Steam Deck, the median of 7 pairs, at the same fps.
|
||||
// RX_VK_BC_SCRATCH_FULL=1: all 8 layers, as before.
|
||||
//
|
||||
// The rule alone, no Vulkan: tests/test_bc_scratch_layers.py compiles it.
|
||||
#pragma once
|
||||
#include <algorithm>
|
||||
#include <cstdint>
|
||||
#include <cstdlib>
|
||||
|
||||
namespace bc_scratch {
|
||||
|
||||
inline constexpr uint32_t kDim = 2048, kMost = 8; // (the largest level a BC3 texture may have; the most layers a run may take)
|
||||
|
||||
/// The layers one run decodes and encodes: RX_VK_BC_LAYERS's value, 1 to kMost; not set, or not a number, 1
|
||||
inline uint32_t run_layers(const char* asked) {
|
||||
return asked ? std::clamp<uint32_t>((uint32_t)std::max(atoi(asked), 0), 1u, kMost) : 1u;
|
||||
}
|
||||
/// The scratch's layers: a run's, or all of them when full (RX_VK_BC_SCRATCH_FULL)
|
||||
inline uint32_t layers(uint32_t run, bool full) { return full ? kMost : std::clamp(run, 1u, kMost); }
|
||||
/// The output buffer's bytes: BC3's 16 bytes a block of 4 x 4, every layer
|
||||
inline uint64_t out_bytes(uint32_t layers) { return (uint64_t)(kDim / 4) * (kDim / 4) * layers * 16; }
|
||||
/// What the scratch holds, in MB: the image and its output buffer, and the second image when there is one
|
||||
inline uint64_t mb(uint32_t layers, bool second) {
|
||||
const uint64_t image = (uint64_t)kDim * kDim * 4 * layers;
|
||||
return (image + out_bytes(layers) + (second ? image : 0)) >> 20;
|
||||
}
|
||||
|
||||
// ---- (Triforce, 11-21, the performance audit's TEX-09, after Source A's) the small scratch ----
|
||||
// A run is a decode, an encode, a copy and three full barriers, and one layer a run made a 1365 x 1365 x 90 terrain array
|
||||
// of 11 levels about 990 of them: the GPU spent most of that time between the runs (Source A, Hyrule Field's entry on an
|
||||
// RX 9070 XT: 50 ms of GPU an array at one layer a run, 16-17 ms at 8; the lab's own array took 114.8 ms of an RTX 5090 in
|
||||
// one upload command buffer). Levels of kSmallDim or less each way, the arrays' second level down, go small_layers() layers
|
||||
// a run through a scratch of their own, kSmallDim x kSmallDim x 8: 32 MB, and the output buffer 8 MB for 4. One layer a
|
||||
// run was for dark lines between Hyrule Field's ground tiles on an RX 6600, layers after a run's first wrong at the small
|
||||
// levels, cause never found: so on AMD's cards the default stays 1 until the picture is checked there, and each layer of a
|
||||
// run is copied by a region of its own, at the encoder's own layer stride. A level of a texture encoded at another size
|
||||
// keeps the first scratch. RX_VK_BC_SMALL_LAYERS=<n>: the layers a run, 1 to 8; 1: no small scratch, as before.
|
||||
inline constexpr uint32_t kSmallDim = 1024;
|
||||
|
||||
/// The small scratch's layers, a run's of a small level: RX_VK_BC_SMALL_LAYERS, 1 to kMost; not set, kMost, and 1 on AMD's
|
||||
/// cards (vendor 0x1002)
|
||||
inline uint32_t small_layers(const char* asked, uint32_t vendor) {
|
||||
if (asked && *asked) return std::clamp<uint32_t>((uint32_t)std::max(atoi(asked), 0), 1u, kMost);
|
||||
return vendor == 0x1002 ? 1u : kMost;
|
||||
}
|
||||
/// Whether a level decoded at w x h goes through the small scratch: there is one, the level fits it, and the texture is not
|
||||
/// one encoded at another size (which has the second scratch)
|
||||
inline bool small_level(uint32_t w, uint32_t h, bool have_small, bool resized) {
|
||||
return have_small && !resized && w <= kSmallDim && h <= kSmallDim;
|
||||
}
|
||||
/// The small scratch's part of the output buffer, and its image, in bytes
|
||||
inline uint64_t small_out_bytes(uint32_t layers) { return (uint64_t)(kSmallDim / 4) * (kSmallDim / 4) * layers * 16; }
|
||||
inline uint64_t small_image_bytes(uint32_t layers) { return (uint64_t)kSmallDim * kSmallDim * 4 * layers; }
|
||||
/// Where layer k of a run begins in the encoder's output: host_shaders/bc_encoder.comp writes block
|
||||
/// (z * blocks_y + y) * blocks_x + x, 16 bytes a block, 8 where only the colour half is written (an image kept as BC1)
|
||||
inline uint64_t layer_offset(uint32_t k, uint32_t w, uint32_t h, bool bc1) {
|
||||
return (uint64_t)k * ((w + 3) / 4) * ((h + 3) / 4) * (bc1 ? 8 : 16);
|
||||
}
|
||||
|
||||
} // namespace bc_scratch
|
||||
87
gpu/vk/bugshot_pixels.h
Normal file
87
gpu/vk/bugshot_pixels.h
Normal file
@@ -0,0 +1,87 @@
|
||||
// (Triforce) The pixels of a bug capture's picture (vk_bugshot.cpp): the window's image as the GPU copied it out, into the
|
||||
// R, G, B, 255 bytes a PNG is written from. Apart from Vulkan, so tests/test_bug_capture.py compiles it alone. The formats
|
||||
// are the ones create_swapchain picks (vk_main.cpp): 8-bit RGBA or BGRA, UNORM or sRGB (the bytes are sRGB codes either
|
||||
// way), and with the Insert menu's HDR a 10-bit HDR10 image, ST 2084 codes of Rec. 2020 light. That one is brought to
|
||||
// sRGB with the picture's paper white as white, so a PNG viewer shows what an SDR screen would show of the game:
|
||||
// highlights above white and colours outside sRGB are cut off.
|
||||
#pragma once
|
||||
|
||||
#include <cmath>
|
||||
#include <cstddef>
|
||||
#include <cstdint>
|
||||
#include <vector>
|
||||
|
||||
namespace bugshot {
|
||||
|
||||
enum Kind { NONE, RGBA8, BGRA8, A2B10G10R10, A2R10G10B10 };
|
||||
|
||||
/// What a window image's format is to this conversion, by its VkFormat number (vk_bugshot.cpp checks the numbers against
|
||||
/// Vulkan's names); NONE: not one it reads
|
||||
inline Kind kind_of(int vk_format) {
|
||||
switch (vk_format) {
|
||||
case 37: case 43: return RGBA8; // (R8G8B8A8_UNORM, R8G8B8A8_SRGB)
|
||||
case 44: case 50: return BGRA8; // (B8G8R8A8_UNORM, B8G8R8A8_SRGB)
|
||||
case 64: return A2B10G10R10; // (A2B10G10R10_UNORM_PACK32: red in the low ten bits)
|
||||
case 58: return A2R10G10B10; // (A2R10G10B10_UNORM_PACK32: blue in the low ten bits)
|
||||
default: return NONE;
|
||||
}
|
||||
}
|
||||
|
||||
/// ST 2084's curve the display's way: a code 0 to 1 as nits, 0 to 10000
|
||||
inline double pq_to_nits(double code) {
|
||||
const double m1 = 2610.0 / 16384.0, m2 = 2523.0 / 4096.0 * 128.0;
|
||||
const double c1 = 3424.0 / 4096.0, c2 = 2413.0 / 4096.0 * 32.0, c3 = 2392.0 / 4096.0 * 32.0;
|
||||
const double p = std::pow(code < 0 ? 0 : code > 1 ? 1 : code, 1.0 / m2);
|
||||
return 10000.0 * std::pow((p > c1 ? p - c1 : 0) / (c2 - c3 * p), 1.0 / m1);
|
||||
}
|
||||
|
||||
/// Linear light 0 to 1 as an sRGB code 0 to 255
|
||||
inline uint8_t srgb8(double linear) {
|
||||
const double x = linear < 0 ? 0 : linear > 1 ? 1 : linear;
|
||||
const double e = x <= 0.0031308 ? 12.92 * x : 1.055 * std::pow(x, 1.0 / 2.4) - 0.055;
|
||||
return (uint8_t)(e * 255.0 + 0.5);
|
||||
}
|
||||
|
||||
/// n texels at px, four bytes each as the GPU wrote them, to R, G, B, 255 in place. hdr10: the image holds ST 2084 codes of
|
||||
/// Rec. 2020 light (a 10-bit format only), and white_nits is the picture's paper white. False: not a format this reads
|
||||
inline bool to_rgba8(uint8_t* px, size_t n, Kind kind, bool hdr10, double white_nits) {
|
||||
if (kind == NONE) return false;
|
||||
if (kind == RGBA8 || kind == BGRA8) {
|
||||
for (size_t i = 0; i < n; i++) {
|
||||
uint8_t* q = px + i * 4;
|
||||
if (kind == BGRA8) {
|
||||
const uint8_t b = q[0];
|
||||
q[0] = q[2], q[2] = b;
|
||||
}
|
||||
q[3] = 255;
|
||||
}
|
||||
return true;
|
||||
}
|
||||
// (ten bits a channel. HDR: a table from a code to its light, white at 1, and one from light to an sRGB code, so a 4K
|
||||
// picture is a few table reads a texel and not nine pow calls)
|
||||
constexpr int kSteps = 4096;
|
||||
std::vector<float> light(hdr10 ? 1024 : 0);
|
||||
std::vector<uint8_t> code8(hdr10 ? kSteps + 1 : 0);
|
||||
const double white = white_nits > 1 ? white_nits : 1;
|
||||
for (size_t v = 0; v < light.size(); v++) light[v] = (float)(pq_to_nits((double)v / 1023.0) / white);
|
||||
for (size_t v = 0; v < code8.size(); v++) code8[v] = srgb8((double)v / kSteps);
|
||||
const auto out8 = [&](float x) { return code8[x <= 0 ? 0 : x >= 1 ? kSteps : (int)(x * kSteps + 0.5f)]; };
|
||||
for (size_t i = 0; i < n; i++) {
|
||||
uint8_t* q = px + i * 4;
|
||||
const uint32_t v = (uint32_t)q[0] | (uint32_t)q[1] << 8 | (uint32_t)q[2] << 16 | (uint32_t)q[3] << 24;
|
||||
const uint32_t lo = v & 1023, mid = (v >> 10) & 1023, hi = (v >> 20) & 1023;
|
||||
const uint32_t r = kind == A2B10G10R10 ? lo : hi, g = mid, b = kind == A2B10G10R10 ? hi : lo;
|
||||
if (!hdr10) { // (sRGB codes with ten bits: the nearest of eight)
|
||||
q[0] = (uint8_t)((r * 255 + 511) / 1023), q[1] = (uint8_t)((g * 255 + 511) / 1023), q[2] = (uint8_t)((b * 255 + 511) / 1023);
|
||||
} else { // (Rec. 2020 light to Rec. 709's primaries, ITU-R BT.2087's matrix, then sRGB's curve)
|
||||
const float lr = light[r], lg = light[g], lb = light[b];
|
||||
q[0] = out8(1.6605f * lr - 0.5876f * lg - 0.0728f * lb);
|
||||
q[1] = out8(-0.1246f * lr + 1.1329f * lg - 0.0083f * lb);
|
||||
q[2] = out8(-0.0182f * lr - 0.1006f * lg + 1.1187f * lb);
|
||||
}
|
||||
q[3] = 255;
|
||||
}
|
||||
return true;
|
||||
}
|
||||
|
||||
} // namespace bugshot
|
||||
164
gpu/vk/built_check.h
Normal file
164
gpu/vk/built_check.h
Normal file
@@ -0,0 +1,164 @@
|
||||
// Whether the caches still have what the light precompile's record says it built (no game, no GPU: the rules; vk_draw.cpp
|
||||
// built_check_cold, precompile_run, built_play_build). After lib_check.h, which asks the same of the stage libraries, and
|
||||
// predict_check.h, which asks it of the predicted pipelines.
|
||||
//
|
||||
// (Triforce, 11-22, the performance audit's PIPE-21, after Source A's and Source B's.) A light precompile (little RAM:
|
||||
// precompile_keeps_all) builds the noted pipelines for the caches, lets them go, and records their keys in <notes>.built,
|
||||
// the compute programs' as one key. A later start trusts that record on its stamp alone: what it lists is left to its first
|
||||
// draw or dispatch, to be built from the caches there. When the caches lost it (the pipeline cache file deleted or refused,
|
||||
// the play file lost, a driver's own cache cleared or past its cap), each of those first draws compiles in play while the
|
||||
// record still says built: a draw left out or waiting for a whole build, a dispatch waiting 50 ms and then left out. Source A
|
||||
// measured a cleared NVIDIA cache: 702,312 draws left out and 21.3 s of draw waits in 140 s of play, against 456 and none
|
||||
// once everything was built again before the title.
|
||||
//
|
||||
// The check. Before the record is trusted, up to kGraphics of the recorded notes (spread over them from another place each
|
||||
// start) are asked of Triforce's own caches as a draw asks, and up to kComputes of the listed compute programs as a dispatch
|
||||
// asks, where the device can ask without compiling (FAIL_ON_PIPELINE_COMPILE_REQUIRED): the answer is the caches' own.
|
||||
// Not a scratch cache of the check's own: AMD's Windows driver keeps no cache of its own (pcache_bound.h), so a question
|
||||
// put to the driver alone would always be cold there. Without the flag, a pipeline is built twice and timed, as lib_check
|
||||
// does: the second comes from the cache by construction, and a first over kRatio times it and kSlackMs more was compiled.
|
||||
// Under kAnswers answers the sample says nothing (the record is trusted, as before); at half of them or more cold it is cold.
|
||||
//
|
||||
// Cold: the record is not trusted this start, everything noted and every listed compute program is built again before the
|
||||
// title (as for a fresh record), and the record is written anew after it. <notes>.built.check keeps, for the stamp, that the
|
||||
// check built them again (kStrikes 0): cold again right after that, the caches do not keep them on this PC (often a driver's
|
||||
// shader cache that is too small), and they are left to their first draws, as before this check, with a hint in the log. A
|
||||
// warm sample forgets that.
|
||||
//
|
||||
// In play, a counter of what the check is for: recorded pipelines whose first build took kSlowUs or more, and recorded
|
||||
// compute programs the dispatch's ask did not find, logged at most every kSayMs while it grows.
|
||||
//
|
||||
// RX_VK_BUILT_CHECK: 1 by default; 0 no check and no counter, as before; 2 always cold (the lab's and the tests').
|
||||
#pragma once
|
||||
|
||||
#include <algorithm>
|
||||
#include <cstdint>
|
||||
#include <cstdio>
|
||||
#include <cstdlib>
|
||||
#include <cstring>
|
||||
#include <string>
|
||||
#include <vector>
|
||||
|
||||
namespace built_check {
|
||||
|
||||
inline constexpr int kGraphics = 24; // (recorded notes asked: Source A asks 48, Source B six computes)
|
||||
inline constexpr int kLook = 2 * kGraphics; // (notes looked at for them: one may have lost its program)
|
||||
inline constexpr int kComputes = 6; // (listed compute programs asked)
|
||||
inline constexpr int kAnswers = 16; // (fewer answers say nothing)
|
||||
inline constexpr double kRatio = 4.0; // (without the flag: a first build against the second, compiled above)
|
||||
inline constexpr double kSlackMs = 3.0; // (and this much more)
|
||||
inline constexpr int kStrikes = 0; // (built again once, and cold again right after: left to the draws)
|
||||
inline constexpr int64_t kSlowUs = 40000; // (a play build of a recorded pipeline over this was compiled)
|
||||
inline constexpr int64_t kSayMs = 10000; // (the counter's line: at most this often)
|
||||
|
||||
/// RX_VK_BUILT_CHECK: 0 off, 1 on (unset too), 2 always cold
|
||||
inline int mode(const char* env) {
|
||||
if (!env || !*env) return 1;
|
||||
const int v = std::atoi(env);
|
||||
return v <= 0 ? 0 : v >= 2 ? 2 : 1;
|
||||
}
|
||||
|
||||
/// One pipeline asked. asked: asked of the caches only; answered_cold: they said a compile is required; made: it was made
|
||||
/// (both times, not asked). One the driver refuses says nothing of the caches
|
||||
struct Probe {
|
||||
bool asked = false, answered_cold = false, made = false;
|
||||
double first_ms = 0, again_ms = 0;
|
||||
};
|
||||
inline bool counts(const Probe& p) { return p.answered_cold || p.made; }
|
||||
inline bool cold(const Probe& p) {
|
||||
return p.answered_cold || (!p.asked && p.made && p.first_ms > kRatio * p.again_ms + kSlackMs);
|
||||
}
|
||||
|
||||
struct Tally {
|
||||
int got = 0, cold = 0, answered = 0;
|
||||
};
|
||||
inline Tally tally(const std::vector<Probe>& probes) {
|
||||
Tally t;
|
||||
for (const Probe& p : probes)
|
||||
if (counts(p)) t.got++, t.cold += cold(p), t.answered += p.answered_cold;
|
||||
return t;
|
||||
}
|
||||
|
||||
enum class Verdict { Unknown, Warm, Cold };
|
||||
/// Under kAnswers answers nothing; half of them or more cold: cold
|
||||
inline Verdict verdict(const Tally& t) {
|
||||
if (t.got < kAnswers) return Verdict::Unknown;
|
||||
return t.cold * 2 >= t.got ? Verdict::Cold : Verdict::Warm;
|
||||
}
|
||||
|
||||
/// k places spread over n, the first at seed's place in the first step: each start asks other pipelines
|
||||
inline std::vector<size_t> spots(size_t n, size_t k, uint64_t seed) {
|
||||
std::vector<size_t> at;
|
||||
if (!n || !k) return at;
|
||||
const size_t step = std::max<size_t>(1, n / k);
|
||||
for (size_t i = (size_t)(seed % step); i < n && at.size() < k; i += step) at.push_back(i);
|
||||
return at;
|
||||
}
|
||||
|
||||
// ---- <notes>.built.check: "built-check <stamp> <strikes>"
|
||||
|
||||
struct State {
|
||||
bool found = false;
|
||||
uint64_t stamp = 0;
|
||||
int strikes = 0;
|
||||
};
|
||||
inline State state_read(const std::string& path) {
|
||||
State s;
|
||||
FILE* f = fopen(path.c_str(), "rb");
|
||||
if (!f) return s;
|
||||
char line[96]{};
|
||||
const size_t n = fread(line, 1, sizeof line - 1, f);
|
||||
fclose(f);
|
||||
line[n] = 0;
|
||||
unsigned long long stamp = 0;
|
||||
int strikes = 0;
|
||||
if (sscanf(line, "built-check %16llx %d", &stamp, &strikes) == 2 && strikes >= 0)
|
||||
s.found = true, s.stamp = stamp, s.strikes = strikes;
|
||||
return s;
|
||||
}
|
||||
/// Written whole beside the old one, then put in place (a start that ends in the middle leaves the old file)
|
||||
inline bool state_write(const std::string& path, uint64_t stamp, int strikes) {
|
||||
char line[96];
|
||||
const int n = snprintf(line, sizeof line, "built-check %016llx %d\n", (unsigned long long)stamp, strikes);
|
||||
const std::string tmp = path + ".tmp";
|
||||
FILE* f = fopen(tmp.c_str(), "wb");
|
||||
if (!f) return false;
|
||||
bool ok = fwrite(line, 1, (size_t)n, f) == (size_t)n;
|
||||
ok = fclose(f) == 0 && ok;
|
||||
if (ok) {
|
||||
remove(path.c_str()); // (rename does not replace on Windows)
|
||||
ok = rename(tmp.c_str(), path.c_str()) == 0;
|
||||
}
|
||||
if (!ok) remove(tmp.c_str());
|
||||
return ok;
|
||||
}
|
||||
|
||||
enum class Do { Nothing, Again, Stop };
|
||||
/// What a start does with the verdict, the record of the last starts and the stamp now. always: mode 2, built again at every
|
||||
/// start whatever was recorded. *next: the strikes to record once they are built again (Again only)
|
||||
inline Do decide(Verdict v, const State& had, uint64_t stamp, bool always, int* next) {
|
||||
if (always) return *next = 0, Do::Again;
|
||||
if (v != Verdict::Cold) return Do::Nothing;
|
||||
const bool same = had.found && had.stamp == stamp; // (built again for this stamp, and cold again since)
|
||||
if (same && had.strikes >= kStrikes) return Do::Stop;
|
||||
*next = same ? had.strikes + 1 : 0;
|
||||
return Do::Again;
|
||||
}
|
||||
|
||||
// ---- the counter: play builds of recorded pipelines the caches did not have
|
||||
|
||||
/// Whether a build of that many microseconds was compiled rather than found
|
||||
inline bool slow(int64_t us) { return us >= kSlowUs; }
|
||||
/// Whether the counter's line is due: said_ms, when it was last written (0: never)
|
||||
inline bool say(int64_t now_ms, int64_t said_ms) { return !said_ms || now_ms - said_ms >= kSayMs; }
|
||||
/// keys sorted without repeats, for has()
|
||||
inline std::vector<uint64_t> sorted(std::vector<uint64_t> keys) {
|
||||
std::sort(keys.begin(), keys.end());
|
||||
keys.erase(std::unique(keys.begin(), keys.end()), keys.end());
|
||||
return keys;
|
||||
}
|
||||
inline bool has(const std::vector<uint64_t>& sorted_keys, uint64_t key) {
|
||||
return std::binary_search(sorted_keys.begin(), sorted_keys.end(), key);
|
||||
}
|
||||
|
||||
} // namespace built_check
|
||||
275
gpu/vk/cache_set.h
Normal file
275
gpu/vk/cache_set.h
Normal file
@@ -0,0 +1,275 @@
|
||||
// (Triforce) One earlier set of the shader caches, for a change of a setting: the rule (vk_cache_set.cpp does the files).
|
||||
//
|
||||
// Why. Everything the game keeps of its shaders is kept for one stamp (vk_shader.cpp disk_stamp, program_cache_stamp): the
|
||||
// program cache (triforce_programs.bin), the pipeline cache (mkw_vk_pipelines.bin by its .for, and the play file), the
|
||||
// pipeline notes and the prepare's record. A start of another stamp drops all of them whole, and prepares every shader again:
|
||||
// on AMD's driver 1.5 to 2 hours (91,600 stage libraries at 16 a second, a cold precompile, the translation). Most stamps
|
||||
// change with the translator, and then nothing of the old set is any use again. Some change with a setting: the menu's Robust
|
||||
// buffers row (the translation's out-of-bounds buffer reads, vk_coopmat.cpp fetch_bounds_buffers), the mesh shaders row, a GPU
|
||||
// of a laptop's two, an RX_VK_ switch of the translation. Switching such a setting back made the same two hours again.
|
||||
// So: when the stamp changed and the build part did not (the translator and its C++ library, vk_shader.cpp
|
||||
// program_cache_build_part: what only a new build changes), the set there is kept, renamed to <name>.<stamp>.<ext>, when the
|
||||
// disk has room for it and for the new one (twice the set and 10 GB free). One set is kept, no more. A start whose stamp is the
|
||||
// kept set's puts it back (every piece, or none: a set with a piece missing or of another size is refused and deleted), and
|
||||
// the set it found takes its place as the kept one. A start of another build deletes the kept set: it is of no use to it.
|
||||
// What a start found, and what changed (triforce_cache_set.txt beside the caches, the parts of the stamp), is in the log.
|
||||
// The driver matches what it is handed by its own hash of the whole creation: a set put back is the very data the driver made
|
||||
// for that stamp, and the caches' own checks (the .for files, the program cache's head, the pipeline cache's device and UUID
|
||||
// header) still read it as they read any start's.
|
||||
// A set put back has the pipeline notes of its own time: what the player met while the other set was there is in the
|
||||
// player's state list (triforce_pipeline_states.bin, not a piece: it is good for any stamp). When that list is not the one
|
||||
// the set was kept with, kStatesPending asks the prepare before the title to go through the lists (vk_draw.cpp prepare_run,
|
||||
// state_list_prepare), as a prepare of every shader would have: those pipelines are built then, not at their first draw.
|
||||
// Files of a kept set's names that no record names (a start that ended in a swap, a record that cannot be read) are deleted
|
||||
// at the next start: no set is left on the disk that nothing knows of.
|
||||
// RX_VK_CACHE_SET=0: no set kept or put back (a change of stamp drops the caches, as before), and a kept set is deleted;
|
||||
// RX_VK_PCACHE_KEEP_OLD and RX_VK_NO_PROGRAM_CACHE leave it off too (a kept set of this build stays, another build's is
|
||||
// deleted). tests/test_cache_set.py (cache_set_files.h does the files)
|
||||
#pragma once
|
||||
#include <cstdint>
|
||||
#include <cstdio>
|
||||
#include <cstdlib>
|
||||
#include <cstring>
|
||||
#include <string>
|
||||
#include <utility>
|
||||
#include <vector>
|
||||
|
||||
namespace cache_set {
|
||||
|
||||
// (every file kept for one stamp, as the game folder names it: vk_shader.cpp disk_init's two, vk_draw.cpp pcache_init's (the
|
||||
// carry's .base too, pcache_carry.h) and pcache_play_open's, the prepare's record and what prepare_run keeps beside it, the notes
|
||||
// and what vk_draw.cpp keeps beside them by the stamp. The first is the set's own: a set is the one its program cache's head names)
|
||||
inline constexpr const char* kPieces[] = {
|
||||
"triforce_programs.bin", "triforce_programs.idx",
|
||||
"mkw_vk_pipelines.bin", "mkw_vk_pipelines.bin.for",
|
||||
"mkw_vk_pipelines.bin.base", // (the carry's history of that file, by its stamp: pcache_carry.h)
|
||||
"mkw_vk_pipelines.bin.own", // ((Triforce) that file holds the driver's part's own alone: vk_draw.cpp pcache_own_mark)
|
||||
"mkw_vk_pipelines_play.bin", "mkw_vk_pipelines_play.bin.for",
|
||||
"triforce_prepared.bin", "triforce_prepared.bin.check",
|
||||
"triforce_prepared.bin.warming", "triforce_prepared.bin.warming.skip",
|
||||
"triforce_pipelines.bin", "triforce_pipelines.bin.predicted",
|
||||
"triforce_pipelines.bin.plan", "triforce_pipelines.bin.built",
|
||||
"triforce_states_applied.bin",
|
||||
};
|
||||
inline constexpr size_t kPieceCount = sizeof kPieces / sizeof kPieces[0];
|
||||
inline constexpr const char* kRecord = "triforce_cache_set.txt"; // (what the set there was made for)
|
||||
inline constexpr const char* kKept = "triforce_cache_set.kept.txt"; // (the kept set: what it was made for, and its pieces)
|
||||
inline constexpr const char* kStates = "triforce_pipeline_states.bin"; // (the player's state list, vk_draw.cpp: not a piece)
|
||||
inline constexpr const char* kStatesPending = "triforce_cache_set.states.txt"; // (a set put back: the lists to go through)
|
||||
inline constexpr uint32_t kProgramsMagic = 0x31505348; // (vk_shader.cpp kDiskMagic: the program cache's head)
|
||||
inline constexpr uint64_t kMargin = 10ull << 30; // (free beyond twice the set)
|
||||
|
||||
/// RX_VK_CACHE_SET: on unless it is 0
|
||||
inline bool on(const char* env) { return !(env && env[0] == '0' && !env[1]); }
|
||||
|
||||
/// A piece's name in the kept set: its name to the first '.', the stamp's 16 digits, the rest
|
||||
/// (mkw_vk_pipelines.bin.for: mkw_vk_pipelines.0123456789abcdef.bin.for)
|
||||
inline std::string kept_name(const std::string& live, uint64_t stamp) {
|
||||
char hex[20];
|
||||
std::snprintf(hex, sizeof hex, "%016llx", (unsigned long long)stamp);
|
||||
const size_t dot = live.find('.');
|
||||
if (dot == std::string::npos) return live + "." + hex;
|
||||
return live.substr(0, dot) + "." + hex + live.substr(dot);
|
||||
}
|
||||
|
||||
/// Whether a file's name is a piece's name in a kept set (kept_name's), and the stamp it names
|
||||
inline bool kept_piece_of(const std::string& file, uint64_t* stamp) {
|
||||
for (const char* p : kPieces) {
|
||||
const std::string live = p;
|
||||
const size_t dot = live.find('.');
|
||||
const std::string stem = live.substr(0, dot), rest = live.substr(dot);
|
||||
if (file.size() != stem.size() + 17 + rest.size() || file.compare(0, stem.size(), stem) || file[stem.size()] != '.' ||
|
||||
file.compare(file.size() - rest.size(), rest.size(), rest))
|
||||
continue;
|
||||
uint64_t v = 0;
|
||||
bool hex = true;
|
||||
for (size_t i = stem.size() + 1; i < stem.size() + 17 && hex; i++) {
|
||||
const char c = file[i];
|
||||
hex = (c >= '0' && c <= '9') || (c >= 'a' && c <= 'f');
|
||||
v = v << 4 | (uint64_t)(c <= '9' ? c - '0' : c - 'a' + 10);
|
||||
}
|
||||
if (!hex) continue;
|
||||
*stamp = v;
|
||||
return true;
|
||||
}
|
||||
return false;
|
||||
}
|
||||
|
||||
/// The stamp a program cache's first 16 bytes name (magic, the build that wrote it, the stamp); false: no program cache's head
|
||||
inline bool programs_head(const uint8_t* b, size_t n, uint64_t* stamp) {
|
||||
if (n < 16) return false;
|
||||
uint32_t magic;
|
||||
std::memcpy(&magic, b, 4);
|
||||
if (magic != kProgramsMagic) return false;
|
||||
std::memcpy(stamp, b + 8, 8);
|
||||
return true;
|
||||
}
|
||||
|
||||
/// What a set was made for (the stamp, the build part, and the parts of the stamp by name for the log) and, for the kept
|
||||
/// one, its pieces as they were renamed: each one's name in the game folder and its size
|
||||
struct Record {
|
||||
uint64_t stamp = 0, build = 0;
|
||||
std::vector<std::pair<std::string, uint64_t>> parts;
|
||||
std::string env; // (the stamp's RX_VK_ and EDEN_ switches that were set, by name)
|
||||
uint64_t states = 0; // (the kept one: the player's state list as it was when kept, cache_set_files.h states_id; 0: none)
|
||||
std::vector<std::pair<std::string, uint64_t>> pieces;
|
||||
uint64_t bytes() const {
|
||||
uint64_t n = 0;
|
||||
for (const auto& p : pieces) n += p.second;
|
||||
return n;
|
||||
}
|
||||
};
|
||||
|
||||
inline std::string format(const Record& r) {
|
||||
std::string s = "triforce-cache-set 1\n";
|
||||
char line[160];
|
||||
std::snprintf(line, sizeof line, "stamp %016llx\nbuild %016llx\n", (unsigned long long)r.stamp, (unsigned long long)r.build);
|
||||
s += line;
|
||||
for (const auto& [name, v] : r.parts) {
|
||||
std::snprintf(line, sizeof line, "part %s %016llx\n", name.c_str(), (unsigned long long)v);
|
||||
s += line;
|
||||
}
|
||||
s += "env " + (r.env.empty() ? std::string("-") : r.env) + "\n";
|
||||
if (r.states) {
|
||||
std::snprintf(line, sizeof line, "states %016llx\n", (unsigned long long)r.states);
|
||||
s += line;
|
||||
}
|
||||
for (const auto& [name, size] : r.pieces) {
|
||||
std::snprintf(line, sizeof line, "piece %s %llu\n", name.c_str(), (unsigned long long)size);
|
||||
s += line;
|
||||
}
|
||||
return s;
|
||||
}
|
||||
|
||||
/// format's text read back; false when it is not one (no head, no stamp or build, a piece that is not one of kPieces)
|
||||
inline bool parse(const std::string& text, Record* out) {
|
||||
Record r;
|
||||
bool head = false, stamp = false, build = false;
|
||||
size_t at = 0;
|
||||
while (at < text.size()) {
|
||||
size_t end = text.find('\n', at);
|
||||
if (end == std::string::npos) end = text.size();
|
||||
std::string line = text.substr(at, end - at);
|
||||
at = end + 1;
|
||||
if (!line.empty() && line.back() == '\r') line.pop_back();
|
||||
if (line.empty()) continue;
|
||||
const size_t sp = line.find(' ');
|
||||
const std::string key = line.substr(0, sp), rest = sp == std::string::npos ? "" : line.substr(sp + 1);
|
||||
if (!head) {
|
||||
if (line != "triforce-cache-set 1") return false;
|
||||
head = true;
|
||||
continue;
|
||||
}
|
||||
unsigned long long v = 0;
|
||||
char name[96];
|
||||
if (key == "stamp" && std::sscanf(rest.c_str(), "%16llx", &v) == 1) {
|
||||
r.stamp = v, stamp = true;
|
||||
} else if (key == "build" && std::sscanf(rest.c_str(), "%16llx", &v) == 1) {
|
||||
r.build = v, build = true;
|
||||
} else if (key == "part" && std::sscanf(rest.c_str(), "%95s %16llx", name, &v) == 2) {
|
||||
r.parts.push_back({name, v});
|
||||
} else if (key == "states" && std::sscanf(rest.c_str(), "%16llx", &v) == 1) {
|
||||
r.states = v;
|
||||
} else if (key == "env") {
|
||||
r.env = rest == "-" ? "" : rest;
|
||||
} else if (key == "piece" && std::sscanf(rest.c_str(), "%95s %llu", name, &v) == 2) {
|
||||
bool known = false;
|
||||
for (const char* p : kPieces) known = known || !std::strcmp(p, name);
|
||||
if (!known) return false;
|
||||
r.pieces.push_back({name, v});
|
||||
}
|
||||
}
|
||||
if (!head || !stamp || !build) return false;
|
||||
*out = std::move(r);
|
||||
return true;
|
||||
}
|
||||
|
||||
/// The parts by name whose value is not the same in a and b (a part only one of them has counts as changed), for the log
|
||||
inline std::vector<std::string> changed(const Record& a, const Record& b) {
|
||||
std::vector<std::string> out;
|
||||
if (a.build != b.build) out.push_back("build");
|
||||
const auto find = [](const Record& r, const std::string& name) -> const uint64_t* {
|
||||
for (const auto& p : r.parts)
|
||||
if (p.first == name) return &p.second;
|
||||
return nullptr;
|
||||
};
|
||||
for (const auto& p : a.parts) {
|
||||
const uint64_t* v = find(b, p.first);
|
||||
if (!v || *v != p.second) out.push_back(p.first);
|
||||
}
|
||||
for (const auto& p : b.parts)
|
||||
if (!find(a, p.first)) out.push_back(p.first);
|
||||
return out;
|
||||
}
|
||||
|
||||
/// Room for a set kept beside the new one: twice the set and kMargin free, counting what deleting the old kept set frees
|
||||
inline bool room(uint64_t free_bytes, uint64_t freed, uint64_t set_bytes) {
|
||||
const uint64_t have = free_bytes + freed < free_bytes ? ~0ull : free_bytes + freed;
|
||||
if (set_bytes > (~0ull - kMargin) / 2) return false;
|
||||
return have >= 2 * set_bytes + kMargin;
|
||||
}
|
||||
|
||||
/// What a start knows: its stamp and build part; the set there (its program cache's head, and the record of what it was made
|
||||
/// for when the record names that stamp); the kept set (its record, and whether every piece is there whole); the disk
|
||||
struct Facts {
|
||||
bool on = false; // (keeping and putting back allowed)
|
||||
bool off_forgets = false; // (off by RX_VK_CACHE_SET=0: a kept set of this build is deleted too, for its space)
|
||||
uint64_t stamp = 0, build = 0; // (this start's)
|
||||
bool live = false; // (a program cache there, its head read)
|
||||
uint64_t live_stamp = 0; // (what its head names)
|
||||
bool live_named = false; // (the record names live_stamp: live_build is known)
|
||||
uint64_t live_build = 0, live_bytes = 0;
|
||||
bool kept = false; // (a kept set's record)
|
||||
uint64_t kept_stamp = 0, kept_build = 0, kept_bytes = 0;
|
||||
bool kept_whole = false; // (each of its pieces there at its size, its program cache's head naming kept_stamp)
|
||||
bool space_known = false;
|
||||
uint64_t free_bytes = 0;
|
||||
};
|
||||
enum class Do {
|
||||
Same, // (the set there is this stamp's: nothing to do)
|
||||
Restore, // (the kept set is this stamp's: put back; keep_live: the set there becomes the kept one, else it is deleted)
|
||||
Keep, // (the set there is kept, renamed; the caches start empty for this stamp)
|
||||
Drop, // (nothing renamed: the caches drop the set there, as before)
|
||||
};
|
||||
struct Plan {
|
||||
Do what = Do::Drop;
|
||||
bool forget = false; // (the kept set is deleted first)
|
||||
bool keep_live = false; // (Restore: the set there is kept in its place)
|
||||
};
|
||||
|
||||
inline Plan decide(const Facts& f) {
|
||||
Plan p;
|
||||
const bool same = f.live && f.live_stamp == f.stamp;
|
||||
if (!f.on) {
|
||||
// (off: nothing kept or put back. A kept set of another build is of no use to any build again: deleted; one of this
|
||||
// build too when the off is RX_VK_CACHE_SET=0's, the switch that frees its space)
|
||||
p.what = same ? Do::Same : Do::Drop;
|
||||
p.forget = f.kept && (f.kept_build != f.build || f.off_forgets);
|
||||
return p;
|
||||
}
|
||||
bool kept_ours = f.kept && f.kept_build == f.build;
|
||||
// (another build's kept set is of no use to this one, nor to any later: deleted. A kept set that names the set there
|
||||
// is what a start that ended while it swapped the two left: its pieces are not where it says)
|
||||
p.forget = f.kept && (!kept_ours || (same && f.kept_stamp == f.stamp));
|
||||
if (same) {
|
||||
p.what = Do::Same;
|
||||
return p;
|
||||
}
|
||||
const bool live_ours = f.live && f.live_named && f.live_build == f.build;
|
||||
if (kept_ours && f.kept_stamp == f.stamp) {
|
||||
if (f.kept_whole) {
|
||||
p.what = Do::Restore;
|
||||
p.keep_live = live_ours;
|
||||
return p;
|
||||
}
|
||||
p.forget = true; // (a piece missing or cut: refused)
|
||||
}
|
||||
if (live_ours && f.space_known && room(f.free_bytes, f.kept ? f.kept_bytes : 0, f.live_bytes)) {
|
||||
p.what = Do::Keep;
|
||||
p.forget = f.kept; // (one set kept: the one there takes the old one's place)
|
||||
return p;
|
||||
}
|
||||
return p;
|
||||
}
|
||||
|
||||
} // namespace cache_set
|
||||
306
gpu/vk/cache_set_files.h
Normal file
306
gpu/vk/cache_set_files.h
Normal file
@@ -0,0 +1,306 @@
|
||||
// (Triforce) The kept set of the shader caches in the game folder: what a start finds there, and what it renames, deletes and
|
||||
// writes for cache_set.h's rule. The standard library and file_whole.h alone, nothing of the device: vk_cache_set.cpp gives it
|
||||
// this start's record (the stamp and its parts) and the switches, and tests/test_cache_set.py runs it on a folder of its own.
|
||||
#pragma once
|
||||
#include <cstdint>
|
||||
#include <cstdio>
|
||||
#include <filesystem>
|
||||
#include <string>
|
||||
#include <utility>
|
||||
#include <vector>
|
||||
|
||||
#include "cache_set.h"
|
||||
#include "file_whole.h"
|
||||
|
||||
namespace cache_set {
|
||||
|
||||
using Log = void (*)(const char* fmt, ...);
|
||||
|
||||
inline std::string stamp_hex(uint64_t s) {
|
||||
char hex[20];
|
||||
std::snprintf(hex, sizeof hex, "%016llx", (unsigned long long)s);
|
||||
return hex;
|
||||
}
|
||||
inline bool read_text(const std::string& path, std::string* out) {
|
||||
FILE* f = std::fopen(path.c_str(), "rb");
|
||||
if (!f) return false;
|
||||
char buf[4096];
|
||||
out->clear();
|
||||
for (size_t n; (n = std::fread(buf, 1, sizeof buf, f)) > 0 && out->size() < (1u << 20);) out->append(buf, n);
|
||||
std::fclose(f);
|
||||
return true;
|
||||
}
|
||||
inline bool read_record(const std::string& path, Record* r) {
|
||||
std::string text;
|
||||
return read_text(path, &text) && parse(text, r);
|
||||
}
|
||||
/// The stamp a program cache's head names
|
||||
inline bool head_of(const std::string& path, uint64_t* stamp) {
|
||||
FILE* f = std::fopen(path.c_str(), "rb");
|
||||
if (!f) return false;
|
||||
uint8_t b[16];
|
||||
const size_t n = std::fread(b, 1, sizeof b, f);
|
||||
std::fclose(f);
|
||||
return programs_head(b, n, stamp);
|
||||
}
|
||||
inline bool size_of(const std::string& path, uint64_t* size) {
|
||||
std::error_code ec;
|
||||
if (!std::filesystem::is_regular_file(path, ec) || ec) return false;
|
||||
const auto n = std::filesystem::file_size(path, ec);
|
||||
if (ec) return false;
|
||||
*size = (uint64_t)n;
|
||||
return true;
|
||||
}
|
||||
/// The pieces in the game folder by their names there, and their sizes
|
||||
inline std::vector<std::pair<std::string, uint64_t>> pieces_there(const std::string& dir) {
|
||||
std::vector<std::pair<std::string, uint64_t>> out;
|
||||
for (const char* p : kPieces) {
|
||||
uint64_t n = 0;
|
||||
if (size_of(dir + p, &n)) out.push_back({p, n});
|
||||
}
|
||||
return out;
|
||||
}
|
||||
/// The player's state list (kStates) as it is now: its bytes hashed, with its size; 0 when there is none, or no record in it
|
||||
inline uint64_t states_id(const std::string& dir) {
|
||||
FILE* f = std::fopen((dir + kStates).c_str(), "rb");
|
||||
if (!f) return 0;
|
||||
uint64_t h = 0xCBF29CE484222325ull, total = 0;
|
||||
std::vector<uint8_t> buf(1 << 16);
|
||||
for (size_t n; (n = std::fread(buf.data(), 1, buf.size(), f)) > 0; total += n)
|
||||
for (size_t i = 0; i < n; i++) h = (h ^ buf[i]) * 1099511628211ull;
|
||||
std::fclose(f);
|
||||
for (int i = 0; i < 8; i++) h = (h ^ (uint8_t)(total >> (8 * i))) * 1099511628211ull;
|
||||
return total > 12 ? (h | 1) : 0; // (12: the list's head alone)
|
||||
}
|
||||
/// Every piece of the kept set there, at its size, and its program cache's head naming its stamp
|
||||
inline bool kept_whole(const std::string& dir, const Record& kept) {
|
||||
bool programs = false;
|
||||
for (const auto& [name, size] : kept.pieces) {
|
||||
const std::string at = dir + kept_name(name, kept.stamp);
|
||||
uint64_t n = 0;
|
||||
if (!size_of(at, &n) || n != size) return false;
|
||||
if (name == kPieces[0]) {
|
||||
uint64_t stamp = 0;
|
||||
if (!head_of(at, &stamp) || stamp != kept.stamp) return false;
|
||||
programs = true;
|
||||
}
|
||||
}
|
||||
return programs;
|
||||
}
|
||||
/// The kept set deleted: every piece its stamp names (what its record lists, and what a start cut short left), then its record
|
||||
inline void forget(const std::string& dir, const Record& kept) {
|
||||
for (const char* p : kPieces) {
|
||||
std::error_code ec;
|
||||
std::filesystem::remove(dir + kept_name(p, kept.stamp), ec);
|
||||
}
|
||||
std::remove((dir + kKept).c_str());
|
||||
}
|
||||
/// The files of a kept set's names (kept_piece_of) that are not the kept set's (kept: its record read; kept_stamp: its stamp)
|
||||
/// deleted: what a start that ended in a swap left, or a set whose record cannot be read. Their count and bytes
|
||||
inline std::pair<size_t, uint64_t> sweep(const std::string& dir, bool kept, uint64_t kept_stamp) {
|
||||
std::pair<size_t, uint64_t> gone{0, 0};
|
||||
std::error_code ec;
|
||||
std::vector<std::filesystem::path> drop;
|
||||
for (std::filesystem::directory_iterator it(dir.empty() ? std::string(".") : dir, ec), end; !ec && it != end; it.increment(ec)) {
|
||||
std::error_code e2;
|
||||
if (!it->is_regular_file(e2) || e2) continue;
|
||||
uint64_t stamp = 0;
|
||||
if (!kept_piece_of(it->path().filename().string(), &stamp) || (kept && stamp == kept_stamp)) continue;
|
||||
drop.push_back(it->path());
|
||||
}
|
||||
for (const auto& p : drop) {
|
||||
std::error_code sized, removed;
|
||||
const auto n = std::filesystem::file_size(p, sized);
|
||||
if (std::filesystem::remove(p, removed) && !removed) gone.first++, gone.second += sized ? 0 : (uint64_t)n;
|
||||
}
|
||||
return gone;
|
||||
}
|
||||
/// Each file renamed (a file there in the way replaced), or none: one that cannot be leaves those done renamed back
|
||||
inline bool move_all(const std::vector<std::pair<std::string, std::string>>& moves, Log log) {
|
||||
for (size_t i = 0; i < moves.size(); i++) {
|
||||
if (file_writers::replace_file(moves[i].first, moves[i].second)) continue;
|
||||
log("vk: shader caches: %s cannot be renamed to %s: the set is left as it was", moves[i].first.c_str(),
|
||||
moves[i].second.c_str());
|
||||
for (size_t j = i; j-- > 0;) file_writers::replace_file(moves[j].second, moves[j].first);
|
||||
return false;
|
||||
}
|
||||
return true;
|
||||
}
|
||||
inline std::string joined(const std::vector<std::string>& v) {
|
||||
std::string s;
|
||||
for (const auto& x : v) s += (s.empty() ? "" : ", ") + (x == "robust" ? std::string("Robust buffers") : x);
|
||||
return s;
|
||||
}
|
||||
|
||||
/// The switches a start has (vk_cache_set.cpp reads them from the environment)
|
||||
struct Switches {
|
||||
bool on = true; // (RX_VK_CACHE_SET not 0, RX_VK_PCACHE_KEEP_OLD not set)
|
||||
bool off_forgets = false; // (off by RX_VK_CACHE_SET=0: a kept set is deleted)
|
||||
bool space_given = false; // (tests: the disk's free space as given, not as the disk says)
|
||||
uint64_t free_bytes = 0;
|
||||
};
|
||||
|
||||
/// The start's kept set decided and done in dir (its name ends in a separator, or is empty): now is this start's record (the
|
||||
/// stamp, the build part, the parts). What a start does before anything opens a cache
|
||||
inline Plan run(const std::string& dir, const Record& now, const Switches& sw, Log log) {
|
||||
Record had, kept;
|
||||
const bool had_read = read_record(dir + kRecord, &had);
|
||||
Facts f;
|
||||
f.on = sw.on, f.off_forgets = sw.off_forgets;
|
||||
f.stamp = now.stamp, f.build = now.build;
|
||||
f.kept = read_record(dir + kKept, &kept);
|
||||
// (no set is left there that nothing knows of: before the free space is measured, which it frees)
|
||||
const auto [swept, swept_bytes] = sweep(dir, f.kept, kept.stamp);
|
||||
if (swept)
|
||||
log("vk: shader caches: %zu files of a kept set no record names deleted (%llu MB: a start that ended while it swapped "
|
||||
"the sets, or a record that cannot be read)", swept, (unsigned long long)(swept_bytes >> 20));
|
||||
f.live = head_of(dir + kPieces[0], &f.live_stamp);
|
||||
f.live_named = f.live && had_read && had.stamp == f.live_stamp;
|
||||
f.live_build = had.build;
|
||||
const std::vector<std::pair<std::string, uint64_t>> live = pieces_there(dir);
|
||||
for (const auto& p : live) f.live_bytes += p.second;
|
||||
if (f.kept) {
|
||||
f.kept_stamp = kept.stamp, f.kept_build = kept.build, f.kept_bytes = kept.bytes();
|
||||
f.kept_whole = kept_whole(dir, kept);
|
||||
}
|
||||
if (sw.space_given) {
|
||||
f.space_known = true, f.free_bytes = sw.free_bytes;
|
||||
} else {
|
||||
std::error_code ec;
|
||||
const auto space = std::filesystem::space(dir.empty() ? std::string(".") : dir, ec);
|
||||
f.space_known = !ec;
|
||||
f.free_bytes = ec ? 0 : (uint64_t)space.available;
|
||||
}
|
||||
const Plan plan = decide(f);
|
||||
|
||||
if (f.live && f.live_stamp != f.stamp) { // (what changed, when it is known)
|
||||
std::string what = "not known: the set there has no record of what it was made for (a build before this rule)";
|
||||
if (f.live_named) {
|
||||
const std::vector<std::string> parts = changed(had, now);
|
||||
what = parts.empty() ? "a part of the translation's settings not named here (its per-GPU settings)" : joined(parts);
|
||||
if (had.env != now.env) what += "; the switches set: " + (had.env.empty() ? "none" : had.env) + " -> " +
|
||||
(now.env.empty() ? "none" : now.env);
|
||||
}
|
||||
log("vk: shader caches: made for %s, this start is %s: changed: %s", stamp_hex(f.live_stamp).c_str(),
|
||||
stamp_hex(f.stamp).c_str(), what.c_str());
|
||||
}
|
||||
if (plan.forget && f.kept) {
|
||||
forget(dir, kept);
|
||||
log("vk: shader caches: the kept set for %s (%llu MB) deleted: %s", stamp_hex(kept.stamp).c_str(),
|
||||
(unsigned long long)(kept.bytes() >> 20),
|
||||
kept.build != now.build ? "it is another build's"
|
||||
: !f.on ? "RX_VK_CACHE_SET=0"
|
||||
: !f.kept_whole && kept.stamp == now.stamp ? "a piece of it is missing or not its size"
|
||||
: "one set is kept, and this one takes its place");
|
||||
} else if (f.kept && !f.on) {
|
||||
log("vk: shader caches: the kept set for %s (%llu MB, as <name>.%s.<ext>) stays while it is off (RX_VK_PCACHE_KEEP_OLD; "
|
||||
"RX_VK_CACHE_SET=0 deletes it)", stamp_hex(kept.stamp).c_str(), (unsigned long long)(kept.bytes() >> 20),
|
||||
stamp_hex(kept.stamp).c_str());
|
||||
}
|
||||
const auto renames = [&](const std::vector<std::pair<std::string, uint64_t>>& pieces, uint64_t stamp, bool to_kept) {
|
||||
std::vector<std::pair<std::string, std::string>> m;
|
||||
for (const auto& p : pieces) {
|
||||
const std::string a = dir + p.first, b = dir + kept_name(p.first, stamp);
|
||||
m.push_back(to_kept ? std::make_pair(a, b) : std::make_pair(b, a));
|
||||
}
|
||||
return m;
|
||||
};
|
||||
if (plan.what == Do::Keep) {
|
||||
// (its record first: a start that ends while the pieces are renamed leaves a set the next one refuses and deletes)
|
||||
Record k = had;
|
||||
k.pieces = live;
|
||||
k.states = states_id(dir);
|
||||
if (!file_writers::write_whole(dir + kKept, format(k))) {
|
||||
log("vk: shader caches: %s cannot be written: the set there is not kept", kKept);
|
||||
} else if (!move_all(renames(live, f.live_stamp, true), log)) {
|
||||
std::remove((dir + kKept).c_str());
|
||||
} else {
|
||||
log("vk: shader caches: the set for %s kept (%llu MB, %zu files, as <name>.%s.<ext>): a start for it again puts it "
|
||||
"back, and its shaders are not prepared again (RX_VK_CACHE_SET=0: not kept). The caches start empty for "
|
||||
"this stamp", stamp_hex(f.live_stamp).c_str(), (unsigned long long)(f.live_bytes >> 20), live.size(),
|
||||
stamp_hex(f.live_stamp).c_str());
|
||||
}
|
||||
} else if (plan.what == Do::Restore) {
|
||||
bool ok = true;
|
||||
std::vector<std::pair<std::string, std::string>> aside;
|
||||
if (plan.keep_live) {
|
||||
aside = renames(live, f.live_stamp, true);
|
||||
ok = move_all(aside, log);
|
||||
}
|
||||
if (ok && !move_all(renames(kept.pieces, kept.stamp, false), log)) {
|
||||
ok = false;
|
||||
for (size_t j = aside.size(); j-- > 0;) file_writers::replace_file(aside[j].second, aside[j].first);
|
||||
}
|
||||
if (!ok) {
|
||||
log("vk: shader caches: the kept set for %s is not put back: the caches drop the set there, as before",
|
||||
stamp_hex(kept.stamp).c_str());
|
||||
} else {
|
||||
// (the pieces the kept set did not have are of the set there: theirs are another stamp's, or none)
|
||||
if (!plan.keep_live)
|
||||
for (const auto& p : live) {
|
||||
bool theirs = false;
|
||||
for (const auto& q : kept.pieces) theirs = theirs || q.first == p.first;
|
||||
if (!theirs) std::remove((dir + p.first).c_str());
|
||||
}
|
||||
const uint64_t states = states_id(dir);
|
||||
bool kept_now = false;
|
||||
if (plan.keep_live) {
|
||||
Record k = had;
|
||||
k.pieces = live;
|
||||
k.states = states;
|
||||
kept_now = file_writers::write_whole(dir + kKept, format(k));
|
||||
if (!kept_now) {
|
||||
for (const auto& p : live) std::remove((dir + kept_name(p.first, f.live_stamp)).c_str());
|
||||
std::remove((dir + kKept).c_str());
|
||||
}
|
||||
} else {
|
||||
std::remove((dir + kKept).c_str());
|
||||
}
|
||||
// (the player met pipelines while the set put back was not there: its notes lack them. The prepare before the
|
||||
// title goes through the state lists, and they are built before the game, as a prepare of every shader would)
|
||||
const bool lists = states && states != kept.states;
|
||||
if (lists && !file_writers::write_whole(dir + kStatesPending, "put back " + stamp_hex(kept.stamp) + "\n"))
|
||||
log("vk: shader caches: %s cannot be written: the pipelines met since the set was kept are built in play",
|
||||
kStatesPending);
|
||||
log("vk: shader caches: the kept set for %s put back (%llu MB, %zu files): its shaders are not prepared again%s%s",
|
||||
stamp_hex(kept.stamp).c_str(), (unsigned long long)(kept.bytes() >> 20), kept.pieces.size(),
|
||||
kept_now ? (", and the set for " + stamp_hex(f.live_stamp) + " is kept in its place").c_str()
|
||||
: plan.keep_live ? ", and the set there could not be kept: deleted" : "",
|
||||
lists ? "; the pipeline state list changed since it was kept: its pipelines are noted before the title" : "");
|
||||
}
|
||||
} else if (plan.what == Do::Drop && f.live && f.live_stamp != f.stamp) {
|
||||
const char* why = !f.on ? "RX_VK_CACHE_SET=0 or RX_VK_PCACHE_KEEP_OLD"
|
||||
: !f.live_named ? "what it was made for is not known"
|
||||
: f.live_build != f.build ? "another build of the translator made it: it is of no use again"
|
||||
: "the disk has no room for it beside a new one (twice the set and 10 GB free)";
|
||||
log("vk: shader caches: the set for %s (%llu MB) is not kept: %s; free %llu MB", stamp_hex(f.live_stamp).c_str(),
|
||||
(unsigned long long)(f.live_bytes >> 20), why, (unsigned long long)(f.free_bytes >> 20));
|
||||
}
|
||||
// (what the set there is made for from now on: this start's caches are written for its stamp)
|
||||
std::string text;
|
||||
const std::string want = format(now);
|
||||
if (!read_text(dir + kRecord, &text) || text != want)
|
||||
if (!file_writers::write_whole(dir + kRecord, want)) log("vk: shader caches: %s cannot be written", kRecord);
|
||||
return plan;
|
||||
}
|
||||
|
||||
/// A start that leaves the caches alone (RX_VK_NO_PROGRAM_CACHE: no program cache to say which set is there): a kept set of
|
||||
/// another build is still deleted, and files of a kept set's names that no record names
|
||||
inline void tidy(const std::string& dir, uint64_t build, Log log) {
|
||||
Record kept;
|
||||
const bool known = read_record(dir + kKept, &kept);
|
||||
const auto [swept, swept_bytes] = sweep(dir, known, kept.stamp);
|
||||
if (swept)
|
||||
log("vk: shader caches: %zu files of a kept set no record names deleted (%llu MB)", swept,
|
||||
(unsigned long long)(swept_bytes >> 20));
|
||||
if (known && kept.build != build) {
|
||||
forget(dir, kept);
|
||||
log("vk: shader caches: the kept set for %s (%llu MB) deleted: it is another build's", stamp_hex(kept.stamp).c_str(),
|
||||
(unsigned long long)(kept.bytes() >> 20));
|
||||
} else if (known) {
|
||||
log("vk: shader caches: the kept set for %s (%llu MB) stays: RX_VK_NO_PROGRAM_CACHE", stamp_hex(kept.stamp).c_str(),
|
||||
(unsigned long long)(kept.bytes() >> 20));
|
||||
}
|
||||
}
|
||||
|
||||
} // namespace cache_set
|
||||
64
gpu/vk/cap_spare.h
Normal file
64
gpu/vk/cap_spare.h
Normal file
@@ -0,0 +1,64 @@
|
||||
// (Triforce, 11-22, the performance audit's MEM-33, after Source A's) Which spare capture buffer a capture takes, how
|
||||
// much the spares and the ones being made ahead may hold together, and a capture's size in steps.
|
||||
//
|
||||
// A capture's buffer comes from the spares (retired buffers, and the 32 to 192 MB ones made ahead on a worker), or is
|
||||
// made new on the game's submitting thread. Three things were so before:
|
||||
// - the new buffer was made with cap_free_mu held (AMD's driver: about 0.8 ms a MB), and the replay's take, holding
|
||||
// pools_mu, waits on that mutex to retire a buffer;
|
||||
// - any spare that fit was taken, so a 2 MB capture could sit in a load's 192 MB spare for the 24 submissions a capture
|
||||
// is kept, and the load then made its own on the game's thread;
|
||||
// - a spare made ahead was made once: the one a load took was not made again.
|
||||
// The rules here are the decisions alone (no Vulkan, no lock): tests/test_capture_spare_reuse.py compiles them, and
|
||||
// vk_memory.cpp's cap_buf_get, cap_prewarm_class and cap_buf_free ask them. RX_VK_CAP_SPARE_FIT=0: as before.
|
||||
#pragma once
|
||||
#include <cstddef>
|
||||
#include <cstdint>
|
||||
|
||||
namespace cap_spare {
|
||||
|
||||
constexpr uint64_t kMB = 1ull << 20;
|
||||
constexpr uint64_t kBig = 32 * kMB; // (a load's spare: the classes made ahead start here)
|
||||
constexpr uint64_t kHeld = 768 * kMB; // (the most the spares hold, whatever the card)
|
||||
constexpr size_t kMost = 12; // (and the most there are)
|
||||
|
||||
/// A capture of size class `cls` leaves the spare of `size` bytes alone: a big one (32 MB or more) over twice its class.
|
||||
/// any_size: no spare is left alone (spares capped at 32 MB, the switch off, or the fallback after a failed allocation)
|
||||
constexpr bool skips(uint64_t size, uint64_t cls, bool any_size) { return !any_size && size >= kBig && size > 2 * cls; }
|
||||
|
||||
/// The spare a capture of `need` bytes (class `cls`) takes of n spares, size_of(i) the i-th one's bytes: the smallest that
|
||||
/// fits and is not left alone, the first of equals. n: none
|
||||
template <class SizeOf>
|
||||
constexpr size_t pick(size_t n, const SizeOf& size_of, uint64_t need, uint64_t cls, bool any_size) {
|
||||
size_t best = n;
|
||||
for (size_t i = 0; i < n; i++) {
|
||||
const uint64_t s = size_of(i);
|
||||
if (s >= need && !skips(s, cls, any_size) && (best == n || s < size_of(best))) best = i;
|
||||
}
|
||||
return best;
|
||||
}
|
||||
|
||||
/// Room for one more spare of `size` bytes to be made ahead: with what the spares hold and what is being made (reserved)
|
||||
/// it stays within what spares may hold (spare_max: the card's, cap_spare_max) and 768 MB
|
||||
constexpr bool room(uint64_t held, uint64_t reserved, uint64_t size, uint64_t spare_max) {
|
||||
const uint64_t limit = spare_max < kHeld ? spare_max : kHeld;
|
||||
return held + reserved <= limit && size <= limit - (held + reserved);
|
||||
}
|
||||
|
||||
/// A buffer of `size` bytes given back is kept as a spare: fewer than 12 with the ones being made (workers), none over 256
|
||||
/// MB (keep_old, RX_VK_CAPTURE_KEEP_OLD: 64), and with the bytes being made (reserved) within the limit. With no workers
|
||||
/// and nothing reserved this is the rule as it was
|
||||
constexpr bool keeps(size_t spares, size_t workers, uint64_t held, uint64_t reserved, uint64_t size, uint64_t spare_max, bool keep_old) {
|
||||
const uint64_t limit = keep_old ? spare_max : (spare_max < kHeld ? spare_max : kHeld);
|
||||
return spares + workers < kMost && size <= (keep_old ? 64 * kMB : 256 * kMB) && held + reserved <= limit &&
|
||||
size <= limit - (held + reserved);
|
||||
}
|
||||
|
||||
/// (RX_VK_CAP_STEP_MB=<n>, 0: off) a capture over 16 and up to 32 MB in steps of step_mb MB, not the whole 32 MB class,
|
||||
/// where spares are capped at 32 MB (a card of 6.5 GB or less). cls: its class as it is (cap_buf_size)
|
||||
constexpr uint64_t stepped(uint64_t need, uint64_t cls, uint64_t step_mb, bool capped) {
|
||||
if (!step_mb || !capped || cls != kBig || need <= 16 * kMB || need > kBig) return cls;
|
||||
const uint64_t step = step_mb * kMB, v = (need + step - 1) / step * step;
|
||||
return v < cls ? v : cls;
|
||||
}
|
||||
|
||||
} // namespace cap_spare
|
||||
129
gpu/vk/capture_behind.h
Normal file
129
gpu/vk/capture_behind.h
Normal file
@@ -0,0 +1,129 @@
|
||||
/* (Triforce) A submit-time capture's copies behind the game's ordering point (vk_memory.cpp capture_make, RX_VK_CAPTURE_BEHIND).
|
||||
With the replay on the submitting thread (RX_VK_ASYNC off, the default) the game's thread was held for the whole capture before
|
||||
its replay began: the scan of the write marks, then the copy of every page they named. At the Castle Town Market that was
|
||||
~0.35 ms a submit, six submits a frame: 2.1 ms of a 10.6 ms frame, outside the replay (the frame trace of 2026-11-15: scan 0.10
|
||||
ms over ~1,240 ranges, ~2,000 pages copied, 7.8 MB). The scan is the ordering point: the pages written before the submit, their
|
||||
marks taken. The copy need not be on the game's thread before the replay: the submitting thread replays the submission before it
|
||||
returns to the game, so nothing it writes after its submit can reach the copy, and what reads the copy comes after the scan: the
|
||||
replay's reads as submitted, the late pages' second copy and the texture shadow (take_capture), and the GPU's copies into the
|
||||
mirrors once a recording is submitted. So the copy is cut in parts of consecutive pages that the capture's copy helpers take as
|
||||
soon as the scan is done, while the submitting thread replays. A reader of a page whose part nobody has taken yet copies that part
|
||||
itself; one whose part is being copied waits for it. Before a recording goes to the GPU, and before the submitting thread goes
|
||||
back to the game, every part is done. Each page is copied once, by whoever claims its part.
|
||||
Plain C++, no Vulkan: tests/test_capture_behind.py compiles it with stand-ins. */
|
||||
#pragma once
|
||||
|
||||
#include <atomic>
|
||||
#include <condition_variable>
|
||||
#include <cstdint>
|
||||
#include <memory>
|
||||
#include <mutex>
|
||||
#include <vector>
|
||||
|
||||
namespace mkwvk::capture_behind {
|
||||
|
||||
constexpr int64_t kPartPages = 32; // (128 KB a part: what a reader that needs one of its pages copies itself, ~10 us)
|
||||
|
||||
struct Part {
|
||||
int64_t i0 = 0, i1 = 0; // (the capture's indices [i0, i1): consecutive, none of them copied ahead)
|
||||
std::atomic<uint8_t> state{0}; // (0 nobody's yet, 1 being copied, 2 copied)
|
||||
};
|
||||
|
||||
/// A capture's pages left to copy, in parts. An index is the page's place in the capture (its slot in the capture's buffer).
|
||||
/// copy(i0, i1) is the caller's: it copies indices [i0, i1), and is run once for each part, by the thread that claims it
|
||||
struct Work {
|
||||
std::unique_ptr<Part[]> parts;
|
||||
int nparts = 0;
|
||||
std::vector<int32_t> part_of; // (by index: its part; -1: copied ahead, not one of this work's)
|
||||
std::atomic<int> left{0}; // (the parts not copied yet)
|
||||
std::mutex mu; // (the waits for a part another thread copies)
|
||||
std::condition_variable cv;
|
||||
|
||||
/// The parts: the indices i < n with ahead[i] == 0 (ahead: null for none), in runs of at most part_pages cut at every index
|
||||
/// copied ahead
|
||||
void plan(const uint8_t* ahead, int64_t n, int64_t part_pages = kPartPages) {
|
||||
if (part_pages < 1) part_pages = 1;
|
||||
part_of.assign((size_t)(n > 0 ? n : 0), -1);
|
||||
std::vector<std::pair<int64_t, int64_t>> runs;
|
||||
for (int64_t i = 0; i < n;) {
|
||||
if (ahead && ahead[i]) {
|
||||
i++;
|
||||
continue;
|
||||
}
|
||||
int64_t j = i + 1;
|
||||
while (j < n && !(ahead && ahead[j]) && j - i < part_pages) j++;
|
||||
runs.push_back({i, j});
|
||||
i = j;
|
||||
}
|
||||
nparts = (int)runs.size();
|
||||
parts.reset(new Part[(size_t)nparts]);
|
||||
for (int k = 0; k < nparts; k++) {
|
||||
parts[k].i0 = runs[(size_t)k].first, parts[k].i1 = runs[(size_t)k].second;
|
||||
for (int64_t i = parts[k].i0; i < parts[k].i1; i++) part_of[(size_t)i] = k;
|
||||
}
|
||||
left.store(nparts, std::memory_order_release);
|
||||
}
|
||||
/// Part k is this thread's to copy (nobody had taken it)
|
||||
bool claim(int k) {
|
||||
uint8_t free = 0;
|
||||
return parts[k].state.compare_exchange_strong(free, 1, std::memory_order_acq_rel);
|
||||
}
|
||||
/// Part k copied, by the thread that claimed it: its bytes are there for every thread that sees it so
|
||||
void copied(int k) {
|
||||
{
|
||||
std::lock_guard lk(mu);
|
||||
parts[k].state.store(2, std::memory_order_release);
|
||||
left.fetch_sub(1, std::memory_order_acq_rel);
|
||||
}
|
||||
cv.notify_all();
|
||||
}
|
||||
bool done() const { return left.load(std::memory_order_acquire) == 0; }
|
||||
/// Index i is copied (or was copied ahead)
|
||||
bool ready(int64_t i) const {
|
||||
if (i < 0 || (size_t)i >= part_of.size()) return true;
|
||||
const int32_t k = part_of[(size_t)i];
|
||||
return k < 0 || parts[k].state.load(std::memory_order_acquire) == 2;
|
||||
}
|
||||
/// Part k copied on return: by this thread when nobody had taken it, else by whoever took it, waited for. 1: this thread
|
||||
/// copied it, 2: it waited, 0: it was copied already
|
||||
template <class Copy> int make_ready(int k, Copy&& copy) {
|
||||
if (parts[k].state.load(std::memory_order_acquire) == 2) return 0;
|
||||
if (claim(k)) {
|
||||
copy(parts[k].i0, parts[k].i1);
|
||||
copied(k);
|
||||
return 1;
|
||||
}
|
||||
std::unique_lock lk(mu);
|
||||
cv.wait(lk, [&] { return parts[k].state.load(std::memory_order_acquire) == 2; });
|
||||
return 2;
|
||||
}
|
||||
/// Index i copied on return (make_ready of its part)
|
||||
template <class Copy> int ensure(int64_t i, Copy&& copy) {
|
||||
if (i < 0 || (size_t)i >= part_of.size()) return 0;
|
||||
const int32_t k = part_of[(size_t)i];
|
||||
return k < 0 ? 0 : make_ready(k, copy);
|
||||
}
|
||||
/// Every part copied on return: those nobody had taken by this thread, in order, then the others waited for. The parts this
|
||||
/// thread copied
|
||||
template <class Copy> int finish(Copy&& copy) {
|
||||
int made = 0;
|
||||
for (int k = 0; k < nparts; k++)
|
||||
if (parts[k].state.load(std::memory_order_acquire) == 0 && claim(k)) {
|
||||
copy(parts[k].i0, parts[k].i1);
|
||||
copied(k);
|
||||
made++;
|
||||
}
|
||||
std::unique_lock lk(mu);
|
||||
cv.wait(lk, [&] { return done(); });
|
||||
return made;
|
||||
}
|
||||
};
|
||||
|
||||
/// Whether a capture's copies go behind: helpers to take them, enough pages for them (copy_split's own threshold: under it the
|
||||
/// submitting thread copies all at once, cheaper than waking a helper), and few to copy ahead (the late pages and the texture
|
||||
/// storage: take_capture reads those at once; many of them, as at a scene's load, and the capture is copied whole as before)
|
||||
inline bool worth_behind(int helpers, int64_t pages, int64_t ahead, int64_t min_pages = 128, int64_t max_ahead = 512) {
|
||||
return helpers > 0 && pages - ahead >= min_pages && ahead <= max_ahead;
|
||||
}
|
||||
|
||||
} // namespace mkwvk::capture_behind
|
||||
110
gpu/vk/capture_tables.h
Normal file
110
gpu/vk/capture_tables.h
Normal file
@@ -0,0 +1,110 @@
|
||||
// (Triforce, 11-21, the upscaler audit's UPS-02) The stored draws' two tables, kept from one frame to the next.
|
||||
//
|
||||
// While NVIDIA DLSS, AMD FSR 4 or AMD FSR 3.1 upscales, every plain draw into the scene stores its positions and is drawn
|
||||
// again into the motion vectors (vk_temporal.cpp capture_motion). Two tables go with that, both filled again every frame:
|
||||
// the draws of a model so far this frame (which of them a draw is: the n-th this frame is the n-th the frame before), and
|
||||
// the frame before's draws by their key (where each one's positions of that frame are). Both were std::unordered_map,
|
||||
// emptied and filled each frame: a node from the heap for every entry and back to it a frame later, about 1,400 of each
|
||||
// a frame in the owner's log of 11-21 ("captured motion ... of 1444 stored a frame ... 0.650 ms CPU a frame"), on the
|
||||
// thread that records the frame. Here they keep their memory: the count in one open-addressed array (flat_map.h) that is
|
||||
// emptied in place, the frame before's draws in two lists that take turns, found through such an array. A frame of about
|
||||
// the size of the one before takes nothing from the heap.
|
||||
//
|
||||
// RX_VK_CAPTURE_TABLES=0: std::unordered_map again, a node an entry, as before (kept()).
|
||||
//
|
||||
// No Vulkan: tests/test_capture_tables.py compiles this and counts the heap's calls.
|
||||
#pragma once
|
||||
#include <cstddef>
|
||||
#include <cstdint>
|
||||
#include <cstdlib>
|
||||
#include <unordered_map>
|
||||
#include <utility>
|
||||
#include <vector>
|
||||
|
||||
#include "flat_map.h"
|
||||
|
||||
namespace capture_tables {
|
||||
|
||||
/// RX_VK_CAPTURE_TABLES (value, null when unset): kept unless it is 0
|
||||
inline bool kept(const char* value) { return !value || !*value || std::atoi(value) != 0; }
|
||||
|
||||
/// An array emptied for the next frame: in place, so the frame fills it without the heap; given back when it is far
|
||||
/// larger than what the frame put in it (a frame of many draws once, then few), so emptying it does not walk all of that
|
||||
template <class K, class V>
|
||||
inline void empty_for_next(FlatMap<K, V>& m) {
|
||||
if (m.capacity() > 8 * (m.size() + 512)) m.clear();
|
||||
else m.reset();
|
||||
}
|
||||
|
||||
/// The room a frame's list is given at its start: the frame before's count (kept), so it grows once and not a dozen times
|
||||
/// from nothing; 0, as before, with the switch off
|
||||
inline size_t room(size_t taken, bool keep) { return keep ? taken : 0; }
|
||||
|
||||
/// How many of a model's draws came so far this frame
|
||||
class Nth {
|
||||
public:
|
||||
explicit Nth(bool keep = true) : keep_(keep) {}
|
||||
/// The draws of `model` before this one, which is counted
|
||||
int next(uint64_t model) { return keep_ ? flat_[model]++ : node_[model]++; }
|
||||
void clear() {
|
||||
if (keep_) empty_for_next(flat_);
|
||||
else node_.clear();
|
||||
}
|
||||
size_t size() const { return keep_ ? flat_.size() : node_.size(); }
|
||||
|
||||
private:
|
||||
bool keep_;
|
||||
FlatMap<uint64_t, int> flat_;
|
||||
std::unordered_map<uint64_t, int> node_;
|
||||
};
|
||||
|
||||
/// A frame's entries by key, and the frame before's: turn() makes what was put since the last turn the frame before's,
|
||||
/// put() notes one of this frame's (the last with a key stands), before() finds the frame before's
|
||||
template <class V>
|
||||
class Before {
|
||||
public:
|
||||
explicit Before(bool keep = true) : keep_(keep) {}
|
||||
void turn() {
|
||||
if (!keep_) {
|
||||
node_old_ = std::move(node_now_);
|
||||
node_now_.clear();
|
||||
return;
|
||||
}
|
||||
std::swap(now_, old_), std::swap(now_at_, old_at_);
|
||||
now_.clear(); // (its memory kept)
|
||||
empty_for_next(now_at_);
|
||||
}
|
||||
void put(uint64_t key, const V& v) {
|
||||
if (!keep_) {
|
||||
node_now_[key] = v;
|
||||
return;
|
||||
}
|
||||
uint32_t& at = now_at_[key]; // (the entry's place in now_, plus one: 0 is none yet)
|
||||
if (at) now_[at - 1] = v;
|
||||
else now_.push_back(v), at = (uint32_t)now_.size();
|
||||
}
|
||||
/// The frame before's entry of `key`, null when it had none: good until the next turn() or clear()
|
||||
const V* before(uint64_t key) {
|
||||
if (!keep_) {
|
||||
const auto it = node_old_.find(key);
|
||||
return it == node_old_.end() ? nullptr : &it->second;
|
||||
}
|
||||
const auto* s = old_at_.find(key);
|
||||
return s ? &old_[s->second - 1] : nullptr;
|
||||
}
|
||||
/// Nothing of either frame (the chain let go: its memory too)
|
||||
void clear() {
|
||||
std::vector<V>().swap(now_), std::vector<V>().swap(old_);
|
||||
now_at_.clear(), old_at_.clear();
|
||||
node_now_.clear(), node_old_.clear();
|
||||
}
|
||||
size_t size() const { return keep_ ? now_.size() : node_now_.size(); }
|
||||
|
||||
private:
|
||||
bool keep_;
|
||||
std::vector<V> now_, old_;
|
||||
FlatMap<uint64_t, uint32_t> now_at_, old_at_;
|
||||
std::unordered_map<uint64_t, V> node_now_, node_old_;
|
||||
};
|
||||
|
||||
} // namespace capture_tables
|
||||
85
gpu/vk/checkpoint_kinds.h
Normal file
85
gpu/vk/checkpoint_kinds.h
Normal file
@@ -0,0 +1,85 @@
|
||||
/* (Triforce) The kinds of RX_VK_CHECKPOINTS markers (vk_draw.cpp checkpoint, checkpoint_mark) and what lies after each, for the
|
||||
device lost's report. The owner's test11 and test12 and two lab runs of five lost the device at the title the same way: the
|
||||
top and the bottom of the pipe both at "write-back of pool ...", the next marker the next frame's first dispatch, never
|
||||
reached. Between those two markers there was a pool's write-back copies, the write-back diff pass, the end of that recording,
|
||||
the whole present (its recording, the blit to the window's image after its acquire, the menu) and the next recording's start
|
||||
and uploads: none of them had a marker, and the report could not say which the GPU stopped in. Now each of those has one, and
|
||||
the report says what the work after the last marker reached is (gap). Plain functions: tests/test_checkpoint_kinds.py. */
|
||||
#pragma once
|
||||
|
||||
#include <cstdint>
|
||||
|
||||
namespace mkwvk::checkpoint_kinds {
|
||||
|
||||
/// A marker's kind as the report names it; its key follows the name in the report (a program, a pool, a texture, a recording's
|
||||
/// timeline value, a swapchain image's index or a chunk count, as the name says)
|
||||
inline const char* name(char c) {
|
||||
switch (c) {
|
||||
case 'd': return "draw of program";
|
||||
case 'c': return "dispatch of program";
|
||||
case 'i': return "INDIRECT dispatch of program";
|
||||
case 'm': return "mirror sync of pool";
|
||||
case 'b': return "write-back of pool";
|
||||
case 'u': return "upload of texture";
|
||||
case 'g': return "GPU unswizzle of texture";
|
||||
case 'w': return "write-back diff pass, chunks";
|
||||
case 's': return "mirror syncs' pass, chunks"; // ((Triforce, 11-21, MEM-48))
|
||||
case 'r': return "recording begun, value";
|
||||
case 'v': return "uploads of the recording, value";
|
||||
case 'z': return "recording ended, value";
|
||||
case 'p': return "present blit begun, image";
|
||||
case 'M': return "menu drawn over the present, image";
|
||||
case 'P': return "present blit ended, image";
|
||||
default: return "?";
|
||||
}
|
||||
}
|
||||
|
||||
/// What the GPU had to do after a marker of this kind, up to the next marker recorded: where a device lost whose last marker
|
||||
/// reached at the top of the pipe is this one stopped
|
||||
inline const char* gap(char c) {
|
||||
switch (c) {
|
||||
case 'd': case 'c': case 'i':
|
||||
return "that draw or dispatch itself (its marker is set just before it), or what its command buffer records after it before the "
|
||||
"next marker: barriers, copies, clears, queries";
|
||||
case 'm': return "that mirror sync's copies (guest memory -> the pool's VRAM mirror), in the recording's uploads";
|
||||
case 'b':
|
||||
return "that pool's write-back copies (its mirror -> guest memory, and the shadow's fill), then the write-back diff pass if "
|
||||
"one follows (w): with no w after it, the end of the recording (z)";
|
||||
case 'u': case 'g': return "that texture's upload or unswizzle itself";
|
||||
case 's':
|
||||
return "the mirror syncs' pass (one compute dispatch in the recording's uploads, reading a capture's buffer and writing "
|
||||
"the pool's VRAM mirror by device address; RX_VK_MSYNC_PASS=0: copy regions), then the sync's copies from guest memory";
|
||||
case 'w':
|
||||
return "the write-back diff pass (one compute dispatch over every pool's chunks, reading mirrors and shadows and writing guest "
|
||||
"memory by device address), then its barrier and the end of the recording (z)";
|
||||
case 'r': return "the recording's first commands: its barrier, and the replay's work before its first marked draw or dispatch";
|
||||
case 'v': return "the recording's uploads: mirror pieces filled, mirror syncs, the page table's changes, texture uploads";
|
||||
case 'z':
|
||||
return "the next submission: its waits on the queue (the mirrors' new pieces bound, a swapchain image acquired) or its first "
|
||||
"commands (r, v, or the present's p next)";
|
||||
case 'p': return "the present's blit: the game's picture scaled into the window's image, the HDR pass and the bars, before the menu (M)";
|
||||
case 'M': return "the menu's drawing (Dear ImGui) over the present, then its last barrier (P)";
|
||||
case 'P': return "the present of that image, then the next submission's waits and first commands";
|
||||
default: return "not known (a marker of no named kind)";
|
||||
}
|
||||
}
|
||||
|
||||
/// The two stages' last markers (sequence numbers; 0: none reached): one sentence of what they mean together
|
||||
inline const char* reading(uint32_t top, uint32_t bottom) {
|
||||
if (!top) return "no marker was reached at the top of the pipe";
|
||||
if (bottom == top)
|
||||
return "every command up to the last marker reached is done and nothing after it began: the GPU stopped in what follows that "
|
||||
"marker (below), or in a wait before the next one";
|
||||
if (bottom < top)
|
||||
return "the work between the bottom's last marker and the top's was running: the GPU stopped in it (a draw or dispatch that "
|
||||
"did not end, or a fault)";
|
||||
return "the bottom of the pipe is past the top: the markers are not in order (several queues or threads), read them with care";
|
||||
}
|
||||
/// (Triforce, P4) The two stages' last markers are of two threads' rings (vk_draw.cpp checkpoint_next: each recording thread
|
||||
/// numbers its own): their numbers say nothing of each other's order
|
||||
inline const char* apart() {
|
||||
return "the two are of two threads' recordings (each numbers its own markers: the replay thread's, a helper thread's, the "
|
||||
"present's): their numbers say nothing of their order, read each with the markers around it (below)";
|
||||
}
|
||||
|
||||
} // namespace mkwvk::checkpoint_kinds
|
||||
128
gpu/vk/cmd_work.h
Normal file
128
gpu/vk/cmd_work.h
Normal file
@@ -0,0 +1,128 @@
|
||||
// (Triforce) What the replay's barrier rules ask of the command buffer they record into, kept per command buffer: how many
|
||||
// commands went into it, and where its last full barrier, its last draw pass's end and its last dispatch were (vk_work.h counts
|
||||
// the commands; vk_texture.cpp full_barrier, pass_start_barrier and pass_resume_barrier, vk_draw.cpp dispatch read and set the
|
||||
// marks). Apart from Vulkan's calls so that tests/test_barrier_rules.py, test_pass_barrier.py and test_dispatch_rule_threads.py can
|
||||
// run it.
|
||||
//
|
||||
// After NX2navi's per-context state (runtime/vk_exec.h VkExec.recorded, barrier_at, bar_dirty, bar_passes, bar_dispatches;
|
||||
// vk_exec.c barrier_to_why, chunk_record, group_put): there a recording context counts its own commands, so its rules hold inside
|
||||
// the worker threads' chunks. Here one count served every thread (work_epoch, an atomic add on every command from every recording
|
||||
// thread), so each rule stood aside while the Replay helper threads recorded (most frames at the Market: about 500 full barriers a
|
||||
// frame, no pass start in its fragment stages only). A command buffer is recorded by one thread at a time (Vulkan asks for that),
|
||||
// and its state here is only touched by whoever records into it, so nothing in it needs to be atomic; the table that finds it is.
|
||||
// Another thread's commands go into its own command buffers and leave this one's marks alone; a command buffer two threads take
|
||||
// turns at (a parallel replay's segment init command buffer, vk_texture.cpp InitRedirect) has every turn's commands counted in it.
|
||||
//
|
||||
// RX_VK_BARRIER_CONTEXT=0: the one count of every thread's commands, and the rules standing aside while the helper threads record,
|
||||
// as before (vk_texture.cpp rules_aside).
|
||||
#pragma once
|
||||
|
||||
#include <vulkan/vulkan_core.h>
|
||||
|
||||
#include <atomic>
|
||||
#include <cstddef>
|
||||
#include <cstdint>
|
||||
#include <cstdlib>
|
||||
#include <cstring>
|
||||
|
||||
namespace mkwvk {
|
||||
|
||||
extern std::atomic<uint64_t> work_epoch; // (vk_texture.cpp: RX_VK_BARRIER_CONTEXT=0's count, every thread's commands)
|
||||
inline thread_local uint64_t work_mine = 0; // (this thread's commands, into any command buffer: work_between)
|
||||
inline std::atomic<uint64_t> work_submits{0}; // (every queue submit through vk_work.h, any thread's: work_between)
|
||||
|
||||
/// RX_VK_BARRIER_CONTEXT=0: the global count, as before; anything else, or unset: per command buffer. Read at the first command
|
||||
/// (0: not read yet, 1: per command buffer, 2: global)
|
||||
inline std::atomic<int> work_way{0};
|
||||
inline bool work_global() {
|
||||
int w = work_way.load(std::memory_order_relaxed);
|
||||
if (!w) [[unlikely]] {
|
||||
const char* e = std::getenv("RX_VK_BARRIER_CONTEXT");
|
||||
w = e && std::strcmp(e, "0") == 0 ? 2 : 1;
|
||||
work_way.store(w, std::memory_order_relaxed);
|
||||
}
|
||||
return w == 2;
|
||||
}
|
||||
|
||||
/// One command buffer's state. The marks are counts of it (work_at) taken right after the command they mark; a mark equal to the
|
||||
/// count now: nothing was recorded into it since
|
||||
struct CmdWork {
|
||||
uint64_t n = 0; // (its commands, and its begins: work_begun)
|
||||
uint64_t barrier_at = UINT64_MAX; // (right after its last full barrier: full_barrier leaves out a second one there)
|
||||
uint64_t pass_end_at = UINT64_MAX; // (right after its last draw pass ended, unless the replay split that pass for a hazard)
|
||||
uint64_t dispatch_at = UINT64_MAX; // (right after its last dispatch: RX_VK_DISPATCH_BARRIER)
|
||||
bool known = true; // (false: the stand-in for a command buffer the table had no room for, never right after anything)
|
||||
};
|
||||
|
||||
/// The table: a slot a command buffer, taken at its first command or begin by whichever thread comes first and kept for good (the
|
||||
/// replay makes its command buffers once: a few hundred at most). Full: cmd_work_none, which no rule relaxes on
|
||||
constexpr size_t kCmdWorkSlots = 1024;
|
||||
struct alignas(64) CmdWorkSlot { // (a cache line each: two threads recording two command buffers share none)
|
||||
std::atomic<VkCommandBuffer> cmd{VK_NULL_HANDLE};
|
||||
CmdWork w;
|
||||
};
|
||||
inline CmdWorkSlot cmd_work_slots[kCmdWorkSlots];
|
||||
inline thread_local CmdWork cmd_work_none{0, UINT64_MAX, UINT64_MAX, UINT64_MAX, false}; // (vk_texture.cpp full_barrier tells it once)
|
||||
inline CmdWork* cmd_work_find(VkCommandBuffer cmd) {
|
||||
if (cmd == VK_NULL_HANDLE) return &cmd_work_none;
|
||||
const size_t h = (size_t)(((uint64_t)(uintptr_t)cmd * 0x9E3779B97F4A7C15ull) >> 54); // (10 bits: kCmdWorkSlots)
|
||||
for (size_t k = 0; k < kCmdWorkSlots; k++) {
|
||||
CmdWorkSlot& s = cmd_work_slots[(h + k) % kCmdWorkSlots];
|
||||
VkCommandBuffer at = s.cmd.load(std::memory_order_acquire);
|
||||
if (at == VK_NULL_HANDLE && s.cmd.compare_exchange_strong(at, cmd, std::memory_order_acq_rel)) return &s.w;
|
||||
if (at == cmd) return &s.w; // (its own, or taken by another thread just now)
|
||||
}
|
||||
return &cmd_work_none;
|
||||
}
|
||||
/// cmd's state (the last one this thread asked for is at hand: most commands go into the command buffer of the one before)
|
||||
inline thread_local VkCommandBuffer cmd_work_last = VK_NULL_HANDLE;
|
||||
inline thread_local CmdWork* cmd_work_last_w = nullptr;
|
||||
inline CmdWork& cmd_work(VkCommandBuffer cmd) {
|
||||
if (cmd != cmd_work_last || !cmd_work_last_w) [[unlikely]] cmd_work_last_w = cmd_work_find(cmd), cmd_work_last = cmd;
|
||||
return *cmd_work_last_w;
|
||||
}
|
||||
|
||||
/// The count a mark is taken in: the command buffer's own; RX_VK_BARRIER_CONTEXT=0, every thread's (work_epoch), as before
|
||||
inline uint64_t work_at(const CmdWork& w) {
|
||||
return work_global() ? work_epoch.load(std::memory_order_relaxed) : w.n;
|
||||
}
|
||||
/// Whether nothing was recorded into the command buffer since the mark was taken
|
||||
inline bool work_right_after(const CmdWork& w, uint64_t mark) {
|
||||
return w.known && mark == work_at(w);
|
||||
}
|
||||
|
||||
/// A command that does or orders GPU work, into cmd (vk_work.h's macros; by hand after a library recorded into cmd itself)
|
||||
inline void note_work(VkCommandBuffer cmd) {
|
||||
// (RX_VK_BARRIER_CONTEXT=0: an add, not a load and a store: a thread preempted between the two would store a stale count over
|
||||
// another thread's, which could make the count equal the one a barrier was recorded at with work recorded since)
|
||||
if (work_global()) work_epoch.fetch_add(1, std::memory_order_relaxed);
|
||||
else cmd_work(cmd).n++;
|
||||
work_mine++;
|
||||
}
|
||||
inline VkCommandBuffer work_on(VkCommandBuffer cmd) {
|
||||
note_work(cmd);
|
||||
return cmd;
|
||||
}
|
||||
/// cmd begun (vkBeginCommandBuffer): a new recording starts afresh, whatever its handle held before (its last command may have
|
||||
/// been a full barrier, and the work submitted since then is not after it)
|
||||
inline VkCommandBuffer work_begun(VkCommandBuffer cmd) {
|
||||
CmdWork& w = cmd_work(cmd);
|
||||
w.n++;
|
||||
w.barrier_at = w.pass_end_at = w.dispatch_at = UINT64_MAX;
|
||||
return cmd;
|
||||
}
|
||||
/// A submit to the queue (vkQueueSubmit2), any thread's
|
||||
inline VkQueue work_submitted(VkQueue q) {
|
||||
work_submits.fetch_add(1, std::memory_order_relaxed);
|
||||
return q;
|
||||
}
|
||||
/// Whether anything was recorded or submitted between two points of this thread's recording: equal before and after, nothing.
|
||||
/// What crosses command buffers asks this (vk_replay.cpp maybe_split: a pass going on in the next part; finish_recording: the
|
||||
/// early tail's write-back): this thread's own commands, into any command buffer, and every submit to the queue, any thread's
|
||||
/// (another thread's recording reaches the queue between two of this thread's command buffers only by a submit).
|
||||
/// RX_VK_BARRIER_CONTEXT=0: every thread's commands, as before
|
||||
inline uint64_t work_between() {
|
||||
return work_global() ? work_epoch.load(std::memory_order_relaxed) : work_mine + work_submits.load(std::memory_order_relaxed);
|
||||
}
|
||||
|
||||
} // namespace mkwvk
|
||||
120
gpu/vk/commit_room.h
Normal file
120
gpu/vk/commit_room.h
Normal file
@@ -0,0 +1,120 @@
|
||||
// (Triforce, 11-22, the performance audit's OS-05, after Source A's) The textures give commit back near Windows' commit limit,
|
||||
// and the page files' room to grow counts as commit still to come.
|
||||
//
|
||||
// Windows charges every block of a discrete card's memory to the process's commit, as it does the game's own memory. Where the
|
||||
// system's commit runs out (RAM and the page files full) an allocation is refused, and the game ends unless the refusal passes
|
||||
// (OS-04's wait, HOS_COMMIT_WAIT_MS, helps only if commit comes back). Source A takes the shortfall under a reserve of free
|
||||
// commit (2,560 MB) off the texture budget on any card, so that eviction and compaction give blocks back. Its first rule fired
|
||||
// while a page file Windows manages could still grow (292,526 textures retired in 90 minutes, black textures); the fix counts
|
||||
// each page file's room to grow, capped by its drive's free space, as commit still free.
|
||||
//
|
||||
// The rules here, apart from Windows so that tests/test_commit_room.py can run them:
|
||||
// - reserve_of: RX_VK_TEX_COMMIT_RESERVE_MB, the reserve. Off (0) unless set: the entry's 2560 is not measured on this tree
|
||||
// - growth_on: RX_VK_TEX_COMMIT_GROWTH, whether the growth is counted: 0 never, another number always, unset where a reserve is
|
||||
// - low_line, read_growth: the growth is read only while less is free than the higher of the reserve and the line of
|
||||
// tex_pressure::commit_low (whose fall-back to the fixed size then gets the growth too: vk_texture.cpp commit_now)
|
||||
// - cap: the textures at most what they hold less the shortfall under the reserve, never under 1 GB (Source A has no floor,
|
||||
// and its false cut went to 0)
|
||||
// - growth: the page files' room to grow, from the records of NtQuerySystemInformation's class 0x90 (undocumented: each
|
||||
// record is checked against the buffer, and one that is not inside it ends the walk), a drive's page files capped together
|
||||
// by its free space. vk_texture.cpp page_file_growth asks Windows, at most once a second, and answers 0 on any failure.
|
||||
// Written from the audit's description: Source A's file is not on this PC.
|
||||
#pragma once
|
||||
|
||||
#include <algorithm>
|
||||
#include <cstddef>
|
||||
#include <cstdint>
|
||||
#include <cstdlib>
|
||||
#include <cstring>
|
||||
|
||||
namespace commit_room {
|
||||
|
||||
inline constexpr uint64_t kFloor = 1ull << 30; // (the textures are never capped under 1 GB)
|
||||
inline constexpr int64_t kMostMb = 65536; // (the reserve at most)
|
||||
|
||||
/// RX_VK_TEX_COMMIT_RESERVE_MB in bytes: 0 (unset, empty, no number, 0 or less): no cut, as before; at most kMostMb MB
|
||||
inline uint64_t reserve_of(const char* env) {
|
||||
if (!env || !*env) return 0;
|
||||
const long long mb = std::atoll(env);
|
||||
return mb <= 0 ? 0 : (uint64_t)std::min<long long>(mb, kMostMb) << 20;
|
||||
}
|
||||
|
||||
/// RX_VK_TEX_COMMIT_GROWTH: unset or empty, counted where a reserve is set; 0 (or no number), never; another number, always
|
||||
inline bool growth_on(const char* env, uint64_t reserve) {
|
||||
if (!env || !*env) return reserve != 0;
|
||||
return std::atoi(env) != 0;
|
||||
}
|
||||
|
||||
/// tex_pressure::commit_low's line in bytes: want_mb MB; -1 (less than 0) the larger of 4 GB and a tenth of the limit; 0, or a
|
||||
/// limit not known (0), none
|
||||
inline uint64_t low_line(uint64_t limit, int64_t want_mb) {
|
||||
if (!limit || !want_mb) return 0;
|
||||
return want_mb > 0 ? (uint64_t)want_mb << 20 : std::max<uint64_t>(4096ull << 20, limit / 10);
|
||||
}
|
||||
|
||||
/// Whether the page files' growth is read at this look: counted (on), and less free (avail) than the higher of the reserve and
|
||||
/// commit_low's line (low)
|
||||
inline bool read_growth(bool on, uint64_t avail, uint64_t reserve, uint64_t low) { return on && avail < std::max(reserve, low); }
|
||||
|
||||
/// The textures' cap: what they hold (total) less the shortfall of what is free (avail, the growth in it) under the reserve,
|
||||
/// never under kFloor. UINT64_MAX (no cap) with no reserve, or with the reserve free
|
||||
inline uint64_t cap(uint64_t total, uint64_t avail, uint64_t reserve) {
|
||||
if (!reserve || avail >= reserve) return UINT64_MAX;
|
||||
const uint64_t shortfall = reserve - avail;
|
||||
return std::max(kFloor, total > shortfall ? total - shortfall : 0);
|
||||
}
|
||||
|
||||
/// NtQuerySystemInformation's class for the page files' records (SystemPageFileInformationEx)
|
||||
inline constexpr unsigned kPageFileClass = 0x90;
|
||||
|
||||
/// One record of that class as 64-bit Windows lays it out; sizes in pages
|
||||
struct PageFileRecord {
|
||||
uint32_t next; // (bytes from this record to the next; 0: the last)
|
||||
uint32_t total_pages; // (its size now)
|
||||
uint32_t in_use_pages, peak_pages;
|
||||
uint16_t name_bytes, name_room; // (the name: UTF-16, such as \??\C:\pagefile.sys, in the same buffer)
|
||||
const uint16_t* name;
|
||||
uint32_t min_pages, max_pages; // (the size it may grow to)
|
||||
};
|
||||
static_assert(sizeof(void*) != 8 || sizeof(PageFileRecord) == 40, "the record as 64-bit Windows lays it out");
|
||||
|
||||
/// The drive letter of a page file's name ('C' for \??\C:\pagefile.sys), 0 where it has none
|
||||
inline char drive_of(const uint16_t* name, size_t chars) {
|
||||
for (size_t i = 0; i + 1 < chars; i++) {
|
||||
const uint16_t c = name[i];
|
||||
if (name[i + 1] == ':' && ((c >= 'A' && c <= 'Z') || (c >= 'a' && c <= 'z'))) return (char)(c >= 'a' ? c - 32 : c);
|
||||
}
|
||||
return 0;
|
||||
}
|
||||
|
||||
/// The page files' room to grow in bytes, from the records in buf[0, len): a record's (max - size) pages of `page` bytes, the
|
||||
/// records of one drive summed and capped by that drive's free space (free_of('C'), bytes). A record that is not wholly inside
|
||||
/// the buffer, whose name is not, or whose next one would overlap it, ends the walk: what came before it counts
|
||||
template <class FreeOf>
|
||||
inline uint64_t growth(const uint8_t* buf, size_t len, uint64_t page, FreeOf free_of) {
|
||||
uint64_t by_drive[26] = {};
|
||||
const uintptr_t base = (uintptr_t)buf;
|
||||
for (size_t at = 0; len >= sizeof(PageFileRecord) && at <= len - sizeof(PageFileRecord);) {
|
||||
PageFileRecord r;
|
||||
std::memcpy(&r, buf + at, sizeof r);
|
||||
char drive = 0;
|
||||
if (r.name_bytes) {
|
||||
const uintptr_t n = (uintptr_t)r.name;
|
||||
if (n < base || n - base > len || r.name_bytes > len - (n - base)) break;
|
||||
uint16_t chars[64] = {};
|
||||
const size_t count = std::min<size_t>(r.name_bytes / 2, 64);
|
||||
std::memcpy(chars, buf + (n - base), count * 2);
|
||||
drive = drive_of(chars, count);
|
||||
}
|
||||
if (drive && r.max_pages > r.total_pages) by_drive[drive - 'A'] += (uint64_t)(r.max_pages - r.total_pages) * page;
|
||||
if (!r.next) break;
|
||||
if (r.next < sizeof r || r.next > len - at) break;
|
||||
at += r.next;
|
||||
}
|
||||
uint64_t sum = 0;
|
||||
for (int d = 0; d < 26; d++)
|
||||
if (by_drive[d]) sum += std::min<uint64_t>(by_drive[d], free_of((char)('A' + d)));
|
||||
return sum;
|
||||
}
|
||||
|
||||
} // namespace commit_room
|
||||
125
gpu/vk/compute_pin.h
Normal file
125
gpu/vk/compute_pin.h
Normal file
@@ -0,0 +1,125 @@
|
||||
// Which compute programs are held to 32 lanes (requiredSubgroupSize), by what the program can tell (no game, no GPU: a rule).
|
||||
//
|
||||
// (Triforce, 11-22, the performance audit's PIPE-52, after Source A's and Source B's.) The console's programs are written
|
||||
// for warps of 32 lanes, so every compute pipeline asked for subgroups of 32 (vk_draw.cpp pin_subgroup). Most programs
|
||||
// cannot tell how many lanes run beside them: they never vote, never read a lane mask or their lane's index, never
|
||||
// shuffle between lanes. Mesa's Intel driver compiles a pinned program for 32 lanes only (SIMD32), which on its large
|
||||
// programs is minutes and gigabytes: Source A measured a 730 KB program at 0.8-1 GB over 23-25 s pinned and at most 300 MB
|
||||
// unpinned (an Iris Xe, Mesa 26.2.4), with 271 of 1046 variants still pinned by this rule.
|
||||
//
|
||||
// A program is pinned when it can see its lanes:
|
||||
// - a vote, a lane mask, the lane's index, a shuffle, FSWZADD or a matrix multiply (Shader::Info's flags);
|
||||
// - a barrier of Subgroup scope in its module: this translator's own case, the barrier its lowering puts between a
|
||||
// warp's shared-memory store and load, which takes the subgroup to be the warp. No flag records it, so the module is
|
||||
// looked at (subgroup_barrier below), and a module nobody looked at counts as having one;
|
||||
// - shared memory in a workgroup of 32 invocations or fewer, where the workgroup is one warp and its lanes run in step.
|
||||
// Everything else gets the driver's own choice: no required size, and no full-subgroups flag either.
|
||||
//
|
||||
// RX_VK_COMPUTE_PIN, read once:
|
||||
// all every compute program pinned, as it was (the default on every driver but Mesa's Intel one)
|
||||
// lanes the rule above (the default on Mesa's Intel driver)
|
||||
// shared the rule above, and any shared memory pins, whatever the workgroup: for a program whose warps keep
|
||||
// shared memory in step in a workgroup over 32, which nothing else here sees
|
||||
// A doubt pins: an unknown value of the switch is the driver's default, a program without a stage is pinned.
|
||||
// The graphics stages and the mesh fallback's compute stages are not asked here: they stay with pin_subgroup.
|
||||
#pragma once
|
||||
#include <cstddef>
|
||||
#include <cstdint>
|
||||
#include <cstring>
|
||||
|
||||
namespace compute_pin {
|
||||
enum class Mode { All, Lanes, Shared };
|
||||
|
||||
/// What one compute program is, as the rule needs it
|
||||
struct Facts {
|
||||
bool has_stage = false; // (false: nothing known, pinned)
|
||||
bool vote = false; // Shader::Info::uses_subgroup_vote
|
||||
bool mask = false; // uses_subgroup_mask
|
||||
bool lane_index = false; // uses_subgroup_invocation_id
|
||||
bool shuffles = false; // uses_subgroup_shuffles
|
||||
bool fswzadd = false; // uses_fswzadd
|
||||
bool matrices = false; // uses_cooperative_matrix
|
||||
bool subgroup_barrier = true; // its module has an OpControlBarrier of Subgroup scope (true until a look says no)
|
||||
uint32_t shared_memory = 0; // bytes of shared memory
|
||||
uint64_t invocations = 0; // the workgroup's x * y * z
|
||||
};
|
||||
|
||||
/// The program can tell how many lanes run in its subgroup, or which one it is
|
||||
constexpr bool sees_lanes(const Facts& f) {
|
||||
return f.vote || f.mask || f.lane_index || f.shuffles || f.fswzadd || f.matrices || f.subgroup_barrier;
|
||||
}
|
||||
|
||||
/// RX_VK_COMPUTE_PIN's value (nullptr: unset) on this driver
|
||||
inline Mode mode_for(const char* env, bool mesa_intel) {
|
||||
if (env && !std::strcmp(env, "all")) return Mode::All;
|
||||
if (env && !std::strcmp(env, "lanes")) return Mode::Lanes;
|
||||
if (env && !std::strcmp(env, "shared")) return Mode::Shared;
|
||||
return mesa_intel ? Mode::Lanes : Mode::All;
|
||||
}
|
||||
|
||||
inline const char* name(Mode m) {
|
||||
return m == Mode::All ? "all" : m == Mode::Lanes ? "lanes" : "shared";
|
||||
}
|
||||
|
||||
/// The program's pipeline asks for subgroups of 32
|
||||
constexpr bool needed(Mode m, const Facts& f) {
|
||||
if (m == Mode::All || !f.has_stage) return true;
|
||||
if (sees_lanes(f)) return true;
|
||||
if (f.shared_memory) return m == Mode::Shared || f.invocations <= 32;
|
||||
return false;
|
||||
}
|
||||
|
||||
/// Whether a SPIR-V module's entry point is a compute shader's (OpEntryPoint with the GLCompute model): the only modules
|
||||
/// the rule is asked about, so the only ones worth a note. One that does not read as SPIR-V: no
|
||||
inline bool compute_entry(const uint32_t* code, size_t words) {
|
||||
constexpr uint32_t kMagic = 0x07230203, kOpEntryPoint = 15, kOpFunction = 54, kGLCompute = 5;
|
||||
if (!code || words < 5 || code[0] != kMagic) return false;
|
||||
for (size_t at = 5; at < words;) {
|
||||
const uint32_t count = code[at] >> 16, op = code[at] & 0xffffu;
|
||||
if (!count || at + count > words || op == kOpFunction) return false;
|
||||
if (op == kOpEntryPoint && count >= 2) return code[at + 1] == kGLCompute;
|
||||
at += count;
|
||||
}
|
||||
return false;
|
||||
}
|
||||
|
||||
/// Whether a SPIR-V module has an OpControlBarrier whose execution scope is the constant Subgroup (3). A module that does
|
||||
/// not read as SPIR-V counts as having one. words: the module's 32-bit words
|
||||
inline bool subgroup_barrier(const uint32_t* code, size_t words) {
|
||||
constexpr uint32_t kMagic = 0x07230203, kOpConstant = 43, kOpControlBarrier = 224, kScopeSubgroup = 3;
|
||||
if (!code || words < 5 || code[0] != kMagic) return true;
|
||||
// (the scopes are constants, and constants come before the functions: one pass for the barriers' scope ids, kept small,
|
||||
// and their values looked up at the end)
|
||||
constexpr size_t kMost = 64;
|
||||
uint32_t scopes[kMost];
|
||||
size_t n_scopes = 0;
|
||||
for (size_t at = 5; at < words;) {
|
||||
const uint32_t count = code[at] >> 16, op = code[at] & 0xffffu;
|
||||
if (!count || at + count > words) return true;
|
||||
if (op == kOpControlBarrier && count >= 4) {
|
||||
const uint32_t id = code[at + 1];
|
||||
bool seen = false;
|
||||
for (size_t i = 0; i < n_scopes; i++) seen = seen || scopes[i] == id;
|
||||
if (!seen) {
|
||||
if (n_scopes == kMost) return true; // (more scope ids than a module has: a doubt)
|
||||
scopes[n_scopes++] = id;
|
||||
}
|
||||
}
|
||||
at += count;
|
||||
}
|
||||
if (!n_scopes) return false;
|
||||
size_t resolved = 0;
|
||||
for (size_t at = 5; at < words;) {
|
||||
const uint32_t count = code[at] >> 16, op = code[at] & 0xffffu;
|
||||
if (op == kOpConstant && count >= 4) {
|
||||
for (size_t i = 0; i < n_scopes; i++)
|
||||
if (scopes[i] == code[at + 2]) {
|
||||
if (code[at + 3] == kScopeSubgroup) return true;
|
||||
resolved++;
|
||||
}
|
||||
}
|
||||
at += count;
|
||||
}
|
||||
return resolved != n_scopes; // (a scope that is no plain constant: a doubt)
|
||||
}
|
||||
} // namespace compute_pin
|
||||
41
gpu/vk/cpu_count.h
Normal file
41
gpu/vk/cpu_count.h
Normal file
@@ -0,0 +1,41 @@
|
||||
// (Triforce, 11-22, the performance audit's OS-49, after Source B's) The logical processors the renderer sizes its pools and
|
||||
// its threads' counts by: std::thread::hardware_concurrency, or, while the lab's HOS_SIM_CPU=<cores>/<threads> holds the
|
||||
// process to fewer of the PC's cores (host/hos_sim_cpu.h), its <threads>: an affinity mask alone does not change what
|
||||
// hardware_concurrency answers, and the pools would be a 24-thread PC's on 6 cores. The host clears the variable when it did
|
||||
// not apply it, so the two agree. Windows only, as the host's part; unset, hardware_concurrency as before. Every such count in
|
||||
// gpu/ goes through cpu_count() but vk_shader.cpp's (a file the shader cache's stamp is made from):
|
||||
// tests/test_sim_cores.py checks that, and the rule.
|
||||
#pragma once
|
||||
#include <cstdlib>
|
||||
#include <thread>
|
||||
|
||||
namespace cpu_count_rule {
|
||||
/// The threads HOS_SIM_CPU asks for ("6/12": 12; "6": 6), when fewer than the PC's hw (0: not known); else 0: not set, not
|
||||
/// "<cores>[/<threads>]" with each from 1 to 256 and threads not under cores, or not fewer
|
||||
inline unsigned sim_threads(const char* e, unsigned hw) {
|
||||
if (!e || !*e) return 0;
|
||||
char* end = nullptr;
|
||||
const long c = std::strtol(e, &end, 10);
|
||||
if (end == e) return 0;
|
||||
long t = c;
|
||||
if (*end == '/') {
|
||||
const char* s = end + 1;
|
||||
t = std::strtol(s, &end, 10);
|
||||
if (end == s) return 0;
|
||||
}
|
||||
if (*end || c < 1 || c > 256 || t < c || t > 256) return 0;
|
||||
return hw && (unsigned)t >= hw ? 0 : (unsigned)t;
|
||||
}
|
||||
} // namespace cpu_count_rule
|
||||
|
||||
/// The logical processors to size by (read once)
|
||||
inline unsigned cpu_count() {
|
||||
static const unsigned n = [] {
|
||||
const unsigned hw = std::thread::hardware_concurrency();
|
||||
#ifdef _WIN32
|
||||
if (const unsigned sim = cpu_count_rule::sim_threads(std::getenv("HOS_SIM_CPU"), hw)) return sim;
|
||||
#endif
|
||||
return hw;
|
||||
}();
|
||||
return n;
|
||||
}
|
||||
39
gpu/vk/dead_count.h
Normal file
39
gpu/vk/dead_count.h
Normal file
@@ -0,0 +1,39 @@
|
||||
// (Triforce, 11-21, the performance audit's THREAD-21, its first step) What destroying the dead images costs a present.
|
||||
//
|
||||
// A texture let go waits in a list until the GPU is done with it, and texture_frame then destroys every one that is, in
|
||||
// the middle of the present being recorded: a render target's image goes to the pool (cheap), every other image is
|
||||
// destroyed with its views. The audit's port bounds that to a share a present, after the present's submit, as Source B
|
||||
// does on its GPU thread's idle moments. Its own proof asks for the cost first: "if no present spends 3 ms in the loop,
|
||||
// stop". This is that count, and nothing else changes: a line for a present whose loop took kSayMs or more, and every
|
||||
// 600 presents the totals and the worst present.
|
||||
//
|
||||
// No Vulkan here: tests/test_dead_count.py compiles it.
|
||||
#pragma once
|
||||
#include <cstdint>
|
||||
|
||||
namespace dead_count {
|
||||
|
||||
inline constexpr double kSayMs = 2.0; // (a present's loop that gets a line of its own)
|
||||
inline constexpr double kWorthMs = 3.0; // (the audit's line: under it in every present, the bounded step is not built)
|
||||
inline constexpr uint64_t kReport = 600; // (presents between the totals' lines)
|
||||
|
||||
struct Counts {
|
||||
uint64_t presents = 0, destroyed = 0, pooled = 0, over_say = 0, over_worth = 0;
|
||||
double ms = 0, worst_ms = 0;
|
||||
uint64_t worst_destroyed = 0, worst_left = 0;
|
||||
/// A present's loop: n_destroyed images destroyed and n_pooled put in the pool in loop_ms, left still waiting. True
|
||||
/// when the present gets a line of its own
|
||||
bool note(uint64_t n_destroyed, uint64_t n_pooled, double loop_ms, uint64_t left) {
|
||||
presents++, destroyed += n_destroyed, pooled += n_pooled, ms += loop_ms;
|
||||
if (loop_ms > worst_ms) worst_ms = loop_ms, worst_destroyed = n_destroyed, worst_left = left;
|
||||
if (loop_ms >= kWorthMs) over_worth++;
|
||||
if (loop_ms < kSayMs) return false;
|
||||
over_say++;
|
||||
return true;
|
||||
}
|
||||
/// Whether the totals' line is due, at present `present`
|
||||
bool due(uint64_t present) const { return presents != 0 && present % kReport == 0; }
|
||||
void reset() { *this = Counts{}; }
|
||||
};
|
||||
|
||||
} // namespace dead_count
|
||||
333
gpu/vk/debug_names.h
Normal file
333
gpu/vk/debug_names.h
Normal file
@@ -0,0 +1,333 @@
|
||||
// (Triforce) The names on the Debug tab's Game flags, Cutscenes and Scan amiibo pages (vk_menu.cpp draws them), and what
|
||||
// a page keeps: the list in groups, the choice, the cards shown. No Vulkan, no ImGui and no file here:
|
||||
// tests/test_debug_names.py compiles this header and runs it on bytes of its own making.
|
||||
//
|
||||
// The game knows a flag (a Bool of its GameData) by the MurmurHash3 of its name and nothing else; its files list the
|
||||
// Bools by hash (GameData/GameDataList.Product.<n>.byml.zs) and use their names in the event flows
|
||||
// (Event/EventFlow/<event>.bfevfl.zs). So the flags' names are every text of the event flows whose hash is a Bool's, and
|
||||
// the events are the event flows that hold a script (a flowchart: the others are a cutscene's timelines, which a script
|
||||
// plays). After NX2navi's F9 panel, which found its names the same way; no line of it is here, and no name: both lists
|
||||
// are made in memory from the player's own game files each run, and written nowhere.
|
||||
#pragma once
|
||||
|
||||
#include <algorithm>
|
||||
#include <cstdint>
|
||||
#include <cstring>
|
||||
#include <string>
|
||||
#include <vector>
|
||||
|
||||
#include "oot_debug_link.h"
|
||||
|
||||
namespace mkwvk::debug_names {
|
||||
|
||||
/// A BYML file's Data.Bool[i].Hash values, sorted, with no number twice: the Bools of a GameDataList (either byte order;
|
||||
/// the game's is the big-endian one). false when the bytes are not such a file. Nothing outside the bytes is read
|
||||
inline bool gamedata_bools(const std::vector<unsigned char>& b, std::vector<uint32_t>& out) {
|
||||
out.clear();
|
||||
const size_t n = b.size();
|
||||
if (n < 16 || !((b[0] == 'B' && b[1] == 'Y') || (b[0] == 'Y' && b[1] == 'B'))) return false;
|
||||
const bool big = b[0] == 'B';
|
||||
bool bad = false; // (a read outside the file: everything read after it is 0, and the file is refused)
|
||||
auto u8 = [&](size_t o) -> uint32_t { return o < n ? b[o] : (bad = true, 0u); };
|
||||
auto u24 = [&](size_t o) -> uint32_t {
|
||||
if (o > n || n - o < 3) return bad = true, 0u;
|
||||
return big ? (uint32_t)b[o] << 16 | (uint32_t)b[o + 1] << 8 | b[o + 2] : (uint32_t)b[o + 2] << 16 | (uint32_t)b[o + 1] << 8 | b[o];
|
||||
};
|
||||
auto u32 = [&](size_t o) -> uint32_t {
|
||||
if (o > n || n - o < 4) return bad = true, 0u;
|
||||
return big ? (uint32_t)b[o] << 24 | (uint32_t)b[o + 1] << 16 | (uint32_t)b[o + 2] << 8 | b[o + 3]
|
||||
: (uint32_t)b[o + 3] << 24 | (uint32_t)b[o + 2] << 16 | (uint32_t)b[o + 1] << 8 | b[o];
|
||||
};
|
||||
// the keys' table: 0xc2, how many, an offset each (and one past the last) from the table's start, the texts
|
||||
const size_t keys = u32(4), root = u32(12);
|
||||
if (bad || !keys || u8(keys) != 0xc2) return false;
|
||||
const uint32_t key_count = u24(keys + 1);
|
||||
auto key_index = [&](const char* name) -> int64_t {
|
||||
const size_t len = strlen(name);
|
||||
for (uint32_t i = 0; i < key_count && !bad; i++) {
|
||||
const size_t at = keys + (size_t)u32(keys + 4 + 4 * (size_t)i);
|
||||
if (!bad && at < n && n - at > len && !memcmp(b.data() + at, name, len + 1)) return i;
|
||||
}
|
||||
return -1;
|
||||
};
|
||||
const int64_t k_data = key_index("Data"), k_bool = key_index("Bool"), k_hash = key_index("Hash");
|
||||
if (bad || k_data < 0 || k_bool < 0 || k_hash < 0) return false;
|
||||
// a dictionary: 0xc1, how many, then eight bytes each: the key's index (three bytes), the value's type, the value
|
||||
auto member = [&](size_t dict, int64_t key, uint32_t* type) -> uint32_t {
|
||||
if (u8(dict) != 0xc1) return bad = true, 0u;
|
||||
const uint32_t count = u24(dict + 1);
|
||||
for (uint32_t i = 0; i < count && !bad; i++) {
|
||||
const size_t e = dict + 4 + 8 * (size_t)i;
|
||||
if (u24(e) == (uint32_t)key) return *type = u8(e + 3), u32(e + 4);
|
||||
}
|
||||
return *type = 0, 0u;
|
||||
};
|
||||
uint32_t type = 0;
|
||||
const size_t data = member(root, k_data, &type);
|
||||
if (bad || type != 0xc1) return false;
|
||||
const size_t bools = member(data, k_bool, &type);
|
||||
if (bad || type != 0xc0 || u8(bools) != 0xc0) return false;
|
||||
// an array: 0xc0, how many, a type byte each (to a multiple of four), then four bytes each
|
||||
const uint32_t count = u24(bools + 1);
|
||||
const size_t values = bools + 4 + (((size_t)count + 3) & ~(size_t)3);
|
||||
if (bad || count > n / 4) return false;
|
||||
out.reserve(count);
|
||||
for (uint32_t i = 0; i < count; i++) {
|
||||
if (u8(bools + 4 + i) != 0xc1) return out.clear(), false;
|
||||
const uint32_t hash = member(u32(values + 4 * (size_t)i), k_hash, &type);
|
||||
if (bad || (type != 0xd3 && type != 0xd1)) return out.clear(), false; // (an unsigned number, or a signed one)
|
||||
if (hash) out.push_back(hash);
|
||||
}
|
||||
std::sort(out.begin(), out.end());
|
||||
out.erase(std::unique(out.begin(), out.end()), out.end());
|
||||
return !out.empty();
|
||||
}
|
||||
|
||||
/// An event flow that holds a script: a little-endian BFEVFL whose header counts a flowchart at least (0x20)
|
||||
inline bool event_has_script(const std::vector<unsigned char>& f) {
|
||||
return f.size() >= 0x24 && !memcmp(f.data(), "BFEVFL\0\0", 8) && f[0xc] == 0xff && f[0xd] == 0xfe && (f[0x20] | f[0x21] << 8) >= 1;
|
||||
}
|
||||
|
||||
/// The texts of a file that are flags' names: every run of a name's characters, 3 to OOT_NAME_MAX of them with a 0 after,
|
||||
/// whose hash is one of bools (sorted). Added to out (the same name more than once: make_list sorts that out)
|
||||
inline void flag_names(const std::vector<unsigned char>& f, const std::vector<uint32_t>& bools, std::vector<std::string>& out) {
|
||||
const size_t n = f.size();
|
||||
for (size_t i = 0; i < n;) {
|
||||
if (!oot_name_char(f[i])) {
|
||||
i++;
|
||||
continue;
|
||||
}
|
||||
size_t end = i;
|
||||
while (end < n && oot_name_char(f[end])) end++;
|
||||
const size_t len = end - i;
|
||||
if (end < n && !f[end] && len >= 3 && len <= OOT_NAME_MAX &&
|
||||
std::binary_search(bools.begin(), bools.end(), oot_name_hash((const char*)f.data() + i, len)))
|
||||
out.emplace_back((const char*)f.data() + i, len);
|
||||
i = end;
|
||||
}
|
||||
}
|
||||
|
||||
/// The event an event flow's file is: its name without ".bfevfl.zs" and without a version before that (".110"); "" when
|
||||
/// the file is not an event flow's or its name not an event's
|
||||
inline std::string event_of_file(const std::string& file) {
|
||||
static const char kEnd[] = ".bfevfl.zs";
|
||||
const size_t end = sizeof kEnd - 1;
|
||||
if (file.size() <= end || file.compare(file.size() - end, end, kEnd)) return "";
|
||||
std::string name = file.substr(0, file.size() - end);
|
||||
const size_t dot = name.rfind('.');
|
||||
if (dot != std::string::npos && dot + 1 < name.size() && name.find_first_not_of("0123456789", dot + 1) == std::string::npos) name.resize(dot);
|
||||
return oot_name_ok(name.c_str()) ? name : "";
|
||||
}
|
||||
|
||||
// ---- a page's list: the names sorted, each once, in groups by what they begin with ----
|
||||
struct Group {
|
||||
std::string title;
|
||||
int first = 0, count = 0; // (its names: [first, first + count) of the list)
|
||||
bool begins = false; // (its names all begin with its title: not Other, and not a list without groups)
|
||||
};
|
||||
struct List {
|
||||
std::vector<std::string> names;
|
||||
std::vector<uint32_t> hashes; // (each name's, for the flags)
|
||||
std::vector<Group> groups; // (never empty while there are names)
|
||||
};
|
||||
constexpr int kGroupMin = 6; // (a beginning that fewer names share has no group of its own: they are in "Other")
|
||||
/// What a name begins with: its text before the first _ or . (the whole name when it has neither)
|
||||
inline std::string name_prefix(const std::string& name) {
|
||||
const size_t cut = name.find_first_of("_.");
|
||||
return cut == std::string::npos || cut == 0 ? name : name.substr(0, cut);
|
||||
}
|
||||
/// The list of some names (any order, any name any number of times). grouped false: one group of them all (the amiibo
|
||||
/// folder's files)
|
||||
inline List make_list(std::vector<std::string> names, bool grouped = true, const char* one_title = "All") {
|
||||
List list;
|
||||
std::sort(names.begin(), names.end());
|
||||
names.erase(std::unique(names.begin(), names.end()), names.end());
|
||||
names.erase(std::remove(names.begin(), names.end(), std::string()), names.end());
|
||||
if (names.empty()) return list;
|
||||
std::vector<std::pair<std::string, std::string>> keyed; // (the group's title, the name)
|
||||
keyed.reserve(names.size());
|
||||
if (!grouped) {
|
||||
for (std::string& name : names) keyed.emplace_back(one_title, std::move(name));
|
||||
} else {
|
||||
// (sorted, so the names of one beginning are together: counted in one pass)
|
||||
std::vector<std::string> prefixes;
|
||||
prefixes.reserve(names.size());
|
||||
for (const std::string& name : names) prefixes.push_back(name_prefix(name));
|
||||
std::vector<std::string> counted = prefixes;
|
||||
std::sort(counted.begin(), counted.end());
|
||||
for (size_t i = 0; i < names.size(); i++) {
|
||||
const auto range = std::equal_range(counted.begin(), counted.end(), prefixes[i]);
|
||||
const bool own = range.second - range.first >= kGroupMin;
|
||||
// ("\x7f" sorts Other after every group of a beginning: a name's characters are all below it)
|
||||
keyed.emplace_back(own ? prefixes[i] : std::string("\x7f"), std::move(names[i]));
|
||||
}
|
||||
std::sort(keyed.begin(), keyed.end());
|
||||
}
|
||||
for (size_t i = 0; i < keyed.size(); i++) {
|
||||
if (list.groups.empty() || keyed[i].first != keyed[i - 1].first) {
|
||||
Group g;
|
||||
g.title = keyed[i].first == "\x7f" ? "Other" : keyed[i].first;
|
||||
g.begins = grouped && keyed[i].first != "\x7f";
|
||||
g.first = (int)i;
|
||||
list.groups.push_back(std::move(g));
|
||||
}
|
||||
list.groups.back().count++;
|
||||
list.hashes.push_back(oot_name_hash(keyed[i].second.data(), keyed[i].second.size()));
|
||||
list.names.push_back(std::move(keyed[i].second));
|
||||
}
|
||||
return list;
|
||||
}
|
||||
|
||||
/// A name as its card shows it: without the beginning its group is named for, and the _ or . after that (the group's
|
||||
/// title says it once for all of them, and the room goes to where the names differ). The whole name in Other, in a list
|
||||
/// without groups, and where nothing would be left
|
||||
inline std::string card_label(const List& list, int group, int index) {
|
||||
const std::string& name = list.names[index];
|
||||
const Group& g = list.groups[group];
|
||||
if (!g.begins || name.size() <= g.title.size() + 1) return name;
|
||||
return name.substr(g.title.size() + 1);
|
||||
}
|
||||
/// A text that does not fit, cut in its middle: a third of what fits from its start, "..." and the rest from its end (the
|
||||
/// game's names tell each other apart at their ends). fits(text): whether a text fits; it must hold for a shorter one when
|
||||
/// it holds for a longer. The text itself when it fits, "..." when nothing more does. UTF-8 is cut between characters
|
||||
template <class Fits>
|
||||
std::string elide_middle(const std::string& text, Fits fits) {
|
||||
if (fits(text)) return text;
|
||||
const size_t n = text.size();
|
||||
auto cut = [&](size_t keep) { // (keep bytes of the text at most: a head and a tail on character boundaries)
|
||||
size_t head = keep / 3, tail = n - (keep - head);
|
||||
while (head > 0 && ((unsigned char)text[head] & 0xc0) == 0x80) head--;
|
||||
while (tail < n && ((unsigned char)text[tail] & 0xc0) == 0x80) tail++;
|
||||
return text.substr(0, head) + "..." + text.substr(tail);
|
||||
};
|
||||
size_t lo = 0, hi = n ? n - 1 : 0; // (the most bytes kept that fit: lo fits or is 0, more than hi does not)
|
||||
while (lo < hi) {
|
||||
const size_t mid = (lo + hi + 1) / 2;
|
||||
if (fits(cut(mid))) lo = mid;
|
||||
else hi = mid - 1;
|
||||
}
|
||||
return cut(lo);
|
||||
}
|
||||
|
||||
// ---- a page: a list, the group shown, the name chosen in it, and the cards as last drawn ----
|
||||
struct Page {
|
||||
List list;
|
||||
int group = 0; // (the group shown)
|
||||
int choice = 0; // (the name chosen: its place in list.names, inside the group)
|
||||
int columns = 1, capacity = 1; // (as last drawn: the cards in a row, on a page)
|
||||
int shown = 0; // (the page of cards shown: the choice's, or another the wheel turned to)
|
||||
|
||||
bool empty() const { return list.names.empty(); }
|
||||
int begin() const { return empty() ? 0 : list.groups[group].first; }
|
||||
int end() const { return empty() ? 0 : list.groups[group].first + list.groups[group].count; }
|
||||
int pages() const { return std::max(1, (end() - begin() + std::max(1, capacity) - 1) / std::max(1, capacity)); }
|
||||
/// The cards shown: [first, last) of list.names
|
||||
int first() const { return begin() + std::clamp(shown, 0, pages() - 1) * std::max(1, capacity); }
|
||||
int last() const { return std::min(end(), first() + std::max(1, capacity)); }
|
||||
/// The page of cards the choice is on
|
||||
void reveal() { shown = empty() ? 0 : (choice - begin()) / std::max(1, capacity); }
|
||||
/// A new list (the files were read, or the folder listed again): the choice kept by its name when the new list has it
|
||||
void set_list(List next) {
|
||||
const std::string was = !empty() && choice >= 0 && choice < (int)list.names.size() ? list.names[choice] : std::string();
|
||||
list = std::move(next);
|
||||
group = 0, choice = 0, shown = 0;
|
||||
if (empty()) return;
|
||||
const auto at = std::find(list.names.begin(), list.names.end(), was);
|
||||
if (!was.empty() && at != list.names.end()) {
|
||||
choice = (int)(at - list.names.begin());
|
||||
for (int g = 0; g < (int)list.groups.size(); g++)
|
||||
if (choice >= list.groups[g].first && choice < list.groups[g].first + list.groups[g].count) group = g;
|
||||
}
|
||||
reveal();
|
||||
}
|
||||
/// Another group: dir -1 the one before, 1 the one after, round the ends; its first name chosen
|
||||
void turn_group(int dir) {
|
||||
if (empty()) return;
|
||||
const int n = (int)list.groups.size();
|
||||
group = ((group + dir) % n + n) % n;
|
||||
choice = begin(), shown = 0;
|
||||
}
|
||||
/// The choice moved: by cards (Left and Right), rows (Up and Down), pages, or to the group's first or last name. It
|
||||
/// stays in the group; the page shown follows it
|
||||
void move(int cards, int rows, int by_pages, bool home, bool to_end) {
|
||||
if (empty()) return;
|
||||
const int step = cards + rows * std::max(1, columns) + by_pages * std::max(1, capacity);
|
||||
choice = home ? begin() : to_end ? end() - 1 : std::clamp(choice + step, begin(), end() - 1);
|
||||
reveal();
|
||||
}
|
||||
/// A card clicked: chosen, when it is one of the group's
|
||||
bool choose(int index) {
|
||||
if (empty() || index < begin() || index >= end()) return false;
|
||||
choice = index;
|
||||
return true;
|
||||
}
|
||||
/// The wheel: the page of cards shown, the choice where it is
|
||||
void turn_shown(int by) { shown = std::clamp(shown + by, 0, pages() - 1); }
|
||||
/// The hashes of the names shown now, and of the chosen one when the wheel turned its page away, at most OOT_WATCH_MAX:
|
||||
/// the flags whose states the page asks for
|
||||
std::vector<uint32_t> watched() const {
|
||||
std::vector<uint32_t> hashes;
|
||||
for (int i = first(); i < last() && (int)hashes.size() < OOT_WATCH_MAX; i++) hashes.push_back(list.hashes[i]);
|
||||
if (!empty() && (choice < first() || choice >= last()) && (int)hashes.size() < OOT_WATCH_MAX) hashes.push_back(list.hashes[choice]);
|
||||
return hashes;
|
||||
}
|
||||
};
|
||||
|
||||
/// The flags' states as the game's side last said them, for the cards shown
|
||||
struct States {
|
||||
std::vector<uint32_t> hashes; // (the list asked for)
|
||||
std::string states; // (a character each: 1, 0, - or ?; empty until an answer for this list came)
|
||||
/// The list asked for changed: true when a new W line is to be sent. The states known are forgotten then
|
||||
bool ask(const std::vector<uint32_t>& now) {
|
||||
if (now == hashes) return false;
|
||||
hashes = now, states.clear();
|
||||
return true;
|
||||
}
|
||||
/// A line of states: kept when it answers the list asked for
|
||||
bool take(const char* line) {
|
||||
std::string got(hashes.size(), '?');
|
||||
if (!oot_states_parse(line, oot_watch_sum(hashes.data(), (int)hashes.size()), (int)hashes.size(), got.data())) return false;
|
||||
states = std::move(got);
|
||||
return true;
|
||||
}
|
||||
/// A flag's state: 1, 0, - (the save data has no such flag) or ? (not known: no save loaded, or no answer yet)
|
||||
char of(uint32_t hash) const {
|
||||
for (size_t i = 0; i < hashes.size() && i < states.size(); i++)
|
||||
if (hashes[i] == hash) return states[i];
|
||||
return '?';
|
||||
}
|
||||
};
|
||||
|
||||
/// The lines a page waits to send, a byte a read (HOS_OPT_DEBUG_ASK, HOS_OPT_AMIIBO_ASK): each ends with a line feed,
|
||||
/// 0 while there is none. The game's side reads them only while the game runs, so of the lines that only say what the
|
||||
/// page shows now (those beginning with `latest`: W) the newest alone is kept, and no more than kLinesMax wait
|
||||
struct Lines {
|
||||
static constexpr int kLinesMax = 32;
|
||||
std::vector<std::string> waiting;
|
||||
size_t at = 0; // (how much of waiting[0] was read)
|
||||
void push(const std::string& line, char latest = 0) {
|
||||
for (size_t i = at ? 1 : 0; latest && i < waiting.size();) { // (an older one not yet begun: dropped)
|
||||
if (!waiting[i].empty() && waiting[i][0] == latest) waiting.erase(waiting.begin() + (long)i);
|
||||
else i++;
|
||||
}
|
||||
if ((int)waiting.size() < kLinesMax) waiting.push_back(line);
|
||||
}
|
||||
int take() {
|
||||
if (waiting.empty()) return 0;
|
||||
if (at < waiting[0].size()) return (unsigned char)waiting[0][at++];
|
||||
waiting.erase(waiting.begin()), at = 0;
|
||||
return '\n';
|
||||
}
|
||||
/// The same line as the last one waiting: a key held on a request repeats it, and one of them is enough
|
||||
bool repeats(const std::string& line) const { return !waiting.empty() && waiting.back() == line; }
|
||||
/// The lines waiting are given up but those beginning with `keep` and one already half read (the Debug tools were
|
||||
/// turned off: what was asked of the game and not yet taken is not done after all; a W still says what is shown)
|
||||
void drop_but(char keep) {
|
||||
for (size_t i = at ? 1 : 0; i < waiting.size();) {
|
||||
if (waiting[i].empty() || waiting[i][0] != keep) waiting.erase(waiting.begin() + (long)i);
|
||||
else i++;
|
||||
}
|
||||
}
|
||||
};
|
||||
|
||||
} // namespace mkwvk::debug_names
|
||||
42
gpu/vk/debug_pending.h
Normal file
42
gpu/vk/debug_pending.h
Normal file
@@ -0,0 +1,42 @@
|
||||
// (Triforce) The Inventory page's change asked of the game and not yet taken by it (vk_menu.cpp keeps it in cheat_action;
|
||||
// host/hos_main.c takes it at Link's update, which does not run while the game's own Options screen has the game paused).
|
||||
// The arithmetic is here so that tests/test_debug_pending.py can run it.
|
||||
//
|
||||
// An action is OOT_CHEAT_ITEM_ACTION + item * 256 + op: op 0..127 adds op - 64 to the item's amount, op 200 unlocks it (1)
|
||||
// or relocks it (0). Only one action waits at a time. Before, each press replaced the last: ten presses of Right on an
|
||||
// amount in the paused Options screen made it one more, and the card showed the amount as it was; a second press of Unlock
|
||||
// unlocked it all the same. Now a press adds to the same item's waiting change, and the card shows what it will be.
|
||||
#pragma once
|
||||
|
||||
#include <algorithm>
|
||||
|
||||
#include "oot_cheat_items.h"
|
||||
|
||||
namespace mkwvk::debug_pending {
|
||||
|
||||
/// The change to item's amount the waiting action asks (0: none, or the action is not that item's amount)
|
||||
inline int amount_delta(int pending, int item) {
|
||||
const int base = OOT_CHEAT_ITEM_ACTION + item * 256;
|
||||
return pending >= base && pending < base + 128 ? pending - base - 64 : 0;
|
||||
}
|
||||
/// Whether the waiting action is item's unlock or relock
|
||||
inline bool toggles(int pending, int item) { return pending == OOT_CHEAT_ITEM_ACTION + item * 256 + 200; }
|
||||
|
||||
/// The action after a step (dir 1 or -1) on item's amount, count what the game last said it has (0 or more): the item's
|
||||
/// waiting change and the step together, kept to what the game allows (0 to 999) and to what one action says (-64 to 63).
|
||||
/// Another item's waiting action is replaced, as before
|
||||
inline int amount_step(int pending, int item, int count, int dir) {
|
||||
const int delta = std::clamp(amount_delta(pending, item) + dir, std::max(-64, -count), std::min(63, 999 - count));
|
||||
return OOT_CHEAT_ITEM_ACTION + item * 256 + 64 + delta;
|
||||
}
|
||||
/// The action after a press of Unlock or Relock on item: a second press before the game took the first takes it back (0)
|
||||
inline int toggle(int pending, int item) { return toggles(pending, item) ? 0 : OOT_CHEAT_ITEM_ACTION + item * 256 + 200; }
|
||||
|
||||
/// What item's card shows: count, what the game last said, with the waiting change made (count < 0, unknown: as it is)
|
||||
inline int shown(int pending, int item, int count) {
|
||||
if (count < 0) return count;
|
||||
if (toggles(pending, item)) return count > 0 ? 0 : 1;
|
||||
return std::clamp(count + amount_delta(pending, item), 0, 999);
|
||||
}
|
||||
|
||||
} // namespace mkwvk::debug_pending
|
||||
36
gpu/vk/demote_watch.h
Normal file
36
gpu/vk/demote_watch.h
Normal file
@@ -0,0 +1,36 @@
|
||||
/* (Triforce) When a pool's trial in guest memory ends early (vk_memory.cpp demote_trials). A pool the CPU and the GPU both write
|
||||
is tried out of the mirrors: its pieces (or the whole pool) are bound in guest memory, 10 frames to settle, then up to 90
|
||||
measured, and the GPU's time a frame is compared with the window before. Only the measured frames were judged, and only after
|
||||
ten of them: in the owner's log (RTX 5090, 2026-11-15, test4) a 16 MB pool took the GPU 106 ms a frame in guest memory
|
||||
against 7.4 before, and the game's thread waited for that GPU in every capture (memory sync 99 ms a submit, 18 ms a frame
|
||||
over 120 presents): some twenty frames of 100 ms, the same again at its second trial, and a 32 MB pool's two trials at 39 ms
|
||||
against 7.7. A fresh install tries every such pool (a pool that failed twice is kept in triforce_mirror_keep.txt for later
|
||||
runs), which is what a benchmark plays. So a trial far over its baseline ends as soon as a few frames say so, settling or
|
||||
measured, the way a slower one ends: mirrored again and measured once more that way, and judged against that (phase 4), so
|
||||
an early end that was wrong costs a second measure, not the pool. Of the 84 trials in the owner's five logs the four costly
|
||||
ones were 4.7 to 19 times their baseline; every other was under 2.3 times but one, 3.6 times, whose mirrored-again frames were
|
||||
as slow (its baseline was off; it stayed in guest memory): under this rule's line too. RX_VK_DEMOTE_QUICK=0: as before. Plain
|
||||
functions: tests/test_demote_watch.cpp checks them. */
|
||||
#pragma once
|
||||
|
||||
#include <cstdint>
|
||||
#include <cstdlib>
|
||||
|
||||
namespace mkwvk::demote_watch {
|
||||
|
||||
constexpr uint64_t kFrames = 3; // (the frames since the trial's start, or its measure's, before one is judged: the GPU's time comes
|
||||
// in as submissions finish, a frame or two late, so the first frames weigh less than they cost)
|
||||
constexpr double kTimes = 3.0; // (over kTimes the baseline and kOverMs more: see above. The trial's own rule, twice the
|
||||
constexpr double kOverMs = 5.0; // baseline and 1 ms after 10 measured frames, stays as it was)
|
||||
|
||||
/// Whether a trial whose frames so far took ms_a_frame of GPU time each, over frames of them, ends now: far slower than base_ms
|
||||
inline bool gross(double base_ms, double ms_a_frame, uint64_t frames) {
|
||||
return base_ms > 0 && frames >= kFrames && ms_a_frame > base_ms * kTimes + kOverMs;
|
||||
}
|
||||
|
||||
/// RX_VK_DEMOTE_QUICK: on unless "0"
|
||||
inline bool quick_on(const char* env) {
|
||||
return !(env && *env && std::atoi(env) == 0 && env[0] == '0');
|
||||
}
|
||||
|
||||
} // namespace mkwvk::demote_watch
|
||||
67
gpu/vk/desc_counts.h
Normal file
67
gpu/vk/desc_counts.h
Normal file
@@ -0,0 +1,67 @@
|
||||
// A program's descriptor counts kept with the program (no game, no GPU: a rule; vk_shader.h Program::desc_n, vk_draw.cpp
|
||||
// build_descriptors).
|
||||
//
|
||||
// (Triforce, 11-22, the performance audit's DESC-22, after Source A's.) build_descriptors sums the program's buffers,
|
||||
// images and texel views at every build, six lists a stage, only to size three reserve() calls that keep the pointers
|
||||
// it hands the writes stable. The sums are fixed once the program is translated: its stages are filled before it is
|
||||
// ready and never after. So the three are packed into one word kept on the program, stored at its first build once
|
||||
// ready and read at every later one; a word that is not a count walks as before.
|
||||
//
|
||||
// The word is 32 bits because that is what the program has room for: the padding after Program::ready. The struct must
|
||||
// not grow: sizeof(Program) is in the program cache's stamp, and a change of it throws every player's caches away. So
|
||||
// three counts of 10 bits; a program with more than 1023 of a kind is not kept (kUnset is what its pack gives) and walks
|
||||
// at every build, as before.
|
||||
//
|
||||
// A short count would let a vector move in the middle of a build and leave the writes with stale pointers. It cannot
|
||||
// come from the walk that stores it, but the build checks all the same that none of the three moved, and a program
|
||||
// whose did is marked kNever, walked from then on, and built again at once.
|
||||
//
|
||||
// RX_VK_DESC_COUNTS=0: the walk at every build, as before. The lab's knob "descwalk" does the same in its phases.
|
||||
#pragma once
|
||||
#include <cstddef>
|
||||
#include <cstdint>
|
||||
#include <cstdlib>
|
||||
|
||||
namespace mkwvk::desc_counts {
|
||||
|
||||
inline constexpr uint32_t kUnset = ~0u; // not counted yet, or too many of a kind to pack
|
||||
inline constexpr uint32_t kNever = 0xFFFFFFFEu; // its kept counts were short once: walked at every build from then on
|
||||
inline constexpr uint32_t kMost = 1023; // the most of a kind a word holds (10 bits)
|
||||
|
||||
/// RX_VK_DESC_COUNTS: on unless set to 0
|
||||
inline bool on(const char* env) {
|
||||
return !env || std::atoi(env) != 0;
|
||||
}
|
||||
|
||||
/// The three counts as one word, or kUnset where one does not fit
|
||||
constexpr uint32_t pack(size_t bufs, size_t imgs, size_t views) {
|
||||
if (bufs > kMost || imgs > kMost || views > kMost) return kUnset;
|
||||
return (uint32_t)bufs | (uint32_t)imgs << 10 | (uint32_t)views << 20;
|
||||
}
|
||||
|
||||
/// Whether the word holds counts (its two top bits are clear: kUnset and kNever have them set)
|
||||
constexpr bool holds(uint32_t word) {
|
||||
return (word >> 30) == 0;
|
||||
}
|
||||
|
||||
/// The counts of a word that holds them; false, and nothing written, for one that does not (walk)
|
||||
constexpr bool unpack(uint32_t word, size_t* bufs, size_t* imgs, size_t* views) {
|
||||
if (!holds(word)) return false;
|
||||
*bufs = word & kMost, *imgs = (word >> 10) & kMost, *views = (word >> 20) & kMost;
|
||||
return true;
|
||||
}
|
||||
|
||||
/// Whether a build that walked stores its counts: the rule on, the program translated (its stages do not change after),
|
||||
/// and the word still unset (never over kNever)
|
||||
constexpr bool stores(bool on, bool ready, uint32_t word) {
|
||||
return on && ready && word == kUnset;
|
||||
}
|
||||
|
||||
/// The guard at a build's end: whether one of the three vectors' storage moved since the reserve (before/after: their
|
||||
/// data() pointers), in a build that went by kept counts. A build that walked reserved exactly, as it always did
|
||||
constexpr bool moved(bool walked, const void* bufs0, const void* bufs1, const void* imgs0, const void* imgs1, const void* views0,
|
||||
const void* views1) {
|
||||
return !walked && (bufs0 != bufs1 || imgs0 != imgs1 || views0 != views1);
|
||||
}
|
||||
|
||||
} // namespace mkwvk::desc_counts
|
||||
113
gpu/vk/desc_keep.h
Normal file
113
gpu/vk/desc_keep.h
Normal file
@@ -0,0 +1,113 @@
|
||||
// (Triforce) Per-binding descriptor reuse (vk_draw.cpp build_descriptors), apart from Vulkan so that tests/test_desc_keep.py can run
|
||||
// it. NX2navi's per-binding reuse (runtime/vk_exec_draw.c pb_take and pb_keep, chrg6's) is the model; the rules below are this
|
||||
// renderer's own, from what each of its descriptors is resolved from.
|
||||
//
|
||||
// A draw whose program, bindings, texture slots and uniform updates are all as the draw before it keeps that draw's descriptors whole
|
||||
// (vk_draw.cpp same_desc). In the owner's test7 session that was 0 to 349 draws a frame of 2,300 to 2,870, and "descriptors" was 1.5
|
||||
// to 2.2 ms of a replay of 7.4 to 9.8 ms: a draw that binds one uniform buffer of its own, or anything bound anywhere, built every
|
||||
// descriptor again, each texture through the views' map and the sampler table, each buffer through its pool. Now each descriptor of
|
||||
// a graphics program keeps its last result beside what it was resolved from, for the last kSlots programs a recording thread drew,
|
||||
// and a descriptor whose inputs and stamps are all the same takes that result instead of resolving again. What a result stands on,
|
||||
// by its kind (stamps_of: the stamps are read before the descriptor is resolved, so that one moved by the resolution itself, or by
|
||||
// another thread meanwhile, makes the next one resolve again):
|
||||
// - a sampled texture (TEX): its handle, view type and depth compare as inputs; the slot's version (the slot registered anew), the
|
||||
// samplers' generation (a sampler registration's contents changed), the destroy generation (a texture destroyed or made again:
|
||||
// a kept view would name freed memory, issue 1126), the dummy generation (a frame left a texture for later: a dummy is looked
|
||||
// up again) and the texture epoch (a view's texture is touched once a submission: texture_view_of does it on its first look);
|
||||
// - a user uniform buffer (UBO, c3 on): its address and size; the recording, the stream's inline uniform updates (ub_generation:
|
||||
// an update into the range is a copy with it in) and the bind stamp (gpu_bind_stamp: a pool gone or imported, a mirror made, the
|
||||
// next submission). Only a pool's or a mirror's range is kept: a copy into the recording's stream (an overlay, an unaligned range,
|
||||
// a range read where it was bound) is made again each time, as it is today;
|
||||
// - a stage's own constants (C1): the recording (they are uploaded once a recording, RecordingRes::c1_at);
|
||||
// - a storage buffer the program only reads (SSBO, a single one, not a table): its address and size; the recording and the bind
|
||||
// stamp. A pool's or a mirror's range only, as for UBO.
|
||||
// The rest is resolved every time as before: the driver's banks (c0, c2: kept by their stages' generations already), storage buffers
|
||||
// the program writes (each binding stamps its pages for the write-back), storage buffer tables, texel buffers and storage images.
|
||||
// A failed resolution (no sampler, no buffer) is never kept: the next draw tries again.
|
||||
#pragma once
|
||||
|
||||
#include <array>
|
||||
#include <cstddef>
|
||||
#include <cstdint>
|
||||
#include <vector>
|
||||
|
||||
namespace desc_keep {
|
||||
|
||||
enum Kind : uint8_t { kNone = 0, kTex, kUbo, kC1, kSsbo };
|
||||
|
||||
/// The stamps results stand on, read where a descriptor is about to be resolved (only those of its kind are read: stamps_of)
|
||||
struct Stamps {
|
||||
uint64_t rec = 0; // the recording (vk_draw.cpp desc_rec: moved at each recording's begin)
|
||||
uint64_t bind = 0; // gpu_bind_stamp
|
||||
uint64_t ub = 0; // ub_generation
|
||||
uint64_t samp = 0; // sampler_registrations_generation
|
||||
uint64_t destroy = 0; // texture_destroy_generation
|
||||
uint64_t dummy = 0; // texture_dummy_generation
|
||||
uint64_t epoch = 0; // texture_epoch
|
||||
};
|
||||
|
||||
/// The stamps a result of that kind stands on, in a fixed order (unused ones 0)
|
||||
inline std::array<uint64_t, 4> stamps_of(Kind k, const Stamps& s) {
|
||||
switch (k) {
|
||||
case kTex: return {s.samp, s.destroy, s.dummy, s.epoch};
|
||||
case kUbo: return {s.rec, s.bind, s.ub, 0};
|
||||
case kC1: return {s.rec, 0, 0, 0};
|
||||
case kSsbo: return {s.rec, s.bind, 0, 0};
|
||||
default: return {~0ull, ~0ull, ~0ull, ~0ull};
|
||||
}
|
||||
}
|
||||
|
||||
/// One descriptor's last resolution: what it was resolved from, and its result (a buffer: handle, offset, range; a texture:
|
||||
/// sampler, view as texture_view_of gave it (null: the dummy is drawn), the Texture). flags: a TEX's slot was a registered one
|
||||
struct Elem {
|
||||
uint8_t kind = kNone;
|
||||
uint8_t flags = 0;
|
||||
uint32_t version = 0; // (TEX: the slot's version)
|
||||
uint64_t in[2]{};
|
||||
std::array<uint64_t, 4> at{};
|
||||
uint64_t out[3]{};
|
||||
};
|
||||
|
||||
/// Whether e's result serves a descriptor of kind k with those inputs (version: a TEX's slot version, 0 for the rest), now
|
||||
inline bool take(const Elem& e, Kind k, uint64_t in0, uint64_t in1, uint32_t version, const Stamps& now) {
|
||||
return k != kNone && e.kind == k && e.in[0] == in0 && e.in[1] == in1 && e.version == version && e.at == stamps_of(k, now);
|
||||
}
|
||||
|
||||
/// A descriptor resolved: kept with its inputs and the stamps read before it was resolved (before)
|
||||
inline void keep(Elem& e, Kind k, uint64_t in0, uint64_t in1, uint32_t version, const Stamps& before, uint64_t out0,
|
||||
uint64_t out1, uint64_t out2, uint8_t flags = 0) {
|
||||
e.kind = k, e.flags = flags, e.version = version;
|
||||
e.in[0] = in0, e.in[1] = in1;
|
||||
e.at = stamps_of(k, before);
|
||||
e.out[0] = out0, e.out[1] = out1, e.out[2] = out2;
|
||||
}
|
||||
|
||||
/// Not kept (a failed resolution, or a result that is not to be taken again)
|
||||
inline void drop(Elem& e) { e.kind = kNone; }
|
||||
|
||||
/// (RX_VK_DESC_KEEP=check) Whether a descriptor resolved again came out as its kept result
|
||||
inline bool same(const Elem& e, uint64_t out0, uint64_t out1, uint64_t out2, uint8_t flags = 0) {
|
||||
return e.out[0] == out0 && e.out[1] == out1 && e.out[2] == out2 && e.flags == flags;
|
||||
}
|
||||
|
||||
/// A recording thread's kept results: the last programs drawn, one a slot by the program's address
|
||||
inline constexpr size_t kSlots = 16;
|
||||
struct Prog {
|
||||
const void* prog = nullptr;
|
||||
std::vector<Elem> e; // (by descriptor, in build_descriptors' order)
|
||||
};
|
||||
inline size_t slot_of(const void* prog) {
|
||||
const uintptr_t p = (uintptr_t)prog;
|
||||
return (size_t)((p >> 6) ^ (p >> 13)) % kSlots;
|
||||
}
|
||||
/// prog's kept results, n descriptors: emptied when its slot held another program's, or another count of them
|
||||
inline Prog& prog_of(std::array<Prog, kSlots>& t, const void* prog, size_t n) {
|
||||
Prog& p = t[slot_of(prog)];
|
||||
if (p.prog != prog || p.e.size() != n) {
|
||||
p.prog = prog;
|
||||
p.e.assign(n, Elem{});
|
||||
}
|
||||
return p;
|
||||
}
|
||||
|
||||
} // namespace desc_keep
|
||||
97
gpu/vk/desc_pool_grow.h
Normal file
97
gpu/vk/desc_pool_grow.h
Normal file
@@ -0,0 +1,97 @@
|
||||
// A recording slot's descriptor pool, made at half size and made whole once a recording runs out of it (no game, no GPU:
|
||||
// a rule; vk_draw.cpp make_pool, slot_reset_locked, draw_begin_recording).
|
||||
//
|
||||
// (Triforce, 11-22, the performance audit's DESC-33, after Source A's.) Each of the recording slots (48 at most) keeps a
|
||||
// descriptor pool of 8192 sets and up to 32768 descriptors a type, and the driver may keep host memory for all of that,
|
||||
// 48 times. Few recordings come near it: Source A saw one take at most 289 sets on a desktop card, and no pool at half
|
||||
// size run out in three scenes on a handheld. So a slot's pool starts at half (4096 sets, 16384 uniform buffers and
|
||||
// combined image samplers, and so on down the types).
|
||||
//
|
||||
// A recording that runs out is given a whole extra pool, as it always was (bind_descriptors), and that is the mark: at
|
||||
// the slot's next reset, where the extra pools are destroyed anyway, a slot that has any and whose own pool is not whole
|
||||
// has it destroyed and made again whole. From then on the slot is as it was before this rule. A busy slot so pays one
|
||||
// extra pool, once.
|
||||
//
|
||||
// RX_VK_DESC_POOL_GROW=0: whole pools from the start, as before. RX_VK_DESC_POOL_DIV=<n> still divides both sizes.
|
||||
// What the recordings took (sets and descriptors a type, the most one recording took of each) and how many pools were
|
||||
// added and slots made whole go into the log's "recording streams" line, so a size can be judged from a run's log.
|
||||
#pragma once
|
||||
#include <cstddef>
|
||||
#include <cstdint>
|
||||
#include <cstdlib>
|
||||
|
||||
namespace mkwvk::desc_pool_grow {
|
||||
|
||||
/// RX_VK_DESC_POOL_GROW: on unless set to 0
|
||||
inline bool on(const char* env) {
|
||||
return !env || std::atoi(env) != 0;
|
||||
}
|
||||
|
||||
/// A whole pool's sets, and its descriptors by type in the order of Took::n
|
||||
inline constexpr uint32_t kSets = 8192;
|
||||
enum Kind { kUniform, kStorage, kSampled, kStorageImage, kUniformTexel, kStorageTexel, kKinds };
|
||||
inline constexpr uint32_t kWhole[kKinds] = {32768, 16384, 32768, 4096, 2048, 2048};
|
||||
|
||||
/// The shift a new slot's pool is made with: half with the rule on, whole without it
|
||||
constexpr uint32_t first_shift(bool on) {
|
||||
return on ? 1u : 0u;
|
||||
}
|
||||
|
||||
/// A count of a pool made at `shift` (0 whole, 1 half) and divided by RX_VK_DESC_POOL_DIV's n (1..32): never 0
|
||||
constexpr uint32_t count(uint32_t whole, uint32_t shift, uint32_t div) {
|
||||
const uint32_t n = (whole >> (shift > 8 ? 8 : shift)) / (div ? div : 1);
|
||||
return n ? n : 1;
|
||||
}
|
||||
|
||||
/// Whether a slot's own pool is made again whole at its reset: its last recording ran out (it has extra pools) and the
|
||||
/// pool is not whole yet
|
||||
constexpr bool remake(uint32_t shift, size_t extra_pools) {
|
||||
return shift != 0 && extra_pools != 0;
|
||||
}
|
||||
|
||||
/// Which of the six a Vulkan descriptor type is (VkDescriptorType's numbers, so that this header needs no Vulkan):
|
||||
/// kKinds for any other
|
||||
constexpr int kind(uint32_t vk_type) {
|
||||
switch (vk_type) {
|
||||
case 6: return kUniform; // VK_DESCRIPTOR_TYPE_UNIFORM_BUFFER
|
||||
case 7: return kStorage; // VK_DESCRIPTOR_TYPE_STORAGE_BUFFER
|
||||
case 1: return kSampled; // VK_DESCRIPTOR_TYPE_COMBINED_IMAGE_SAMPLER
|
||||
case 3: return kStorageImage; // VK_DESCRIPTOR_TYPE_STORAGE_IMAGE
|
||||
case 4: return kUniformTexel; // VK_DESCRIPTOR_TYPE_UNIFORM_TEXEL_BUFFER
|
||||
case 5: return kStorageTexel; // VK_DESCRIPTOR_TYPE_STORAGE_TEXEL_BUFFER
|
||||
default: return kKinds;
|
||||
}
|
||||
}
|
||||
|
||||
/// What one recording took of its slot's pools: sets, and descriptors by type
|
||||
struct Took {
|
||||
uint32_t sets = 0;
|
||||
uint32_t n[kKinds] = {};
|
||||
};
|
||||
|
||||
/// One more set allocated, with these writes (W: VkWriteDescriptorSet, or a test's own with the same two fields)
|
||||
template <class W>
|
||||
inline void took(Took& t, const W* writes, size_t count_of) {
|
||||
t.sets++;
|
||||
for (size_t i = 0; i < count_of; i++) {
|
||||
const int k = kind((uint32_t)writes[i].descriptorType);
|
||||
if (k < kKinds) t.n[k] += (uint32_t)writes[i].descriptorCount;
|
||||
}
|
||||
}
|
||||
|
||||
/// The most any recording took of each, raised by one recording's
|
||||
inline void peak(Took& most, const Took& t) {
|
||||
if (t.sets > most.sets) most.sets = t.sets;
|
||||
for (int k = 0; k < kKinds; k++)
|
||||
if (t.n[k] > most.n[k]) most.n[k] = t.n[k];
|
||||
}
|
||||
|
||||
/// Whether what a recording took fits a pool made at `shift` and divided by div (the tests', and the log's reader's)
|
||||
constexpr bool fits(const Took& t, uint32_t shift, uint32_t div) {
|
||||
if (t.sets > count(kSets, shift, div)) return false;
|
||||
for (int k = 0; k < kKinds; k++)
|
||||
if (t.n[k] > count(kWhole[k], shift, div)) return false;
|
||||
return true;
|
||||
}
|
||||
|
||||
} // namespace mkwvk::desc_pool_grow
|
||||
95
gpu/vk/desc_sig.h
Normal file
95
gpu/vk/desc_sig.h
Normal file
@@ -0,0 +1,95 @@
|
||||
// (Triforce) The whole-draw skip by the signature of a draw's descriptor inputs (vk_draw.cpp desc_sig_walk and draw), apart from
|
||||
// Vulkan so that tests/test_desc_sig.py can run it. NX2navi's bind_sig (runtime/vk_exec_draw.c) is the model; its code is not
|
||||
// chrg6's, so this is written anew from what it does, on this renderer's own rules (desc_keep.h).
|
||||
//
|
||||
// A draw keeps the descriptors the draw before it pushed only while nothing at all was bound anew, no inline uniform update came and
|
||||
// no texture slot changed (vk_draw.cpp same_desc): 381 of 2,830 draws a frame in the owner's test18. The others build every
|
||||
// descriptor again, even when what they read is what the draw before them read: a uniform buffer bound for another stage, a texture
|
||||
// bound at a slot the program does not sample. Through the per-binding reuse (desc_keep.h) each of their descriptors then takes the
|
||||
// result it had, and the push finds nothing changed: the build and the compare were the work, 2.0 to 3.3 ms a frame.
|
||||
//
|
||||
// Now such a draw first reads its signature: what each of its descriptors would be resolved from (desc_keep.h's inputs: a texture's
|
||||
// handle and its slot's version, a user uniform buffer's address and size, a read-only storage buffer's address and size words, a
|
||||
// driver bank's stage generation), and the stamps desc_keep.h's rules give each kind it has (stamps_of), read before the inputs.
|
||||
// When it equals the signature of the descriptors at the graphics bind point, which are this program's, the build and the push are
|
||||
// left out: a build would take every one of them from what was kept (each input and stamp is the same) and push nothing.
|
||||
// That holds only where every descriptor of the last build was kept, so the signature stands for
|
||||
// - push programs only (a set's bind point state is not followed), not the mesh fallback, not RenoDX's;
|
||||
// - programs whose every descriptor is of a kind the reuse keeps (shape_of): sampled textures, user uniform buffers (c3 to c18),
|
||||
// a stage's constants (c1), single read-only storage buffers, and the driver's banks (c0, c2), which their own cache keeps by
|
||||
// the stage's generations. Not storage buffers a program writes (each build stamps their pages for the write-back), tables,
|
||||
// texel buffers or storage images;
|
||||
// - builds that dropped nothing (armed): a failed resolution, a copy in the recording's stream (an inline update's overlay, an
|
||||
// unaligned range) or an unbound buffer make the next draw build again.
|
||||
// The textures a hit samples that the GPU wrote (pass hazards: gfx_sampled) are read again from the last build's textures, as a
|
||||
// build would. A signature is the bind point's only once that draw's descriptors were bound: a draw left out between its build and
|
||||
// its bind leaves the last one in place.
|
||||
// RX_VK_DESC_SIG=0: every such draw builds, as before. =check: a hit builds too, and one whose descriptors or sampled textures come
|
||||
// out otherwise than what is pushed is counted and logged ("vk: descriptor signature check").
|
||||
#pragma once
|
||||
|
||||
#include <array>
|
||||
#include <cstddef>
|
||||
#include <cstdint>
|
||||
#include <vector>
|
||||
|
||||
#include "desc_keep.h"
|
||||
|
||||
namespace desc_sig {
|
||||
|
||||
/// The kinds of descriptor a program has, one bit each by desc_keep::Kind (bit kTex, kUbo, kC1, kSsbo); kBank its driver banks
|
||||
enum : uint8_t { kBank = 1u << 7 };
|
||||
constexpr uint8_t bit(desc_keep::Kind k) { return (uint8_t)(1u << k); }
|
||||
|
||||
/// What a program's descriptors are, from its stages' lists (made once a program: they do not change once it is translated)
|
||||
struct Shape {
|
||||
const void* prog = nullptr;
|
||||
bool ok = false; // (every descriptor of a kind the signature stands for)
|
||||
uint8_t kinds = 0;
|
||||
};
|
||||
inline constexpr size_t kShapes = 64;
|
||||
inline size_t shape_slot(const void* prog) {
|
||||
const uintptr_t p = (uintptr_t)prog;
|
||||
return (size_t)((p >> 6) ^ (p >> 14)) % kShapes;
|
||||
}
|
||||
|
||||
/// The signature's head: the program, the recording (c1 and the banks live in its streams), and the stamps of each kind it has, in
|
||||
/// desc_keep.h's order (the same stamps a kept result of that kind stands on)
|
||||
inline void put_head(std::vector<uint64_t>& out, const void* prog, uint8_t kinds, const desc_keep::Stamps& now) {
|
||||
out.push_back((uint64_t)(uintptr_t)prog);
|
||||
out.push_back(now.rec);
|
||||
out.push_back(kinds);
|
||||
for (desc_keep::Kind k : {desc_keep::kTex, desc_keep::kUbo, desc_keep::kC1, desc_keep::kSsbo})
|
||||
if (kinds & bit(k))
|
||||
for (uint64_t v : desc_keep::stamps_of(k, now)) out.push_back(v);
|
||||
}
|
||||
|
||||
/// A signature, and what its build left besides the descriptors: the textures it sampled through a view (draw_sampled keeps those
|
||||
/// the GPU wrote) and the dummy counts it added (the frame statistics')
|
||||
struct Sig {
|
||||
const void* prog = nullptr;
|
||||
bool armed = false; // (every descriptor of that build was kept: a draw with the same signature may take them all)
|
||||
std::vector<uint64_t> w;
|
||||
std::vector<const void*> tex;
|
||||
int64_t dummies = 0, dummies_known = 0;
|
||||
};
|
||||
|
||||
/// Whether a draw of prog whose signature is now may keep the descriptors bound, which bound's draw built
|
||||
inline bool hit(const Sig& bound, const Sig& now, const void* prog) {
|
||||
return bound.armed && bound.prog == prog && now.prog == prog && !now.w.empty() && bound.w == now.w;
|
||||
}
|
||||
/// now's draw built its descriptors and they were bound: its signature is the bind point's (armed: none of them dropped)
|
||||
inline void commit(Sig& bound, Sig& now, bool armed) {
|
||||
std::swap(bound.w, now.w);
|
||||
std::swap(bound.tex, now.tex);
|
||||
bound.prog = now.prog, bound.armed = armed;
|
||||
bound.dummies = now.dummies, bound.dummies_known = now.dummies_known;
|
||||
now.prog = nullptr, now.w.clear();
|
||||
}
|
||||
/// Descriptors bound that no signature stands for (a program the skip does not take, a build that dropped one)
|
||||
inline void forget(Sig& bound) {
|
||||
bound.prog = nullptr, bound.armed = false;
|
||||
bound.w.clear();
|
||||
}
|
||||
|
||||
} // namespace desc_sig
|
||||
@@ -4,6 +4,7 @@
|
||||
#pragma once
|
||||
|
||||
#include <cstdint>
|
||||
#include <cstdlib>
|
||||
#include <string>
|
||||
#include <string_view>
|
||||
|
||||
@@ -16,6 +17,50 @@ inline bool scene_size_fits(uint32_t w, uint32_t h, uint32_t ow, uint32_t oh) {
|
||||
return w && h && ow && oh && w <= ow && h <= oh && (uint64_t)w * 3 >= ow && (uint64_t)h * 3 >= oh;
|
||||
}
|
||||
|
||||
/// ((Triforce, 11-22, the performance audit's PRESENT-38, after Source A's) One quality mode's range of render sizes for
|
||||
/// an output, as NGX answers it (ok: it answered one).
|
||||
struct ModeRange {
|
||||
uint32_t min_w = 0, min_h = 0, max_w = 0, max_h = 0;
|
||||
bool ok = false;
|
||||
};
|
||||
/// How far outside a range a render size may be and still be that mode's: a 64th of the range's largest size on that
|
||||
/// axis, 2 pixels at least. The Render resolution row's shares are twelfths of the output rounded up, NGX's own sizes
|
||||
/// are rounded down: 854x480 against NGX's 853x480 for Ultra Performance at 2560x1440.
|
||||
constexpr uint32_t mode_slack(uint32_t largest) { return largest / 64 > 2 ? largest / 64 : 2; }
|
||||
/// The mode (an index into ranges) that takes a render size of w x h for an output of ow x oh, or -1. The first range
|
||||
/// that holds the size; else, with slack on, the first it is outside of by no more than mode_slack on each axis, and
|
||||
/// only for a size DLSS scales at all (scene_size_fits: the output 1 to 3 times it). *widened: taken by the slack.
|
||||
inline int mode_for(const ModeRange* ranges, size_t n, uint32_t w, uint32_t h, uint32_t ow, uint32_t oh, bool slack, bool* widened) {
|
||||
*widened = false;
|
||||
for (size_t i = 0; i < n; i++) {
|
||||
const ModeRange& r = ranges[i];
|
||||
if (r.ok && w >= r.min_w && h >= r.min_h && w <= r.max_w && h <= r.max_h) return (int)i;
|
||||
}
|
||||
if (!slack || !scene_size_fits(w, h, ow, oh)) return -1;
|
||||
for (size_t i = 0; i < n; i++) {
|
||||
const ModeRange& r = ranges[i];
|
||||
if (!r.ok) continue;
|
||||
const uint32_t sw = mode_slack(r.max_w), sh = mode_slack(r.max_h);
|
||||
if (w + sw >= r.min_w && h + sh >= r.min_h && w <= r.max_w + sw && h <= r.max_h + sh) return *widened = true, (int)i;
|
||||
}
|
||||
return -1;
|
||||
}
|
||||
|
||||
/// ((Triforce, 11-22, the performance audit's PRESENT-32, after Source A's) Whether a feature is made in a one-size mode
|
||||
/// (NGX's DLAA mode: vk_dlss.cpp prepare's `same` list): its render size is the output's. Before, whenever the two sizes
|
||||
/// were equal. NVIDIA DLSS with the Render resolution row at Auto starts at the output's size, so it got a one-size
|
||||
/// feature, and the game's dynamic resolution left that size some frames later: a second feature made in play (Source
|
||||
/// A's save load-in: 119 ms, then 29.5 ms 17 frames on). With the rule (rule: RX_VK_DLSS_PREWARM=1), one-size only where
|
||||
/// NVIDIA DLAA is the Upscaler's choice (dlaa); DLSS at the output's size takes the scaled mode whose range holds it
|
||||
/// (mode_for: Quality's reaches the output's size), and Auto's smaller sizes then draw with that same feature. Off by
|
||||
/// default: DLSS at full size then draws with the scaled mode's network, not DLAA's, a picture nobody has compared yet
|
||||
inline bool dlaa_feature(bool rule, bool dlaa, uint32_t rw, uint32_t rh, uint32_t ow, uint32_t oh) {
|
||||
const bool full = rw == ow && rh == oh;
|
||||
return rule ? full && dlaa : full;
|
||||
}
|
||||
/// RX_VK_DLSS_PREWARM (value, null when unset): the rule above is on only when it is 1
|
||||
inline bool prewarm_rule(const char* text) { return text && *text && std::atoi(text) == 1; }
|
||||
|
||||
/// Whether s is all ASCII: NGX's Linux library takes only such paths (vk_dlss.cpp wide())
|
||||
inline bool ascii(std::string_view s) {
|
||||
for (const char c : s)
|
||||
@@ -84,4 +129,28 @@ inline std::wstring utf8_to_wide(std::string_view s, bool* bad = nullptr) {
|
||||
return out;
|
||||
}
|
||||
|
||||
/// ((Triforce, 11-21, the performance audit's PRESENT-37, after Source A's) The preset DLSS draws with while the DLSS model
|
||||
/// row is stepped through (vk_dlss.cpp prepare). The row goes K, M, L, NVIDIA's default and applies live, so K to L passes
|
||||
/// M: a feature was made for each preset passed, tens of ms of CPU each, and NGX keeps a released feature's video memory for
|
||||
/// some seconds: Source A's 16 GB card at 3440x1440 held 1419 MB of DLSS's for K, M and L at once against 813 MB for L
|
||||
/// alone, and where memory was short NGX refused the next feature. A new preset is taken once the row has held it for
|
||||
/// `frames`; until then the feature already made draws on with its own. With no feature (the first, or one let go) the
|
||||
/// row's preset is taken at once
|
||||
struct PresetSettle {
|
||||
unsigned wanted = ~0u; // (the row's preset, as last seen)
|
||||
uint64_t since = 0; // (the frame it was first seen in)
|
||||
/// The preset for frame `now`: want, the row's; have, the feature's; has: there is a feature
|
||||
unsigned pick(unsigned want, unsigned have, bool has, uint64_t now, uint64_t frames) {
|
||||
if (want != wanted) wanted = want, since = now;
|
||||
if (!has || want == have) return want;
|
||||
return now - since >= frames ? want : have;
|
||||
}
|
||||
};
|
||||
/// RX_VK_DLSS_MODEL_SETTLE's frames: 30 when not set; 0 (and what is no number, or less): at once, as before
|
||||
inline uint64_t settle_frames(const char* text) {
|
||||
if (!text || !*text) return 30;
|
||||
const long v = std::atol(text);
|
||||
return v > 0 ? (uint64_t)v : 0;
|
||||
}
|
||||
|
||||
} // namespace mkwvk::dlss_checks
|
||||
|
||||
58
gpu/vk/dlss_feature.h
Normal file
58
gpu/vk/dlss_feature.h
Normal file
@@ -0,0 +1,58 @@
|
||||
// (Triforce, 11-21, the upscaler audit's UPS-01) The render size NVIDIA DLSS's feature is made for.
|
||||
//
|
||||
// vk_dlss.cpp made every feature for the largest render size its quality mode allows, so that the game's dynamic
|
||||
// resolution could change the render size under it without the feature being made again. NGX answers that largest size as
|
||||
// the output's own (the owner's log of 11-21: "feature 1 made ...: 2944x1243 (2208x932 to 4416x1864) -> 4416x1864, mode
|
||||
// 2, preset M"), so with the Render resolution row at a fixed share, where the render size never changes, the feature was
|
||||
// still made for 2.25 times the pixels it was ever given at Quality, 4 times at Performance and 9 at Ultra Performance,
|
||||
// and the chain's three input images with it. NVIDIA's guide (DLSS Programming Guide 310.6.0, 3.2 and 5.2.8) has the
|
||||
// optimal render size passed to the creation "exactly as given", and says of the second generation transformer's presets
|
||||
// L and M that their cost follows the size the feature is initialized for, which is why it is to be the dynamic range's
|
||||
// largest only where there is a dynamic range. The same log has DLSS's own video memory at 1069 MB for that feature, where
|
||||
// NVIDIA's table has 456 MB for preset M at 4K.
|
||||
//
|
||||
// So: a fixed share gets a feature of its own render size (exact), as AMD FSR 4 gets a context of its own
|
||||
// (temporal_rules.h plan_context). Where the render size is the game's to change (the row's Auto), or once it has changed
|
||||
// under one output in this run, the feature is made for the mode's whole range, as before. DLAA's is the output's either
|
||||
// way.
|
||||
//
|
||||
// RX_VK_DLSS_FEATURE_MAX=1: every feature for its mode's largest render size, as before (whole_range()).
|
||||
//
|
||||
// No NGX, no Vulkan: tests/test_dlss_feature.py compiles this.
|
||||
#pragma once
|
||||
#include <cstdint>
|
||||
#include <cstdlib>
|
||||
|
||||
namespace dlss_feature {
|
||||
|
||||
/// RX_VK_DLSS_FEATURE_MAX (value, null when unset): not 0, every feature for its mode's largest render size, as before
|
||||
inline bool whole_range(const char* value) { return value && *value && std::atoi(value) != 0; }
|
||||
|
||||
/// What a feature is made for: w x h, and whether that is the one render size it takes (exact) or the largest of a range
|
||||
struct Size {
|
||||
uint32_t w, h;
|
||||
bool exact;
|
||||
};
|
||||
|
||||
/// The size to make a feature for, for a frame of rw x rh in a mode whose render sizes go up to max_w x max_h. aa: DLAA,
|
||||
/// the render size the output's (its range is that one size). range: the render size may change under this output (the
|
||||
/// row's Auto, the size seen changing, or the switch)
|
||||
inline Size size(bool aa, bool range, uint32_t rw, uint32_t rh, uint32_t max_w, uint32_t max_h) {
|
||||
if (aa || range || rw > max_w || rh > max_h) return {max_w, max_h, false};
|
||||
return {rw, rh, true};
|
||||
}
|
||||
|
||||
/// Whether a feature made for fw x fh takes a frame of rw x rh: its one size (exact), or any from min_w x min_h up to its own
|
||||
inline bool takes(bool exact, uint32_t fw, uint32_t fh, uint32_t min_w, uint32_t min_h, uint32_t rw, uint32_t rh) {
|
||||
if (exact) return rw == fw && rh == fh;
|
||||
return rw <= fw && rh <= fh && rw >= min_w && rh >= min_h;
|
||||
}
|
||||
|
||||
/// Whether this frame shows that the render size changes under one output: a feature of one size (exact) for the same
|
||||
/// output, preset and colour (same), and a frame of another size. The caller keeps it for the run: the next feature, and
|
||||
/// every one after it, is made for the mode's whole range
|
||||
inline bool found_range(bool same, bool exact, uint32_t fw, uint32_t fh, uint32_t rw, uint32_t rh) {
|
||||
return same && exact && (rw != fw || rh != fh);
|
||||
}
|
||||
|
||||
} // namespace dlss_feature
|
||||
@@ -93,6 +93,36 @@ inline void unjitter(float m[16], const float jitter[2]) {
|
||||
}
|
||||
}
|
||||
|
||||
/// ((Triforce, 11-20) the owner's "DLSS flicker": the whole picture and the shadows, while the camera moves, with NVIDIA
|
||||
/// DLSS frame generation on, and not without it.) Streamline's camera matrices are in the clip space of the depth buffer it
|
||||
/// is given: its depth value is that clip space's z over w, 1 at the near plane with depthInverted. The game's projection
|
||||
/// is NVN's, clip z from -1 at the near plane to 1 at the far one, and the depth frame generation gets is FSR's copy
|
||||
/// (vk_fsr3_inputs.cpp): d = (1 - ndc z) / 2 where the game's buffer has 0 at the near plane (flipped: Camera::flip_depth),
|
||||
/// (ndc z + 1) / 2 where it is reversed already. Given the game's own matrices with that depth, a depth read as clip z put
|
||||
/// what is far a fifth of a metre from the camera and what is near at the far plane, so whatever Streamline works out from
|
||||
/// the depth and the camera's move was wrong as soon as the camera moved, and right while it stood. Only DLSS frame
|
||||
/// generation takes matrices (FSR takes the near and far planes and the field of view), so nothing else showed it.
|
||||
/// The depth's clip space is the game's with z' = (w - z) / 2 (flipped) or (w + z) / 2: Z * clip, Z the identity but for
|
||||
/// its z row, (0 0 -1/2 1/2) or (0 0 1/2 1/2).
|
||||
/// A projection (or a view-projection) into the depth's clip space: Z * m, row major (clip = M * position)
|
||||
inline void depth_clip_projection(const float m[16], bool flipped, float out[16]) {
|
||||
const float s = flipped ? -0.5f : 0.5f;
|
||||
for (int i = 0; i < 16; i++) out[i] = m[i];
|
||||
for (int c = 0; c < 4; c++) out[2 * 4 + c] = s * m[2 * 4 + c] + 0.5f * m[3 * 4 + c];
|
||||
}
|
||||
/// A clip-to-clip matrix of the game's (this frame's clip space to the frame before's: reprojection) between the depth's
|
||||
/// clip spaces: Z * m * inverse(Z), inverse(Z)'s z row (0 0 -2 1) or (0 0 2 -1). The identity stays the identity
|
||||
inline void depth_clip_reprojection(const float m[16], bool flipped, float out[16]) {
|
||||
const float k = flipped ? -2.0f : 2.0f, t = flipped ? 1.0f : -1.0f;
|
||||
float r[16];
|
||||
for (int row = 0; row < 4; row++) { // (m * inverse(Z): the z column scaled, and added to the w column)
|
||||
r[row * 4 + 0] = m[row * 4 + 0], r[row * 4 + 1] = m[row * 4 + 1];
|
||||
r[row * 4 + 2] = k * m[row * 4 + 2];
|
||||
r[row * 4 + 3] = m[row * 4 + 3] + t * m[row * 4 + 2];
|
||||
}
|
||||
depth_clip_projection(r, flipped, out); // (Z * that)
|
||||
}
|
||||
|
||||
/// A jitter in clip space as DLSS takes it: in render pixels, x right and y down (clip y is up)
|
||||
inline void jitter_to_pixels(const float clip[2], unsigned width, unsigned height, float out[2]) {
|
||||
out[0] = clip[0] * 0.5f * (float)width;
|
||||
|
||||
74
gpu/vk/dlssg_pace.h
Normal file
74
gpu/vk/dlssg_pace.h
Normal file
@@ -0,0 +1,74 @@
|
||||
// (Triforce) Who holds the frame limit while NVIDIA DLSS frame generation runs: Reflex's sleep, where the game's frame
|
||||
// starts, or the game's own wait (host/nn_vi.c hos_vi_native_acquire). After NX2navi's runtime/dlssg_pace.h, whose rule
|
||||
// this is. A function of what the game's presenting thread sees and nothing else, so tests/test_dlssg_pace.py drives
|
||||
// it without a GPU (plain C: the test compiles it as C and as C++).
|
||||
//
|
||||
// The period counts the frames the game draws (the Game frame rate cap's, the host's own). Reflex holds it only while
|
||||
// every one of these is so, and the game's own wait holds the whole period otherwise: the caller allows it (the
|
||||
// Reflex frame limit row on, frames being generated now, a Streamline whose Reflex sleeps here and refused no limit);
|
||||
// there is a period (the caller passes 0 while no frame is generated: a menu, a load, a movie); a present went out,
|
||||
// and within DLSSG_HOLD_IDLE_NS (a minimized window gives the driver no present to time a frame by: Reflex's sleep is
|
||||
// left out then, vk_dlssg.cpp frame_start); the limit Reflex should hold is the one Streamline last took; and the
|
||||
// frames have not shown that Reflex does not hold it (for the rest of the run then).
|
||||
//
|
||||
// The frames are checked over windows of DLSSG_PACE_WINDOW presents Reflex held to one period and one multiplier.
|
||||
// Over half of them sooner than three quarters of the period: nothing holds the game, and the game's own wait takes
|
||||
// the limit for the run. Over half of them about the multiplier times the period apart (within half a period): the
|
||||
// driver counted the frames on screen, the generated ones too, so Reflex is given the period over the multiplier from
|
||||
// then on; the same again (a GPU that only managed that rate) gives the limit to the game's own wait for the run. A
|
||||
// frame a load or a hitch held counts as neither.
|
||||
#pragma once
|
||||
#include <stdint.h>
|
||||
|
||||
#define DLSSG_HOLD_IDLE_NS 200000000ull // no present for this long: Reflex's sleep is left out
|
||||
#define DLSSG_PACE_WINDOW 60 // presents over which Reflex's hold of the limit is checked
|
||||
|
||||
enum { DLSSG_PACE_SAME, DLSSG_PACE_FAST, DLSSG_PACE_SHOWN, DLSSG_PACE_SLOW }; // what the last window found
|
||||
|
||||
typedef struct {
|
||||
uint64_t last_ns; // the newest frame start Reflex held, 0 when the game's own wait held it
|
||||
uint64_t last_period; // its period
|
||||
int last_mult; // and its multiplier
|
||||
int window, fast, slow; // frames counted in the window, those too soon, those about the multiplier periods apart
|
||||
int shown; // Reflex counts the frames on screen: it is given the period over the multiplier
|
||||
int bad; // Reflex did not hold the limit: the game's own wait holds it for the rest of the run
|
||||
} DlssgPace;
|
||||
|
||||
// now: the clock where the game's frame starts. period_ns: the period of the cap in effect (0: no frame is generated
|
||||
// now, or no limit). mult: frames on screen per drawn frame now. hosted_ns: the clock after the newest present, 0
|
||||
// before the first. allowed: Reflex may hold the limit. applied_us: the limit Streamline last took. *want_us: the
|
||||
// limit Reflex should hold, 0 none. *found: what a window that ended here found (DLSSG_PACE_*). Returns 1 when Reflex
|
||||
// holds this frame's limit: the game's own wait steps aside, and only then.
|
||||
static inline int dlssg_pace_step(DlssgPace *p, uint64_t now, uint64_t period_ns, int mult, uint64_t hosted_ns,
|
||||
int allowed, uint32_t applied_us, uint32_t *want_us, int *found) {
|
||||
*found = DLSSG_PACE_SAME;
|
||||
if (mult < 1) mult = 1;
|
||||
if (!p->bad && p->last_ns && now >= p->last_ns && p->last_period == period_ns && p->last_mult == mult) {
|
||||
uint64_t d = now - p->last_ns, m = (uint64_t)mult;
|
||||
if (d < period_ns / 4 * 3) p->fast++;
|
||||
else if (m > 1 && d * 2 > period_ns * (2 * m - 1) && d * 2 < period_ns * (2 * m + 1)) p->slow++;
|
||||
if (++p->window >= DLSSG_PACE_WINDOW) {
|
||||
if (p->fast > DLSSG_PACE_WINDOW / 2) p->bad = 1, *found = DLSSG_PACE_FAST;
|
||||
else if (p->slow > DLSSG_PACE_WINDOW / 2 && p->shown) p->bad = 1, *found = DLSSG_PACE_SLOW;
|
||||
else if (p->slow > DLSSG_PACE_WINDOW / 2) p->shown = 1, *found = DLSSG_PACE_SHOWN;
|
||||
p->window = p->fast = p->slow = 0;
|
||||
}
|
||||
} else {
|
||||
p->window = p->fast = p->slow = 0; // a window counts only frames Reflex held to one period and multiplier
|
||||
}
|
||||
uint64_t us = 0;
|
||||
// a present stamped after now went out while this thread read the clock: as recent as one can be
|
||||
int live = hosted_ns && (hosted_ns >= now || now - hosted_ns < DLSSG_HOLD_IDLE_NS);
|
||||
if (allowed && period_ns && !p->bad && live) {
|
||||
uint64_t per = p->shown ? period_ns / (uint64_t)mult : period_ns;
|
||||
us = (per + 500) / 1000;
|
||||
if (!us) us = 1;
|
||||
if (us > 0x7fffffffu) us = 0; // no limit Reflex could be given: the game's own wait keeps it
|
||||
}
|
||||
*want_us = (uint32_t)us;
|
||||
int held = us && applied_us == (uint32_t)us;
|
||||
p->last_ns = held ? now : 0;
|
||||
p->last_period = held ? period_ns : 0;
|
||||
p->last_mult = mult;
|
||||
return held;
|
||||
}
|
||||
38
gpu/vk/dmabuf_gate.h
Normal file
38
gpu/vk/dmabuf_gate.h
Normal file
@@ -0,0 +1,38 @@
|
||||
// (Triforce, 11-22, the performance audit's OS-08, after Source A's) Whether guest memory is offered to the driver as
|
||||
// udmabuf dma-bufs at all, and what the start's line says (vk_memory.cpp import_dma_buf asks at its first call; plain code:
|
||||
// tests/test_dmabuf_driver_gate.py runs it anywhere).
|
||||
//
|
||||
// On Linux the guest window is a memfd and each pool is first offered to the driver as a dma-buf made by udmabuf. NVIDIA's
|
||||
// proprietary driver refuses every one (VK_ERROR_OUT_OF_DEVICE_MEMORY) and the pool ends as a host-pointer import anyway:
|
||||
// 49-50 refused imports a run and a kernel message at each start on Source A's RTX 4090, the first refusal 2.5-3.7 ms, one
|
||||
// of a 64 MB piece 198 ms on the presentation thread during a save load. So with that driver /dev/udmabuf is not opened
|
||||
// and the pools are host-pointer imports from the start. NVK and RADV have other driver ids and keep their dma-bufs.
|
||||
// RX_VK_DMABUF_NVIDIA=1: tried on NVIDIA's driver too, as before (for a later driver that takes them).
|
||||
#pragma once
|
||||
|
||||
namespace dmabuf_gate {
|
||||
|
||||
/// What the start found. fn: the host gave its memfd function (false also with RX_VK_NO_DMABUF). memfd: the window is a
|
||||
/// memfd. ext: the device imports dma-bufs (VK_EXT_external_memory_dma_buf and its fd properties call). nvidia: NVIDIA's
|
||||
/// proprietary driver, and RX_VK_DMABUF_NVIDIA not set
|
||||
struct Facts {
|
||||
bool fn = false, memfd = false, ext = false, nvidia = false;
|
||||
};
|
||||
|
||||
/// Whether /dev/udmabuf is opened: only then can a pool be offered as a dma-buf
|
||||
inline bool try_udmabuf(const Facts& f) { return f.fn && f.memfd && f.ext && !f.nvidia; }
|
||||
|
||||
/// As it was: NVIDIA's driver was offered them too
|
||||
inline bool try_udmabuf_before(const Facts& f) { return f.fn && f.memfd && f.ext; }
|
||||
|
||||
/// The start's "vk: guest memory as dma-bufs: ..." answer. opened: /dev/udmabuf opened (asked only after try_udmabuf)
|
||||
inline const char* answer(const Facts& f, bool opened) {
|
||||
if (try_udmabuf(f) && opened) return "yes";
|
||||
if (!f.fn) return "no (no memfd window, or RX_VK_NO_DMABUF)";
|
||||
if (!f.memfd) return "no (the window is anonymous memory)";
|
||||
if (!f.ext) return "no (VK_EXT_external_memory_dma_buf)";
|
||||
if (f.nvidia) return "no (NVIDIA's driver refuses them: host pointers)";
|
||||
return "no (/dev/udmabuf cannot be opened)";
|
||||
}
|
||||
|
||||
} // namespace dmabuf_gate
|
||||
41
gpu/vk/drawn_page_map.h
Normal file
41
gpu/vk/drawn_page_map.h
Normal file
@@ -0,0 +1,41 @@
|
||||
// (Triforce, 11-21, the performance audit's TEX-21, after Source A's) A texture the GPU drew gets its page map at its first
|
||||
// draw.
|
||||
//
|
||||
// A texture in a virtual pool (its storage mapped 64 KB a page) is checked through its mapping: texture_dirty keeps each
|
||||
// page's guest address (page_map) and says which pages were mapped anew or written since. The map is recorded at the
|
||||
// texture's first check, which for a texture uploaded from memory is its first upload. One the GPU drew first is marked
|
||||
// uploaded by the draw and never checked, so it has no map; once the game finalizes it, its first check finds none, takes
|
||||
// that as "everything was mapped since", and the whole storage is uploaded from memory over what the GPU drew: a HUD
|
||||
// label or a render target gone blank after an area change, and an upload of up to 15 MB each. Source A's sessions from
|
||||
// Hyrule Field to Castle Town: such whole uploads of HUD labels 6, other HUD 4 to 6, render targets 4 to 9, all to 0.
|
||||
// Here: the map is recorded at the first draw that says who drew it, by one walk of the mapping (the storage's size / 64 KB
|
||||
// lookups, once a texture). The later check then gives only the pages mapped anew or written by the CPU.
|
||||
// Not for a write that gives no reason ("?": a clear, the movie upscaler): those keep the path that brings a font cache's
|
||||
// letters back from memory. RX_VK_DRAWN_PAGE_MAP=0: no map at the draw, as before.
|
||||
//
|
||||
// The rule alone: tests/test_drawn_page_map.py compiles it, and texture_dirty cut from the source.
|
||||
#pragma once
|
||||
#include <cstddef>
|
||||
#include <cstdint>
|
||||
#include <cstdlib>
|
||||
#include <cstring>
|
||||
|
||||
namespace drawn_page_map {
|
||||
|
||||
inline constexpr uint64_t kVirtual = 1ull << 40; // (a virtual pool's addresses start here)
|
||||
inline constexpr uint64_t kPage = 1ull << 16; // (and its pages are 64 KB)
|
||||
|
||||
/// RX_VK_DRAWN_PAGE_MAP: off only with 0
|
||||
inline bool on(const char* asked) { return !asked || !*asked || atoi(asked) != 0; }
|
||||
|
||||
/// The pages of a storage of guest_size bytes
|
||||
constexpr size_t pages(uint64_t guest_size) { return (size_t)((guest_size + kPage - 1) / kPage); }
|
||||
|
||||
/// Whether a write by `why` into a texture at gpu, of guest_size bytes of storage, whose map holds map_pages pages now,
|
||||
/// records its map first
|
||||
inline bool seeds(bool switch_on, const char* why, uint64_t gpu, uint64_t guest_size, size_t map_pages) {
|
||||
if (!switch_on || !why || !std::strcmp(why, "?")) return false;
|
||||
return gpu >= kVirtual && guest_size != 0 && map_pages != pages(guest_size);
|
||||
}
|
||||
|
||||
} // namespace drawn_page_map
|
||||
140
gpu/vk/driver_faults.h
Normal file
140
gpu/vk/driver_faults.h
Normal file
@@ -0,0 +1,140 @@
|
||||
// Programs the driver faulted on, kept out of it, and one build of a program at a time (no game, no GPU: the rules and the
|
||||
// gate; vk_draw.cpp fault_quarantine, build_graphics, guarded_graphics). Apart from Vulkan so that
|
||||
// tests/test_driver_faults.py runs it alone, on many made-up PCs.
|
||||
//
|
||||
// (Triforce, 11-22, the performance audit's PIPE-57, after Source A's.) The guards catch a fault inside the driver's
|
||||
// compiler (vk_draw.cpp guarded_graphics), and a fault during the start's own shader work starts the game again
|
||||
// (driver_fault_restart). Nothing remembered the program: the next start asked the driver for it again, faulted again and
|
||||
// restarted again, and in play a second thread could ask for the same program while the driver was still inside the
|
||||
// first. Source A saw that on AMD's Windows driver (an RX 6800 XT): in a freeze, three threads sat in the driver on two
|
||||
// programs' builds while the presenting thread waited for one of them; a preparation stuck at "6625 of 6627, nothing
|
||||
// finished for 30 s" took 32.5 s, and 4.2 s with three faulted programs kept out on the first run, 2.3 s after. Triforce
|
||||
// has the same class of report: "two of 4926 on an RX 5700 XT, stuck at 99%".
|
||||
//
|
||||
// What. On AMD's own Windows driver (or everywhere with RX_VK_DRIVER_FAULTS=2):
|
||||
// - a program whose pipeline build faulted in the driver is not asked of the driver again this run, and its key goes to
|
||||
// triforce_driver_faults.txt beside the executable, one line under the vendor and the driver's version, written by the
|
||||
// fault's handler before the restart can end the process (so a start that faults does not fault again); later starts
|
||||
// with that driver skip those programs, and say so once (delete the file to try again; another driver version ignores
|
||||
// the lines);
|
||||
// - one build of a program at a time: a second waits at the gate, outside the driver, and a probe that asks the caches
|
||||
// only is told a compile is required.
|
||||
// Elsewhere (NVIDIA's, Intel's, Mesa's drivers, AMD's on Linux) nothing of it runs, as before: the NVIDIA fault of 11-13
|
||||
// was a race in the driver, not a bad program. RX_VK_DRIVER_FAULTS=0: nowhere.
|
||||
#pragma once
|
||||
|
||||
#include <condition_variable>
|
||||
#include <cstdint>
|
||||
#include <cstdlib>
|
||||
#include <mutex>
|
||||
#include <string>
|
||||
#include <unordered_set>
|
||||
#include <vector>
|
||||
|
||||
namespace driver_faults {
|
||||
|
||||
/// RX_VK_DRIVER_FAULTS: 0 off, 1 (unset) AMD's Windows driver, 2 every driver
|
||||
inline int mode(const char* env) {
|
||||
if (!env || !*env) return 1;
|
||||
const int v = std::atoi(env);
|
||||
return v <= 0 ? 0 : v >= 2 ? 2 : 1;
|
||||
}
|
||||
/// Whether the quarantine, its file and the gate are on: mode 2 anywhere, mode 1 on Windows with AMD's own driver
|
||||
/// (VK_DRIVER_ID_AMD_PROPRIETARY)
|
||||
inline bool on(int mode, bool windows, bool amd_proprietary) {
|
||||
return mode >= 2 || (mode == 1 && windows && amd_proprietary);
|
||||
}
|
||||
|
||||
inline constexpr int kLine = 35; // ("vvvvvvvv dddddddd kkkkkkkkkkkkkkkk\n")
|
||||
|
||||
/// One line of the file into out (kLine bytes and a 0): made without the C library, as the fault's handler makes it
|
||||
inline int line(char* out, uint32_t vendor, uint32_t driver, uint64_t key) {
|
||||
static const char hex[] = "0123456789abcdef";
|
||||
int n = 0;
|
||||
for (int i = 7; i >= 0; i--) out[n++] = hex[(vendor >> (4 * i)) & 0xF];
|
||||
out[n++] = ' ';
|
||||
for (int i = 7; i >= 0; i--) out[n++] = hex[(driver >> (4 * i)) & 0xF];
|
||||
out[n++] = ' ';
|
||||
for (int i = 15; i >= 0; i--) out[n++] = hex[(key >> (4 * i)) & 0xF];
|
||||
out[n++] = '\n';
|
||||
out[n] = 0;
|
||||
return n;
|
||||
}
|
||||
/// The keys of the file's lines for this vendor and driver version; another's, and a broken line, are passed over
|
||||
inline std::vector<uint64_t> parse(const std::string& text, uint32_t vendor, uint32_t driver) {
|
||||
std::vector<uint64_t> keys;
|
||||
const auto hexval = [](const std::string& s, size_t at, int digits, uint64_t* v) {
|
||||
if (at + (size_t)digits > s.size()) return false;
|
||||
uint64_t x = 0;
|
||||
for (int i = 0; i < digits; i++) {
|
||||
const char c = s[at + i];
|
||||
const int d = c >= '0' && c <= '9' ? c - '0' : c >= 'a' && c <= 'f' ? c - 'a' + 10 : c >= 'A' && c <= 'F' ? c - 'A' + 10 : -1;
|
||||
if (d < 0) return false;
|
||||
x = x << 4 | (uint64_t)d;
|
||||
}
|
||||
*v = x;
|
||||
return true;
|
||||
};
|
||||
size_t at = 0;
|
||||
while (at < text.size()) {
|
||||
size_t end = text.find('\n', at);
|
||||
if (end == std::string::npos) end = text.size();
|
||||
std::string l = text.substr(at, end - at);
|
||||
if (!l.empty() && l.back() == '\r') l.pop_back();
|
||||
uint64_t v = 0, d = 0, k = 0;
|
||||
if (l.size() == (size_t)kLine - 1 && l[8] == ' ' && l[17] == ' ' && hexval(l, 0, 8, &v) && hexval(l, 9, 8, &d) &&
|
||||
hexval(l, 18, 16, &k) && v == vendor && d == driver)
|
||||
keys.push_back(k);
|
||||
at = end + 1;
|
||||
}
|
||||
return keys;
|
||||
}
|
||||
|
||||
/// The run's quarantine and the gate: begin before a program's build, end after it (with whether it faulted)
|
||||
class Quarantine {
|
||||
public:
|
||||
enum class Begin { Go, Busy, Faulted };
|
||||
void load(const std::vector<uint64_t>& keys) {
|
||||
std::lock_guard lk(mu_);
|
||||
bad_.insert(keys.begin(), keys.end());
|
||||
}
|
||||
/// Faulted: not asked of the driver. Busy: another build of it is inside the driver and the caller may not wait (a
|
||||
/// probe of the caches). Go: this build may go in; the caller ends it with end(). may_wait: wait at the gate instead
|
||||
Begin begin(uint64_t key, bool may_wait) {
|
||||
std::unique_lock lk(mu_);
|
||||
for (;;) {
|
||||
if (bad_.count(key)) return Begin::Faulted;
|
||||
if (!busy_.count(key)) break;
|
||||
if (!may_wait) return Begin::Busy;
|
||||
cv_.wait(lk);
|
||||
}
|
||||
busy_.insert(key);
|
||||
return Begin::Go;
|
||||
}
|
||||
/// True when it faulted now for the first time
|
||||
bool end(uint64_t key, bool faulted) {
|
||||
bool fresh = false;
|
||||
{
|
||||
std::lock_guard lk(mu_);
|
||||
busy_.erase(key);
|
||||
if (faulted) fresh = bad_.insert(key).second;
|
||||
}
|
||||
cv_.notify_all();
|
||||
return fresh;
|
||||
}
|
||||
bool faulted(uint64_t key) const {
|
||||
std::lock_guard lk(mu_);
|
||||
return bad_.count(key) != 0;
|
||||
}
|
||||
size_t count() const {
|
||||
std::lock_guard lk(mu_);
|
||||
return bad_.size();
|
||||
}
|
||||
|
||||
private:
|
||||
mutable std::mutex mu_;
|
||||
std::condition_variable cv_;
|
||||
std::unordered_set<uint64_t> bad_, busy_;
|
||||
};
|
||||
|
||||
} // namespace driver_faults
|
||||
@@ -440,13 +440,6 @@ void EmitSetAttribute(EmitContext& ctx, IR::Attribute attr, Id value, [[maybe_un
|
||||
value = ctx.OpBitcast(output->type, value);
|
||||
}
|
||||
|
||||
static constexpr IR::Attribute cd0 = IR::Attribute::ClipDistance0;
|
||||
static constexpr IR::Attribute cd7 = IR::Attribute::ClipDistance7;
|
||||
|
||||
if (attr >= cd0 && attr <= cd7) {
|
||||
const u32 idx = (u32) attr - (u32) cd0;
|
||||
clip_distance_written.set(idx);
|
||||
}
|
||||
ctx.OpStore(output->pointer, value);
|
||||
}
|
||||
|
||||
|
||||
@@ -197,11 +197,9 @@ void EmitPrologue(EmitContext& ctx) {
|
||||
|
||||
if (Sirit::ValidId(ctx.clip_distances)) {
|
||||
for (u32 i = 0; i < ctx.profile.max_user_clip_distances; ++i) {
|
||||
if (!clip_distance_written.test(i)) {
|
||||
const Id idx = ctx.Const(i);
|
||||
const Id element = ctx.OpAccessChain(ctx.output_f32, ctx.clip_distances, idx);
|
||||
ctx.OpStore(element, ctx.Const(0.0f));
|
||||
}
|
||||
const Id idx = ctx.Const(i);
|
||||
const Id element = ctx.OpAccessChain(ctx.output_f32, ctx.clip_distances, idx);
|
||||
ctx.OpStore(element, ctx.Const(0.0f));
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
@@ -2312,7 +2312,11 @@ void EmitContext::DefineMesh(const IR::Program& program) {
|
||||
mesh_attribute_words = (std::max)(
|
||||
mesh_primitive_base + Common::DivCeil(mesh_info.max_primitives, 32U) * primitive_lines * 32,
|
||||
1U);
|
||||
mesh_map_bytes = 4 + mesh_info.Corners() * mesh_info.max_primitives;
|
||||
// Whole blocks of 32 primitives, as the attributes: the program writes the index bytes of every primitive block, lanes past
|
||||
// max_primitives too (a count of 0 wraps its min(count - 1, lane) to every lane). Sized to max_primitives, those writes were
|
||||
// clamped onto the last byte, the last primitive's last corner: with 126 primitives (most of OoT's) a stray vertex, flashing
|
||||
// triangles over Kokiri's foliage and ground (as nxnative found: "primitive index bytes for every primitive block")
|
||||
mesh_map_bytes = 4 + mesh_info.Corners() * Common::DivCeil(mesh_info.max_primitives, 32U) * 32;
|
||||
mesh_shared_u32 = TypePointer(spv::StorageClass::Workgroup, U32[1]);
|
||||
mesh_attributes = DefineVariable(*this, TypeArray(U32[1], Const(mesh_attribute_words)),
|
||||
std::nullopt, spv::StorageClass::Workgroup);
|
||||
|
||||
@@ -19,8 +19,6 @@
|
||||
|
||||
namespace Shader::Backend::SPIRV {
|
||||
|
||||
static std::bitset<8> clip_distance_written;
|
||||
|
||||
using Sirit::Id;
|
||||
|
||||
class VectorTypes {
|
||||
|
||||
@@ -284,6 +284,9 @@ inline LoweredProgram LowerToMaxwellWith(std::span<const Instruction> code, u32
|
||||
if (!entered)
|
||||
for (u32 pc = index + 1; pc < target; ++pc) dead_instructions[pc] = true;
|
||||
}
|
||||
// (hookshot) A shared memory store (STS) not yet followed by a barrier: an LDS after it may read what another lane of the
|
||||
// warp stored (see LDS below)
|
||||
bool shared_store_pending = false;
|
||||
for (u32 index = 0; index < code.size(); ++index) {
|
||||
auto insn = code[index];
|
||||
result.source_instruction_offsets.push_back(static_cast<u32>(result.instructions.size()));
|
||||
@@ -1701,6 +1704,7 @@ inline LoweredProgram LowerToMaxwellWith(std::span<const Instruction> code, u32
|
||||
case 0xb1d: // BAR.SYNC 0 (bit 80: .DEFER_BLOCKING), full workgroup.
|
||||
if (insn.words[1] || (insn.words[2] & ~0x10000U) || native_predicate != 7)
|
||||
return fail(index, "native workgroup barrier mode");
|
||||
shared_store_pending = false;
|
||||
emit(0xf0a8000000000000ULL | predicate | (u64{1} << 43) |
|
||||
(u64{1} << 44) | (u64{1} << 39) | (u64{7} << 39));
|
||||
break;
|
||||
@@ -1757,6 +1761,23 @@ inline LoweredProgram LowerToMaxwellWith(std::span<const Instruction> code, u32
|
||||
}
|
||||
const bool store = opcode != 0x984;
|
||||
const u32 data = store ? insn.Bits(32, 8) : static_cast<u32>(dest);
|
||||
// (hookshot) A warp stores to shared memory (STS) and reads it back (LDS) with no barrier between: the Switch runs a
|
||||
// warp's shared memory accesses in order, Vulkan orders another lane's store and this lane's load only through a
|
||||
// barrier, and the driver's compiler may move them past each other (OoT's 4/3 gameplay network on NVIDIA's cooperative
|
||||
// matrix path: 16x2 bright bars, Hookshot issue 2455). As for LDSM: a subgroup barrier before the first LDS after an
|
||||
// STS, only where the LDS is not predicated (every lane of the warp reaches it). RX_VK_SHARED_NO_LDS_BARRIER=1: none, as before
|
||||
// ((Triforce) both our STS forms, 0x388 and 0x988 with a uniform base, count as the store; the switch read here, as
|
||||
// this file's others are: upstream's env_on is in a header we do not have)
|
||||
static const bool lds_barrier = [] {
|
||||
const char* e = std::getenv("RX_VK_SHARED_NO_LDS_BARRIER");
|
||||
return !(e && *e && *e != '0');
|
||||
}();
|
||||
if (store) {
|
||||
shared_store_pending = true;
|
||||
} else if (shared_store_pending && lds_barrier && native_predicate == 7) {
|
||||
emit(0xef98000000000000ULL | predicate | (u64{1} << 20)); // (MEMBAR, bit 20: barrier_operations.cpp)
|
||||
shared_store_pending = false;
|
||||
}
|
||||
emit((store ? 0xef58000000000000ULL : 0xef48000000000000ULL) |
|
||||
predicate | data | (u64{address} << 8) | (u64{insn.Bits(40, 24)} << 20) |
|
||||
(u64{width} << 48));
|
||||
@@ -1877,6 +1898,9 @@ inline LoweredProgram LowerToMaxwellWith(std::span<const Instruction> code, u32
|
||||
return fail(index, "native matrix load mode");
|
||||
u32 address = insn.Bits(24, 8);
|
||||
const u64 offset = insn.Bits(40, 24);
|
||||
// (Fritchii f130f0c50; ours had it already) An address the results overwrite is copied first, to R254: the paths below
|
||||
// use R246-R253, and R253 is where each matrix's row address is shuffled to, so with the copy there matrices 1-3 of
|
||||
// an .x2/.x4 read matrix 0's rows (OoT has one such LDSM.x4, gpu/vk/tests/ldsm_address_test.cpp)
|
||||
if (uniform_offset) {
|
||||
const u32 uniform = uniform_register(insn.Bits(32, 8));
|
||||
if (uniform == ~0U) return fail(index, "native matrix load uniform offset");
|
||||
|
||||
@@ -146,6 +146,10 @@ void TranslateF2I(TranslatorVisitor& v, u64 insn, const IR::F16F32F64& src_a) {
|
||||
}()};
|
||||
const bool is_signed{f2i.is_signed != 0};
|
||||
const auto [max_bound, min_bound] = ClampBounds(f2i.dest_format, is_signed);
|
||||
// (Fritchii eba45b151, zortos293/Lichii#1051 by myshadowww) INT32_MAX rounds up to 2^31 as an F32 bound, so an input at or
|
||||
// above it reached OpConvertFToS out of range, which is undefined. The clamp stops below 2^31 and the exact INT32_MAX is
|
||||
// selected after the conversion
|
||||
const bool signed_f32_i32{f2i.dest_format == DestFormat::I32 && f2i.src_format == SrcFormat::F32 && is_signed};
|
||||
|
||||
IR::F16F32F64 intermediate;
|
||||
switch (f2i.src_format) {
|
||||
@@ -156,7 +160,7 @@ void TranslateF2I(TranslatorVisitor& v, u64 insn, const IR::F16F32F64& src_a) {
|
||||
break;
|
||||
}
|
||||
case SrcFormat::F32: {
|
||||
const IR::F32 max_val{v.ir.Imm32(static_cast<f32>(max_bound))};
|
||||
const IR::F32 max_val{v.ir.Imm32(signed_f32_i32 ? 0x1.fffffep30f : static_cast<f32>(max_bound))};
|
||||
const IR::F32 min_val{v.ir.Imm32(static_cast<f32>(min_bound))};
|
||||
intermediate = v.ir.FPClamp(rounded_value, min_val, max_val);
|
||||
break;
|
||||
@@ -173,6 +177,10 @@ void TranslateF2I(TranslatorVisitor& v, u64 insn, const IR::F16F32F64& src_a) {
|
||||
|
||||
const size_t bitsize{std::max<size_t>(32, BitSize(f2i.dest_format))};
|
||||
IR::U16U32U64 result{v.ir.ConvertFToI(bitsize, is_signed, intermediate)};
|
||||
if (signed_f32_i32) {
|
||||
result = IR::U32{v.ir.Select(v.ir.FPGreaterThanEqual(op_a, v.ir.Imm32(0x1p31f)),
|
||||
v.ir.Imm32(static_cast<u32>((std::numeric_limits<s32>::max)())), result)};
|
||||
}
|
||||
|
||||
bool handled_special_case = false;
|
||||
const bool special_nan_cases =
|
||||
|
||||
@@ -61,6 +61,20 @@ IR::U32 SmallAbs(TranslatorVisitor& v, const IR::U32& value, int bitsize) {
|
||||
return IR::U32{v.ir.Select(is_least, value, absolute)};
|
||||
}
|
||||
|
||||
// (Fritchii, zortos293 f6a8f5132) U32 -> F32 rounded toward zero. OpConvertUToF rounds to nearest whatever the IR's rounding
|
||||
// asks, so an input of 2^24 or more could round up: 0xffffff80..0xffffffff became 2^32 instead of 0x4f7fffff (a hash times 2^-32
|
||||
// reached 1.0). The top 24 significant bits make the float, the rest are dropped; an input below 2^24 converts exactly
|
||||
IR::F32 ConvertUnsigned32ToFloatRZ(TranslatorVisitor& v, const IR::U32& source) {
|
||||
auto& ir = v.ir;
|
||||
const IR::U32 msb{ir.Select(ir.INotEqual(source, ir.Imm32(0)), ir.FindUMsb(source), ir.Imm32(0))};
|
||||
const IR::U32 shift{ir.ISub(ir.UMax(msb, ir.Imm32(23)), ir.Imm32(23))};
|
||||
const IR::U32 retained{ir.ShiftRightLogical(source, shift)};
|
||||
const IR::U32 bits{ir.BitwiseOr(ir.ShiftLeftLogical(ir.IAdd(msb, ir.Imm32(127)), ir.Imm32(23)),
|
||||
ir.BitwiseAnd(retained, ir.Imm32(0x7fffff)))};
|
||||
const IR::F32 exact{ir.ConvertIToF(32, 32, false, ir.BitwiseAnd(source, ir.Imm32(0xffffff)))};
|
||||
return IR::F32{ir.Select(ir.IGreaterThan(msb, ir.Imm32(23), false), ir.BitCast<IR::F32>(bits), exact)};
|
||||
}
|
||||
|
||||
void I2F(TranslatorVisitor& v, u64 insn, IR::U32U64 src) {
|
||||
const Encoding i2f{insn};
|
||||
if (i2f.cc != 0) {
|
||||
@@ -99,7 +113,12 @@ void I2F(TranslatorVisitor& v, u64 insn, IR::U32U64 src) {
|
||||
.rounding = CastFpRounding(i2f.fp_rounding),
|
||||
.fmz_mode = IR::FmzMode::DontCare,
|
||||
};
|
||||
auto value{v.ir.ConvertIToF(size_t(dst_bitsize), size_t(conversion_src_bitsize), is_signed, src, fp_control)};
|
||||
IR::F16F32F64 value;
|
||||
if (!is_signed && conversion_src_bitsize == 32 && dst_bitsize == 32 && i2f.fp_rounding == FpRounding::RZ) {
|
||||
value = ConvertUnsigned32ToFloatRZ(v, IR::U32{src});
|
||||
} else {
|
||||
value = v.ir.ConvertIToF(size_t(dst_bitsize), size_t(conversion_src_bitsize), is_signed, src, fp_control);
|
||||
}
|
||||
if (i2f.neg != 0) {
|
||||
if (i2f.abs != 0 || !is_signed) {
|
||||
// We know the value is positive
|
||||
|
||||
65
gpu/vk/evict_reason.h
Normal file
65
gpu/vk/evict_reason.h
Normal file
@@ -0,0 +1,65 @@
|
||||
// Whether the write-back of a texture the eviction chose is stale: another image was drawn over its storage since its own
|
||||
// last drawing (no game, no GPU: a rule). tests/test_evict_reason.py
|
||||
//
|
||||
// (Triforce, 11-22, the performance audit's TEX-R404, after Source A's.) Over the budget, the eviction writes each chosen
|
||||
// texture the GPU drew back into the game's memory before it goes, so that one made again over that storage is uploaded
|
||||
// with what the GPU drew. The game draws its HUD labels (110 x 110, 46 x 46, 140 x 140) inside the storage of a
|
||||
// 1920 x 1080 target it no longer uses. The eviction's own tests do not see that: drawn_over looks only at images of the
|
||||
// same start, and target_overlap.h leaves out an image inside the target. So when the target went, its old texels landed
|
||||
// over the labels, the pages counted as written, and the labels, finalized, were uploaded from them as noise (Source A:
|
||||
// in 3 of 3 runs of two minutes in Hyrule Field at a 6 GB allowance, none with the rule).
|
||||
//
|
||||
// The rule: a write-back is left out when another image the GPU drew overlaps the texture's storage and was drawn in a
|
||||
// later frame than the texture's own last drawing. The texture still goes. One drawn in the same frame keeps the
|
||||
// write-back (a frame's targets drawn side by side), and so does one that only touches it. What is given up: the parts
|
||||
// of the texture the newer image does not cover are not written back either; made again there and sampled undrawn, it
|
||||
// starts from what the memory holds, as before the write-back existed. RX_VK_EVICT_STALE_FLUSH=1: written back, as before.
|
||||
#pragma once
|
||||
#include <algorithm>
|
||||
#include <cstddef>
|
||||
#include <cstdint>
|
||||
#include <cstdlib>
|
||||
#include <vector>
|
||||
|
||||
namespace evict_reason {
|
||||
|
||||
/// RX_VK_EVICT_STALE_FLUSH: unset, empty or 0 leaves a stale write-back out; anything else writes it back, as before
|
||||
inline bool flush_stale(const char* env) { return env && *env && std::atoi(env) != 0; }
|
||||
|
||||
/// Whether the storage [lo, hi) and [olo, ohi) share a byte (ranges that touch do not)
|
||||
constexpr bool overlaps(uint64_t lo, uint64_t hi, uint64_t olo, uint64_t ohi) {
|
||||
return lo < hi && olo < ohi && lo < ohi && olo < hi;
|
||||
}
|
||||
|
||||
/// Whether a texture last drawn in frame `last` holds texels older than what is in its storage: `other` is the frame of
|
||||
/// the newest drawing of another image over that storage (0: none)
|
||||
constexpr bool write_back_stale(uint64_t last, uint64_t other) { return other > last; }
|
||||
|
||||
/// A texture whose write-back the eviction is about to make
|
||||
struct Candidate {
|
||||
uint64_t lo = 0, hi = 0; // its storage
|
||||
size_t index = 0; // the caller's own: its place in the eviction's list
|
||||
const void* self = nullptr; // itself, so that it is not taken for another image over its own storage
|
||||
uint64_t newest = 0; // the frame of the newest drawing of another image over its storage (0: none)
|
||||
};
|
||||
|
||||
/// Puts the candidates in the order of their starts, for note(); returns the longest one's size
|
||||
inline uint64_t by_start(std::vector<Candidate>& c) {
|
||||
std::sort(c.begin(), c.end(), [](const Candidate& a, const Candidate& b) { return a.lo < b.lo; });
|
||||
uint64_t longest = 0;
|
||||
for (const Candidate& k : c) longest = std::max(longest, k.hi > k.lo ? k.hi - k.lo : 0);
|
||||
return longest;
|
||||
}
|
||||
|
||||
/// One image the GPU drew, [olo, ohi), last drawn in frame `o_written`: every candidate it overlaps (but itself) takes
|
||||
/// that frame when it is newer than what the candidate has. `c` is in by_start's order, `longest` what by_start gave
|
||||
inline void note(std::vector<Candidate>& c, uint64_t longest, uint64_t olo, uint64_t ohi, uint64_t o_written, const void* other) {
|
||||
if (ohi <= olo) return;
|
||||
// (a candidate over [olo, ohi) starts before ohi and after olo less its size: the search begins at olo less the longest)
|
||||
auto it = std::lower_bound(c.begin(), c.end(), olo > longest ? olo - longest : 0,
|
||||
[](const Candidate& k, uint64_t g) { return k.lo < g; });
|
||||
for (; it != c.end() && it->lo < ohi; ++it)
|
||||
if (it->self != other && overlaps(it->lo, it->hi, olo, ohi) && o_written > it->newest) it->newest = o_written;
|
||||
}
|
||||
|
||||
} // namespace evict_reason
|
||||
73
gpu/vk/fallback_size_policy.h
Normal file
73
gpu/vk/fallback_size_policy.h
Normal file
@@ -0,0 +1,73 @@
|
||||
// Mesh-fallback compute stages too big for two Mesa drivers are not handed to them (no game, no GPU: a rule).
|
||||
//
|
||||
// (Triforce, 11-22, the performance audit's PIPE-56, after Source A's.) A mesh program on a GPU without mesh shaders runs
|
||||
// as compute stages (vk_draw.cpp fallback_compute_pipelines). Every stage went to the driver, and a refusal was only
|
||||
// remembered after the build. Two drivers do not get that far on the largest stages:
|
||||
// - Mesa's RADV on GCN 1 to 4 (GFX6-8: vk_vendor.h radv_gcn). Source A's numbers under drm-shim with a 6 GB cap: the
|
||||
// 79,947-phi stage not built after 45-90 s and killed (25 GB at 150 s), where GFX9 and Navi built it in 13-14 s at
|
||||
// 2.9 GB. Limit: 50,000 OpPhi in one stage (2 of about 800 modules are over it);
|
||||
// - Mesa's Intel driver. An Iris Xe: a 251-call stage hit a 14 GB cap unfinished after 27 minutes. Limits: 20,000 OpPhi
|
||||
// or 200 OpFunctionCall in one stage (10 of 376 stages left out in Source A's full preparation).
|
||||
// A program with a stage over its driver's limit goes into the failed set before anything is built, as a program the
|
||||
// driver refused does: its draws are left out. Which objects those are is not known. No other driver has a limit, and the
|
||||
// switches set none there: NVIDIA, Intel's Windows driver, AMD's own drivers and RADV from Vega on build every stage.
|
||||
//
|
||||
// RX_VK_FALLBACK_PHI_LIMIT=<n>, RX_VK_FALLBACK_CALL_LIMIT=<n>: the two limits on those two drivers; 0: no limit.
|
||||
#pragma once
|
||||
#include <cstddef>
|
||||
#include <cstdint>
|
||||
#include <cstdlib>
|
||||
|
||||
namespace fallback_size {
|
||||
constexpr uint32_t kOpFunctionCall = 57, kOpPhi = 245;
|
||||
constexpr uint32_t kGcnPhis = 50000, kIntelPhis = 20000, kIntelCalls = 200;
|
||||
|
||||
struct Size {
|
||||
uint32_t phis = 0, calls = 0;
|
||||
};
|
||||
struct Limits {
|
||||
uint32_t phis = 0, calls = 0; // (0: none)
|
||||
};
|
||||
enum class Driver { Other, RadvGcn, MesaIntel };
|
||||
|
||||
/// A module's OpPhi and OpFunctionCall instructions. One that does not read as SPIR-V counts 0 of both
|
||||
inline Size count(const uint32_t* code, size_t words) {
|
||||
Size s;
|
||||
if (!code || words < 5 || code[0] != 0x07230203u) return s;
|
||||
for (size_t at = 5; at < words;) {
|
||||
const uint32_t n = code[at] >> 16, op = code[at] & 0xffffu;
|
||||
if (!n || at + n > words) return Size{};
|
||||
s.phis += op == kOpPhi, s.calls += op == kOpFunctionCall;
|
||||
at += n;
|
||||
}
|
||||
return s;
|
||||
}
|
||||
inline uint32_t spirv_phis(const uint32_t* code, size_t words) { return count(code, words).phis; }
|
||||
inline uint32_t spirv_calls(const uint32_t* code, size_t words) { return count(code, words).calls; }
|
||||
|
||||
/// A limit's switch: unset, empty or not a number, the driver's own; 0, none; else the number
|
||||
inline uint32_t parse_limit(const char* env, uint32_t own) {
|
||||
if (!env || !*env) return own;
|
||||
char* end = nullptr;
|
||||
const unsigned long long v = std::strtoull(env, &end, 10);
|
||||
if (end == env || *end) return own;
|
||||
return v > 0xffffffffull ? 0xffffffffu : (uint32_t)v;
|
||||
}
|
||||
|
||||
/// The driver: RADV on a GCN 1 to 4 GPU first (it is never Mesa's Intel driver), then Mesa's Intel driver
|
||||
constexpr Driver driver_of(bool radv_gcn, bool mesa_intel) {
|
||||
return radv_gcn ? Driver::RadvGcn : mesa_intel ? Driver::MesaIntel : Driver::Other;
|
||||
}
|
||||
|
||||
/// The limits on that driver, with the two switches' values (nullptr: unset). None on any other driver, whatever is set
|
||||
inline Limits limits(Driver d, const char* phi_env, const char* call_env) {
|
||||
if (d == Driver::Other) return Limits{};
|
||||
const Limits own = d == Driver::RadvGcn ? Limits{kGcnPhis, 0} : Limits{kIntelPhis, kIntelCalls};
|
||||
return Limits{parse_limit(phi_env, own.phis), parse_limit(call_env, own.calls)};
|
||||
}
|
||||
|
||||
/// The stage is over a limit: its program is not built
|
||||
constexpr bool refused(const Limits& l, const Size& s) {
|
||||
return (l.phis && s.phis > l.phis) || (l.calls && s.calls > l.calls);
|
||||
}
|
||||
} // namespace fallback_size
|
||||
38
gpu/vk/fg_motion.h
Normal file
38
gpu/vk/fg_motion.h
Normal file
@@ -0,0 +1,38 @@
|
||||
// (Triforce, 11-22, the owner's DLSS frame generation flicker, again) Whose motion vectors frame generation is given.
|
||||
//
|
||||
// Since FG-4 (vk_fsr3_inputs.cpp chain_motion), while NVIDIA DLSS or AMD FSR 4 upscales, frame generation's motion is the
|
||||
// upscaler chain's: the camera's from the depth, with every draw of the scene that stored its positions drawn again over
|
||||
// it (vk_temporal.cpp capture_motion). With NVIDIA DLSS frame generation that showed as the owner's flicker of 11-20 again
|
||||
// (the whole picture while the camera moved), on a build with the clip-space fix of that day (dlss_math.h
|
||||
// depth_clip_projection) in it. The owner's runs of 11-22 (DLSS at 50%, 2x frame generation, HDR, before the bloom):
|
||||
// - RX_VK_FG_OWN_MOTION=0, the camera's motion alone: no flicker;
|
||||
// - the chain's motion with the upscaler audit's commits all set back (RX_VK_CAPTURE_TABLES=0, RX_VK_REDRAW_STATE=0,
|
||||
// RX_VK_DLSS_FEATURE_MAX=1): the flicker, so it is not theirs;
|
||||
// - RX_VK_DLSS_PROBE=1 in that run: the redraw's motion of the scene's largest draw within 0.03 px of the camera's.
|
||||
// What in the chain's motion Streamline takes for a flicker is not known: its made frames are not in any log or shot. DLSS
|
||||
// frame generation finds what moves by itself with its own optical flow, and FSR's inputs give it the depth and the
|
||||
// camera, so it is given the camera's motion alone: what the first run showed. AMD's frame generation and FSR 3.1's
|
||||
// interpolation keep the chain's, as since FG-4 (no flicker of theirs is reported).
|
||||
//
|
||||
// RX_VK_DLSSG_OWN_MOTION=1: DLSS frame generation is given the chain's motion again, as before. RX_VK_FG_OWN_MOTION=0 still
|
||||
// gives every generator the camera's alone.
|
||||
//
|
||||
// No Vulkan: tests/test_fg_motion.py compiles this.
|
||||
#pragma once
|
||||
#include <cstdlib>
|
||||
|
||||
namespace fg_motion {
|
||||
|
||||
/// A switch read as the renderer reads its others: set and not 0 (value null when unset)
|
||||
inline bool set_on(const char* value) { return value && *value && std::atoi(value) != 0; }
|
||||
/// RX_VK_FG_OWN_MOTION=0 (value null when unset): the camera's motion alone, for every generator
|
||||
inline bool camera_alone(const char* value) { return value && *value && std::atoi(value) == 0; }
|
||||
|
||||
/// Whether frame generation is given the upscaler chain's motion this frame. dlssg: NVIDIA DLSS frame generation runs in
|
||||
/// this session (vk_dlssg.h active). fg_own: RX_VK_FG_OWN_MOTION; dlssg_own: RX_VK_DLSSG_OWN_MOTION (null when unset)
|
||||
inline bool chain(bool dlssg, const char* fg_own, const char* dlssg_own) {
|
||||
if (camera_alone(fg_own)) return false;
|
||||
return !dlssg || set_on(dlssg_own);
|
||||
}
|
||||
|
||||
} // namespace fg_motion
|
||||
42
gpu/vk/file_whole.h
Normal file
42
gpu/vk/file_whole.h
Normal file
@@ -0,0 +1,42 @@
|
||||
// (Triforce, handoff §8) A whole file at once, for file_writers.h: written as <path>.part and put in the old one's place, so
|
||||
// that a write cut short (the process ended in it, a full disk) leaves the old file or the new one, never part of one
|
||||
#pragma once
|
||||
#include "file_writers.h"
|
||||
#include <cstdio>
|
||||
#include <string>
|
||||
#ifdef _WIN32
|
||||
#ifndef NOMINMAX
|
||||
#define NOMINMAX
|
||||
#endif
|
||||
#ifndef WIN32_LEAN_AND_MEAN
|
||||
#define WIN32_LEAN_AND_MEAN
|
||||
#endif
|
||||
#include <windows.h>
|
||||
#endif
|
||||
|
||||
namespace file_writers {
|
||||
/// rename over an existing file, as each platform does it
|
||||
inline bool replace_file(const std::string& from, const std::string& to) {
|
||||
#ifdef _WIN32
|
||||
return MoveFileExA(from.c_str(), to.c_str(), MOVEFILE_REPLACE_EXISTING) != 0;
|
||||
#else
|
||||
return std::rename(from.c_str(), to.c_str()) == 0;
|
||||
#endif
|
||||
}
|
||||
|
||||
/// The whole file: <path>.part written and closed, then put in path's place; false (and no .part left) when any of it fails,
|
||||
/// and then path is as it was. Inside a Writing of its own (refused once the end has begun)
|
||||
inline bool write_whole(const std::string& path, const void* data, size_t n) {
|
||||
Writing w;
|
||||
if (!w) return false;
|
||||
const std::string part = path + ".part";
|
||||
FILE* f = std::fopen(part.c_str(), "wb");
|
||||
if (!f) return false;
|
||||
bool ok = !n || std::fwrite(data, 1, n, f) == n;
|
||||
ok = std::fclose(f) == 0 && ok;
|
||||
ok = ok && replace_file(part, path);
|
||||
if (!ok) std::remove(part.c_str());
|
||||
return ok;
|
||||
}
|
||||
inline bool write_whole(const std::string& path, const std::string& text) { return write_whole(path, text.data(), text.size()); }
|
||||
} // namespace file_writers
|
||||
96
gpu/vk/file_writers.h
Normal file
96
gpu/vk/file_writers.h
Normal file
@@ -0,0 +1,96 @@
|
||||
// (Triforce, handoff §8) The renderer's file writes made on threads of their own, and the process's end. The window's close ends
|
||||
// the game through ExitProcess (host/hos_relaunch.c), which stops every other thread wherever it is: a thread in the middle of
|
||||
// a file (a bug capture's picture, the pipeline cache's save in play, the order of the pipelines, a hang report, the prepared
|
||||
// record, the predicted pipelines' list, a debug picture or the frame log) left part of it. Now:
|
||||
// Writing one write, from its open to its close or rename, as a scope: counted while it runs, and refused once the end has
|
||||
// begun (on any thread but the one ending the process: its own last saves go on)
|
||||
// stop the process is about to end (HOS_OPT_EXITING, vk_replay.cpp option_set): no write begins from here on, and the
|
||||
// ones under way are waited for, up to a bound (2 s; RX_VK_WRITERS_WAIT_MS=<ms>, 0: no wait, as before). A write
|
||||
// that is still going then is cut by the exit as before; each such file is one write_whole makes, or one
|
||||
// whose readers take its whole records only, so it is the old file or the new one, or a list one record short
|
||||
// write_whole (file_whole.h) a whole file written as <path>.part and put in the old one's place: the old file or the new
|
||||
// one, never part of one, whenever the process ends
|
||||
// The threads themselves are not joined: they are detached and keep no state that outlives the process; what is waited for
|
||||
// is their write, the only part of them an end can damage. The standard library alone (file_whole.h has the
|
||||
// Windows part): tests/test_file_writers.py compiles it alone
|
||||
#pragma once
|
||||
#include <algorithm>
|
||||
#include <atomic>
|
||||
#include <chrono>
|
||||
#include <cstdio>
|
||||
#include <cstdlib>
|
||||
#include <string>
|
||||
#include <thread>
|
||||
|
||||
namespace file_writers {
|
||||
inline std::atomic<bool> g_stopping{false};
|
||||
inline std::atomic<int> g_busy{0}; // (writes under way on other threads than the ending one)
|
||||
inline std::atomic<int> g_refused{0}; // (writes not begun because the end had)
|
||||
inline thread_local bool t_ending = false; // (the thread that called stop: its writes are the exit's own)
|
||||
inline thread_local int t_inside = 0; // (Writings this thread holds: one inside another is part of it, never refused)
|
||||
|
||||
/// One write: `file_writers::Writing w; if (!w) return;` before the file is opened, and the scope ends after its close (and
|
||||
/// rename). The busy count goes up before the stop flag is read and stop sets the flag before it reads the count (both
|
||||
/// sequentially consistent): a write either sees the stop and does not begin, or stop sees it and waits for it. One begun
|
||||
/// inside another on the same thread (a save's own small files) goes on: the outer one is waited for whole
|
||||
class Writing {
|
||||
public:
|
||||
Writing() {
|
||||
if (t_ending || t_inside > 0) {
|
||||
on_ = true;
|
||||
t_inside++, nested_ = true;
|
||||
return;
|
||||
}
|
||||
g_busy.fetch_add(1);
|
||||
if (g_stopping.load()) {
|
||||
g_busy.fetch_sub(1);
|
||||
g_refused.fetch_add(1, std::memory_order_relaxed);
|
||||
return;
|
||||
}
|
||||
on_ = counted_ = true;
|
||||
t_inside++;
|
||||
}
|
||||
~Writing() {
|
||||
if (counted_ || nested_) t_inside--;
|
||||
if (counted_) g_busy.fetch_sub(1);
|
||||
}
|
||||
Writing(const Writing&) = delete;
|
||||
Writing& operator=(const Writing&) = delete;
|
||||
explicit operator bool() const { return on_; }
|
||||
|
||||
private:
|
||||
bool on_ = false, counted_ = false, nested_ = false;
|
||||
};
|
||||
|
||||
/// The end has begun: a loop that writes ends at its next turn
|
||||
inline bool stopping() { return g_stopping.load(std::memory_order_relaxed); }
|
||||
|
||||
/// The bound of stop's wait: 2 s, or RX_VK_WRITERS_WAIT_MS (0: no wait at all, no write refused: as before this)
|
||||
inline int exit_wait_ms() {
|
||||
const char* e = getenv("RX_VK_WRITERS_WAIT_MS");
|
||||
return e && *e ? std::max(0, atoi(e)) : 2000;
|
||||
}
|
||||
|
||||
struct Stopped {
|
||||
int left = 0; // (writes still under way when the wait ended: cut by the exit)
|
||||
double waited_ms = 0; // (how long the wait took)
|
||||
int refused = 0; // (writes not begun since the end began, so far)
|
||||
};
|
||||
|
||||
/// The process is about to end, on this thread: no write begins on another thread from here on, and the ones under way are
|
||||
/// waited for, up to ms. ms 0: nothing is stopped (the old behaviour)
|
||||
inline Stopped stop(int ms) {
|
||||
Stopped s;
|
||||
if (ms <= 0) return s;
|
||||
t_ending = true;
|
||||
g_stopping.store(true);
|
||||
const auto t0 = std::chrono::steady_clock::now();
|
||||
const auto until = t0 + std::chrono::milliseconds(ms);
|
||||
while (g_busy.load() > 0 && std::chrono::steady_clock::now() < until) std::this_thread::sleep_for(std::chrono::milliseconds(1));
|
||||
s.left = g_busy.load();
|
||||
s.waited_ms = std::chrono::duration<double, std::milli>(std::chrono::steady_clock::now() - t0).count();
|
||||
s.refused = g_refused.load();
|
||||
return s;
|
||||
}
|
||||
|
||||
} // namespace file_writers
|
||||
42
gpu/vk/finalized_obsolete.h
Normal file
42
gpu/vk/finalized_obsolete.h
Normal file
@@ -0,0 +1,42 @@
|
||||
// (Triforce, 11-21, the performance audit's TEX-43, after Source A's) A finalized texture the GPU drew whose storage the
|
||||
// game has since written over.
|
||||
//
|
||||
// A texture the GPU drew in one frame is kept after the game finalizes it (texture_sweep keep_finalized, up to
|
||||
// RX_VK_KEEP_FINALIZED_MB): the game may make a texture over that storage again and sample it without drawing it, as the
|
||||
// terrain does with its 260 x 260 tiles. Within the budget only a draw over one let it go, and the game's own CPU writes
|
||||
// to that memory never did: in players' logs about 200 render target images an hour stayed for good, most of them such
|
||||
// tiles, with the blocks they pin. Source A's rule: once the CPU has written three quarters or more of the storage since
|
||||
// the release, the game has put other data there, the image holds nothing that memory holds on the console, and the
|
||||
// sweep lets it go whatever the budget says. Only one of 32 KB or more (a small one shares its pages with what the game
|
||||
// keeps beside it, and a write there says nothing of it) and not one in a virtual pool. Its laps of Hyrule Field at a
|
||||
// 6 GB allowance: render target images 129, 166, 221 against 128, 131, 129; the kept ones 131 against 41.
|
||||
// The answer is kept with the texture: once rewritten it stays so until it is released again, and one not yet rewritten
|
||||
// is looked at again at most every kRecheck presents (the look walks its pages' write generations).
|
||||
// RX_VK_KEEP_OBSOLETE=1: kept, as before.
|
||||
//
|
||||
// The rule alone: tests/test_finalized_obsolete.py compiles it.
|
||||
#pragma once
|
||||
#include <cstdint>
|
||||
#include <cstdlib>
|
||||
|
||||
namespace finalized_obsolete {
|
||||
|
||||
inline constexpr uint64_t kMinBytes = 32u << 10;
|
||||
inline constexpr uint64_t kVirtual = 1ull << 40;
|
||||
inline constexpr uint64_t kRecheck = 60;
|
||||
|
||||
/// RX_VK_KEEP_OBSOLETE: not 0 keeps them, as before
|
||||
inline bool keep(const char* asked) { return asked && atoi(asked) != 0; }
|
||||
|
||||
/// Whether a texture can be one: finalized after the GPU wrote it, in no virtual pool, of 32 KB of storage or more
|
||||
constexpr bool candidate(bool released, bool gpu_written, uint64_t gpu, uint64_t guest_size) {
|
||||
return released && gpu_written && gpu < kVirtual && guest_size >= kMinBytes;
|
||||
}
|
||||
|
||||
/// Whether `written` bytes of a storage of guest_size, written by the CPU since the release, make it rewritten
|
||||
constexpr bool rewritten(uint64_t written, uint64_t guest_size) { return guest_size != 0 && written * 4 >= guest_size * 3; }
|
||||
|
||||
/// Whether one not rewritten when last looked at, in present `checked` (0: never), is looked at again in present `frame`
|
||||
constexpr bool due(uint64_t frame, uint64_t checked) { return !checked || frame - checked >= kRecheck; }
|
||||
|
||||
} // namespace finalized_obsolete
|
||||
74
gpu/vk/first_start.h
Normal file
74
gpu/vk/first_start.h
Normal file
@@ -0,0 +1,74 @@
|
||||
// (Triforce, 11-22, the performance audit's PRESENT-46, after Source A's) A first start sized to the card: the Render
|
||||
// resolution row set once on a weak GPU.
|
||||
//
|
||||
// With no options file every card started the same: the Render resolution at Auto, the game's dynamic resolution. Source A
|
||||
// reads the card and puts it in a tier; its tier 1 is an integrated GPU (a handheld's is one), a card without mesh shaders,
|
||||
// or a card with 4.5 GB of video memory or less, and it draws the 3D at 50% there, or at 33% on an integrated GPU beside
|
||||
// 4 processor threads or fewer (its note: an i3-1115G4's own GPU took 60 to 220 ms a frame at 50%).
|
||||
//
|
||||
// This is that test, for a first start only (vk_menu.cpp menu_first_start): no options file, RX_VK_LOW_SPEC not set,
|
||||
// nothing saved by it, so a start after the player has saved anything is the player's own. Left out, as the audit's
|
||||
// entry says: Source A's preset row, its fingerprint that chooses again on new hardware (it would overwrite the player's
|
||||
// rows), its tuner in play, any change of the upscaler or of the output's size, and its test of the SteamDeck variable,
|
||||
// which took desktop PCs started from that launcher for handhelds: a handheld is found by its integrated GPU alone.
|
||||
//
|
||||
// RX_VK_FIRST_START_AUTO: not set or 0, off (as before; the default until the owner has decided what a first start on a
|
||||
// weak card gets). 1: on. A larger number: on, the card taken as having that many MB of video memory, for the lab.
|
||||
//
|
||||
// No Vulkan, no menu: tests/test_first_start.py compiles this.
|
||||
#pragma once
|
||||
#include <cstdint>
|
||||
#include <cstdlib>
|
||||
|
||||
namespace first_start {
|
||||
|
||||
/// What is read of the card
|
||||
struct Card {
|
||||
bool integrated = false; // (VK_PHYSICAL_DEVICE_TYPE_INTEGRATED_GPU)
|
||||
bool mesh_shaders = true; // (VK_EXT_mesh_shader, as the device was made)
|
||||
uint64_t vram_mb = 0; // (its video memory; 0: not known)
|
||||
unsigned threads = 0; // (the processor's threads; 0: not known)
|
||||
};
|
||||
|
||||
constexpr uint64_t kSmallVramMb = 4608; // 4.5 GB: tier 1 up to here
|
||||
constexpr unsigned kFewThreads = 4; // an integrated GPU beside this many threads or fewer: 33%
|
||||
constexpr int kRowHalf = 4, kRowThird = 6; // opts.resolution's 50% and 33% (upscaler_rows.h)
|
||||
|
||||
enum Why { kKeep = 0, kIntegrated, kNoMeshShaders, kSmallVram };
|
||||
/// The Render resolution row for the card's first start: row 0, the row left as it is (not tier 1)
|
||||
struct Choice {
|
||||
int row = 0;
|
||||
Why why = kKeep;
|
||||
};
|
||||
|
||||
inline Choice choose(const Card& c) {
|
||||
if (c.integrated) return {c.threads && c.threads <= kFewThreads ? kRowThird : kRowHalf, kIntegrated};
|
||||
if (!c.mesh_shaders) return {kRowHalf, kNoMeshShaders};
|
||||
if (c.vram_mb && c.vram_mb <= kSmallVramMb) return {kRowHalf, kSmallVram};
|
||||
return {};
|
||||
}
|
||||
|
||||
inline const char* why_text(Why w) {
|
||||
switch (w) {
|
||||
case kIntegrated: return "an integrated GPU";
|
||||
case kNoMeshShaders: return "no mesh shaders";
|
||||
case kSmallVram: return "4.5 GB of video memory or less";
|
||||
default: return "not a weak card";
|
||||
}
|
||||
}
|
||||
|
||||
/// RX_VK_FIRST_START_AUTO (value, null when unset): -1 off; 0 on, the card's own video memory; otherwise on, the card
|
||||
/// taken as having that many MB
|
||||
inline int64_t switch_mb(const char* text) {
|
||||
if (!text || !*text) return -1;
|
||||
const long long v = std::atoll(text);
|
||||
return v <= 0 ? -1 : v == 1 ? 0 : (int64_t)v;
|
||||
}
|
||||
|
||||
/// Whether the first start's choice is made at all: the switch on, no options file, RX_VK_LOW_SPEC not asked (it sets the
|
||||
/// row itself), and no switch that sets the Render resolution for this run (HOS_DYNRES)
|
||||
inline bool applies(bool has_options, bool low_spec, bool resolution_env, int64_t switch_value) {
|
||||
return switch_value >= 0 && !has_options && !low_spec && !resolution_env;
|
||||
}
|
||||
|
||||
} // namespace first_start
|
||||
94
gpu/vk/flat_map.h
Normal file
94
gpu/vk/flat_map.h
Normal file
@@ -0,0 +1,94 @@
|
||||
// (hookshot) From Hookshot eda64e895 (dukenheimer; FlatMap by Corentin Bonjour, Hookshot PR 2774). tests/test_flat_map.py
|
||||
#pragma once
|
||||
#include <cstddef>
|
||||
#include <cstdint>
|
||||
#include <utility>
|
||||
#include <vector>
|
||||
|
||||
/// An integer key's value, kept in one array: open addressing, linear probing, at most 3/4 full, the slot from a
|
||||
/// multiplicative hash. For the replay's per-thread caches, searched on every draw: std::unordered_map divided by a prime
|
||||
/// and followed a node pointer per lookup. No erase of one key (those caches are only cleared whole). A clear keeps a small
|
||||
/// array and gives a large one back. An insert can move every entry: no reference or pointer is held across one.
|
||||
template <class K, class V>
|
||||
class FlatMap {
|
||||
public:
|
||||
struct Slot {
|
||||
K first{};
|
||||
bool used = false;
|
||||
V second{};
|
||||
};
|
||||
|
||||
Slot* find(K key) {
|
||||
if (!count_) return nullptr;
|
||||
for (size_t i = index(key);; i = (i + 1) & mask_) {
|
||||
Slot& s = slots_[i];
|
||||
if (!s.used) return nullptr;
|
||||
if (s.first == key) return &s;
|
||||
}
|
||||
}
|
||||
Slot* end() const { return nullptr; }
|
||||
|
||||
V& operator[](K key) {
|
||||
if (Slot* s = find(key)) return s->second;
|
||||
if ((count_ + 1) * 4 > slots_.size() * 3) grow();
|
||||
return place(key).second;
|
||||
}
|
||||
|
||||
size_t size() const { return count_; }
|
||||
size_t capacity() const { return slots_.size(); }
|
||||
|
||||
/// (Triforce) every entry, as f(key, value&): no insert inside f (it can move every entry)
|
||||
template <class F>
|
||||
void each(F&& f) {
|
||||
for (Slot& s : slots_)
|
||||
if (s.used) f(s.first, s.second);
|
||||
}
|
||||
|
||||
void clear() {
|
||||
if (count_ && slots_.size() > kKeptSlots) { // (a cache cleared at its size limit: its memory goes back too)
|
||||
count_ = 0;
|
||||
std::vector<Slot>().swap(slots_);
|
||||
mask_ = 0, shift_ = 64;
|
||||
return;
|
||||
}
|
||||
reset();
|
||||
}
|
||||
|
||||
/// Empty, the array kept whatever its size: for a map emptied often that fills again to about the same size
|
||||
/// (RecordingRes::c1_at, at every recording's start), so it isn't given back and grown again from 16 each time
|
||||
void reset() {
|
||||
if (!count_) return;
|
||||
count_ = 0;
|
||||
for (Slot& s : slots_)
|
||||
if (s.used) s = Slot{};
|
||||
}
|
||||
|
||||
private:
|
||||
static constexpr size_t kKeptSlots = 4096;
|
||||
|
||||
size_t index(K key) const { return (size_t)((uint64_t)key * 0x9E3779B97F4A7C15ull >> shift_); }
|
||||
|
||||
Slot& place(K key) {
|
||||
size_t i = index(key);
|
||||
while (slots_[i].used) i = (i + 1) & mask_;
|
||||
Slot& s = slots_[i];
|
||||
s.first = key, s.used = true;
|
||||
count_++;
|
||||
return s;
|
||||
}
|
||||
|
||||
void grow() {
|
||||
std::vector<Slot> old(slots_.empty() ? 16 : slots_.size() * 2);
|
||||
old.swap(slots_);
|
||||
mask_ = slots_.size() - 1;
|
||||
shift_ = 64;
|
||||
for (size_t n = slots_.size(); n > 1; n >>= 1) shift_--;
|
||||
count_ = 0;
|
||||
for (Slot& s : old)
|
||||
if (s.used) place(s.first).second = std::move(s.second);
|
||||
}
|
||||
|
||||
std::vector<Slot> slots_;
|
||||
size_t count_ = 0, mask_ = 0;
|
||||
unsigned shift_ = 64;
|
||||
};
|
||||
78
gpu/vk/focus_anim.h
Normal file
78
gpu/vk/focus_anim.h
Normal file
@@ -0,0 +1,78 @@
|
||||
// (Triforce, 11-19) The Mods and Quit entries under the pause menu's System list, and Quit's dialog, animated as the game
|
||||
// animates its own System page buttons and its own dialogs. His "the mods and quit tabs ... have no animations either" and
|
||||
// "the quit menu submenu also has no animations". Each button of the System list (and of the game's dialogs) is the part
|
||||
// PaCommonBtnThin_00 of the Common archive's layout; a dialog is SystemWindow_00 with three of them, 188 units apart. The
|
||||
// parts play their own animations; here are their times and amounts, read from the layouts' curves (the keys: a frame, a
|
||||
// value and a slope, Hermite between them, 60 frames a second, as mods_entry_anim.h reads the title's):
|
||||
// Select 8 frames: the cursor (N_Cursor_00) shown from the first, its alpha from 0 to 1 by frame 4
|
||||
// UnSelect 14 frames: its alpha from 1 to 0 by frame 8, hidden at 14
|
||||
// In 20 frames: the window's question and buttons (N_In_00) from alpha 0 to 1 by frame 15 (slow to start: its
|
||||
// slopes 0.02 and 0.14), and rising into place from 38 units under it (122 to 142 at frame 5, 160 at 20)
|
||||
// Out 2 frames: their alpha from 1 to 0, straight
|
||||
// The buttons' cursor does not slide from one to the next: the button let go fades its cursor out while the new one fades
|
||||
// its in. No ImGui and no Vulkan: tests/test_pause_mods_anim.py compiles it alone.
|
||||
#pragma once
|
||||
#include <algorithm>
|
||||
#include <cstdint>
|
||||
#include "mods_entry_anim.h" // (the layouts' curves: Key, curve, frames)
|
||||
|
||||
namespace focus_anim {
|
||||
|
||||
using mods_entry::curve;
|
||||
using mods_entry::frames;
|
||||
using mods_entry::Key;
|
||||
using mods_entry::kFrameNs;
|
||||
|
||||
// (PaCommonBtnThin_00's: N_Cursor_00's alpha)
|
||||
constexpr Key kSelect[] = {{0, 0, 0}, {4, 1, 0}};
|
||||
constexpr Key kUnSelect[] = {{0, 1, 0}, {8, 0, 0}};
|
||||
constexpr int64_t kSelectNs = 8 * kFrameNs, kUnSelectNs = 14 * kFrameNs;
|
||||
// (SystemWindow_00's: N_In_00's alpha and its height, its place 160 units up)
|
||||
constexpr Key kInAlpha[] = {{0, 0, 0.02f}, {15, 1, 0.14f}};
|
||||
constexpr Key kInY[] = {{0, 122, 5.6094f}, {5, 142, 2.50235f}, {20, 160, 0}};
|
||||
constexpr Key kOutAlpha[] = {{0, 1, -0.5f}, {2, 0, -0.5f}};
|
||||
constexpr float kInPlace = 160;
|
||||
constexpr int64_t kInNs = 20 * kFrameNs, kOutNs = 2 * kFrameNs;
|
||||
|
||||
/// A button's cursor from one drawing to the next: how much of it shows (N_Cursor_00's alpha, 0 to 1)
|
||||
struct Focus {
|
||||
bool sel = false; // (it had the cursor at the last step)
|
||||
float from = 0; // (shown when the cursor last came or went)
|
||||
int64_t since_ns = 0; // (...and when)
|
||||
float shown = 0;
|
||||
/// One drawing at now: whether the button has the cursor. settle: as it stands once its animation is done (a still,
|
||||
/// a picture that compares, a window just opened)
|
||||
void step(int64_t now, bool selected, bool settle = false) {
|
||||
if (settle || !since_ns) {
|
||||
sel = selected, from = shown = selected ? 1.0f : 0.0f, since_ns = now;
|
||||
return;
|
||||
}
|
||||
if (selected != sel) from = shown, sel = selected, since_ns = now;
|
||||
const float f = frames(now - since_ns);
|
||||
// (a change back half way: from where it is, as mods_entry::Anim, not from the curve's own start)
|
||||
shown = sel ? std::max(from, curve(kSelect, f)) : from * curve(kUnSelect, f);
|
||||
shown = std::clamp(shown, 0.0f, 1.0f);
|
||||
}
|
||||
};
|
||||
|
||||
/// A dialog's window t ns after it opened (In) or after it closed (Out, out = true): its alpha, and how far under its place
|
||||
/// it is (the layout's units, up positive: 0 in place)
|
||||
struct Window {
|
||||
float alpha = 1, drop = 0;
|
||||
};
|
||||
inline Window window_at(int64_t t, bool out = false) {
|
||||
Window w;
|
||||
if (t < 0) t = 0;
|
||||
const float f = frames(t);
|
||||
if (out) {
|
||||
w.alpha = std::clamp(curve(kOutAlpha, f), 0.0f, 1.0f);
|
||||
return w;
|
||||
}
|
||||
w.alpha = std::clamp(curve(kInAlpha, f), 0.0f, 1.0f);
|
||||
w.drop = std::max(0.0f, kInPlace - curve(kInY, f));
|
||||
return w;
|
||||
}
|
||||
/// (the window takes input once its In has played, as the game's)
|
||||
inline bool window_ready(int64_t t) { return t >= kInNs; }
|
||||
|
||||
} // namespace focus_anim
|
||||
93
gpu/vk/foliage_shadow.h
Normal file
93
gpu/vk/foliage_shadow.h
Normal file
@@ -0,0 +1,93 @@
|
||||
// (Triforce) The foliage in the sun's far shadow cascades, made cheaper (vk_draw.cpp draw_cull, for dynamic_state; and
|
||||
// foliage_rate, for the mesh draws), apart from Vulkan so that tests/test_foliage_shadow.py can run its rules. After NX2navi's
|
||||
// 754ded9 (epicmeme546's): 18 to 43 fps at Kokiri Forest's entrance on an RX 7900 XT there, the two halves together.
|
||||
//
|
||||
// The game draws its foliage cards (grass, ferns, leaves) double-sided with mesh programs, into the shadow cascades too: the
|
||||
// sun's three cascades are the layers of one depth array, 1600 x 1600 in the owner's logs. Seen from the sun the stacked,
|
||||
// alpha-tested cards cover the same texels many times over: NX2navi counted about 1,000 million fragments in one such draw of
|
||||
// the far cascade. A cascade here is Triforce's own test for one, as for the cascades' depth bias in dynamic_state: depth only,
|
||||
// square, wider than 768 (the characters' own shadow maps, 768 and less, are never touched).
|
||||
//
|
||||
// 1. The cull (RX_VK_FOLIAGE_SHADOW_CULL, on from layer 1 as NX2navi has it). A draw of a mesh program (native, or through the
|
||||
// compute fallback) into a cascade from the given layer on, that culls neither face, is drawn with its back faces culled (its
|
||||
// front ones with an "f"): half the cards' fragments there. Cull mode is dynamic state, so no other pipeline is made. It is not
|
||||
// quite the same picture: a card that faces away from the sun no longer casts its shadow in those cascades. Where cards stand in
|
||||
// crossed pairs or clumps (the game's foliage almost everywhere) the others cover the same texels; a lone card's shadow can go,
|
||||
// in the middle and far cascades only, where a texel is tens of centimetres to metres of ground and such a card is a fraction of
|
||||
// one. NX2navi kept it so (nothing done for a lone card), and so does this; the near cascade (layer 0) is drawn as the game asks.
|
||||
// RX_VK_FOLIAGE_SHADOW_CULL=off: as the game asks everywhere.
|
||||
//
|
||||
// 2. The shading rate (RX_VK_FOLIAGE_SHADOW_VRS, from layer 2 as NX2navi has it). The same draws, of native mesh programs, from
|
||||
// that layer on shade one fragment per 2 x 2 pixels (VK_KHR_fragment_shading_rate, pipelineFragmentShadingRate): the leaves'
|
||||
// alpha test runs a quarter as often. Depth is still tested and written for each texel; what is coarser is the alpha cut-out, in
|
||||
// 2 x 2 texels of the far cascade, metres of ground each. The rate is dynamic state in the pipelines that can draw such a draw and
|
||||
// in no other: native mesh programs with no colour target and a depth one (pipeline_rate), whose key that already names; they
|
||||
// set it at each draw (1 x 1 into any other target), their stage libraries are keyed apart. Where the device has no
|
||||
// pipelineFragmentShadingRate (older Intel, some APUs), without native mesh shaders, or RX_VK_FOLIAGE_SHADOW_VRS=off, the
|
||||
// extension is not enabled and no pipeline changes.
|
||||
#pragma once
|
||||
|
||||
#include <cstdint>
|
||||
#include <cstdlib>
|
||||
#include <cstring>
|
||||
|
||||
namespace mkwvk::foliage_shadow {
|
||||
|
||||
constexpr int kNone = 0, kFront = 1, kBack = 2; // (NVN's faces: NVN_FACE_NONE, _FRONT, _BACK)
|
||||
constexpr uint32_t kCascadeMin = 769; // (a cascade is wider than 768: vk_draw.cpp dynamic_state's test)
|
||||
constexpr int kCullFrom = 1, kRateFrom = 2; // (NX2navi's layers: the cull from the middle cascade, the rate in the far one)
|
||||
|
||||
struct Rule {
|
||||
int from = -1; // (the first cascade layer it applies to; -1: off)
|
||||
bool front = false; // (the front faces culled, not the back ones)
|
||||
};
|
||||
|
||||
/// RX_VK_FOLIAGE_SHADOW_CULL=<layer>[f]: from that layer of the cascades on (the game has three, 0 to 2), the back faces culled,
|
||||
/// the front ones with an "f" after the number. Unset or empty: from layer 1, the back faces (NX2navi's). Anything not starting
|
||||
/// with a digit ("off", "-1"): off
|
||||
inline Rule parse(const char* e) {
|
||||
Rule r;
|
||||
if (!e || !*e) return r.from = kCullFrom, r;
|
||||
if (*e < '0' || *e > '9') return r;
|
||||
r.from = std::atoi(e);
|
||||
r.front = std::strchr(e, 'f') != nullptr;
|
||||
return r;
|
||||
}
|
||||
|
||||
/// RX_VK_FOLIAGE_SHADOW_VRS=<layer>: from that layer of the cascades on, one fragment per 2 x 2 pixels. Unset or empty: from
|
||||
/// layer 2 (NX2navi's). Anything not starting with a digit ("off", "-1"): off, and the device's extension is not asked for.
|
||||
/// Returns the first layer, -1 off
|
||||
inline int parse_rate(const char* e) {
|
||||
if (!e || !*e) return kRateFrom;
|
||||
if (*e < '0' || *e > '9') return -1;
|
||||
return std::atoi(e);
|
||||
}
|
||||
|
||||
/// A draw into a cascade from layer `from` on (no colour target, a depth target width x height, square and wider than 768, the
|
||||
/// depth view's first layer at from or past it), of a mesh program, that culls neither face as the game draws it
|
||||
inline bool far_foliage(int from, int game, bool mesh_program, int colour_targets, bool depth, uint32_t width, uint32_t height,
|
||||
uint32_t layer) {
|
||||
if (from < 0 || game != kNone || !mesh_program || colour_targets != 0 || !depth) return false;
|
||||
return width == height && width >= kCascadeMin && layer >= (uint32_t)from;
|
||||
}
|
||||
|
||||
/// The face a draw culls: the game's (cull, NVN's face), or under the rule the one it names, for far_foliage's draws
|
||||
inline int cull(const Rule& rule, int game, bool mesh_program, int colour_targets, bool depth, uint32_t width, uint32_t height,
|
||||
uint32_t layer) {
|
||||
if (!far_foliage(rule.from, game, mesh_program, colour_targets, depth, width, height, layer)) return game;
|
||||
return rule.front ? kFront : kBack;
|
||||
}
|
||||
|
||||
/// Whether a pipeline has the shading rate as dynamic state: the rate is on (on), the program a native mesh one, no colour
|
||||
/// target and a depth one. Only what its key names (vk_draw.cpp pipeline_key): every pipeline of a key is made alike
|
||||
inline bool pipeline_rate(bool on, bool native_mesh, int colour_targets, bool depth_format) {
|
||||
return on && native_mesh && colour_targets == 0 && depth_format;
|
||||
}
|
||||
|
||||
/// The fragment size (1 or 2, both ways) a draw through such a pipeline sets: 2 for far_foliage's draws from layer `from` on
|
||||
inline uint32_t rate(int from, int game, bool native_mesh, int colour_targets, bool depth, uint32_t width, uint32_t height,
|
||||
uint32_t layer) {
|
||||
return far_foliage(from, game, native_mesh, colour_targets, depth, width, height, layer) ? 2u : 1u;
|
||||
}
|
||||
|
||||
} // namespace mkwvk::foliage_shadow
|
||||
23
gpu/vk/fps_cap.h
Normal file
23
gpu/vk/fps_cap.h
Normal file
@@ -0,0 +1,23 @@
|
||||
/* The Game frame rate cap row's "Monitor's rate" (vk_menu.cpp), as plain code: the cap a display's refresh rate gives.
|
||||
host/nn_vi.c fps_max asks the renderer for it at the start (HOS_OPT_DISPLAY_CAP, vk_replay.cpp's option), and the row
|
||||
says the same number beside its name. After NX2navi's fps_limit = display. No Vulkan, no menu:
|
||||
tests/test_fps_cap_monitor.py checks it. */
|
||||
#pragma once
|
||||
|
||||
namespace mkwvk::fps_cap {
|
||||
|
||||
/// The row's value for Monitor's rate (frame_rate_cap= in the options file): after 60, 90, 120, 144, 165 and 240 (0 to
|
||||
/// 5); 45, 40 and 30 are -1 to -3
|
||||
constexpr int kMonitor = 6;
|
||||
/// The caps the game's pacing and its time step take (host/nn_vi.c): a display under 30 Hz or over 240 gets the nearest
|
||||
constexpr int kLowest = 30, kHighest = 240;
|
||||
|
||||
/// The cap for a display's refresh rate in mHz (59,940 is 60, 143,856 is 144): its rate rounded to whole frames a
|
||||
/// second, kLowest to kHighest. 0 when the rate is not known (mhz 0 or less): the host then keeps the console's 60.
|
||||
inline int of_display(int mhz) {
|
||||
if (mhz <= 0) return 0;
|
||||
const int hz = (int)(((long long)mhz + 500) / 1000);
|
||||
return hz < kLowest ? kLowest : hz > kHighest ? kHighest : hz;
|
||||
}
|
||||
|
||||
} // namespace mkwvk::fps_cap
|
||||
48
gpu/vk/frame_trace.h
Normal file
48
gpu/vk/frame_trace.h
Normal file
@@ -0,0 +1,48 @@
|
||||
// (Triforce, wp20-onepct) The per-frame trace's split of a frame (RX_VK_FRAME_TRACE=1: triforce_frames.csv beside the game, one
|
||||
// line a frame, vk_replay.cpp status_tick). Its first columns say what the replay did (the draws' parts, sampled and summed over
|
||||
// every thread that records: the parallel replay's helpers too); they could not say what made a slow frame that was not the
|
||||
// draws': the lab's Market frames over the 99th percentile were 7-21% explained by them (r37, r38, r41). These columns, after
|
||||
// the first ones, split every frame the way the stall line splits a frame of 50 ms or more (stall_watch.h): the replay's own
|
||||
// time (its submits and the present phase, on every thread that replays: on the game's threads with the synchronous replay),
|
||||
// the present's own, the dispatches and the queue submits within the replay, the replay thread's waits for its helpers'
|
||||
// segments, and what is left: outside the replay and the present, the game's own update. Of that, the game's threads spend
|
||||
// some in our work: their fence waits and the captures at their submits; and the waits for write-backs still on the GPU (the
|
||||
// captures' late pages, the reads and bindings put off), with the part of them after the writer's recording was done on the
|
||||
// GPU (the submissions' timeline held back by work stopped at an event wait: vk_replay.cpp vk_timeline_wait), and the queue
|
||||
// submits made while a background import ran. Plain functions: tests/test_frame_trace.py
|
||||
#pragma once
|
||||
#include <cstddef>
|
||||
#include <cstdio>
|
||||
|
||||
namespace mkwvk::frame_trace {
|
||||
|
||||
/// One frame's split, in milliseconds
|
||||
struct Split {
|
||||
double interval = 0; // (present to present)
|
||||
double replay = 0; // (the submits and the present phase, every thread that replays: the stall line's "replay")
|
||||
double present = 0; // (the present's own: the swapchain's image, its fence, the present call)
|
||||
double dispatch = 0, qsubmit = 0; // (in replay: the dispatches, the queue submits)
|
||||
double par_wait = 0; // (in replay: the replay thread's waits for the helper threads' segments)
|
||||
double game_wait = 0; // (the game's fence waits, every thread)
|
||||
double captures = 0; // (the submit-time captures on the game's threads)
|
||||
double wb_wait = 0; // (the waits for write-backs still on the GPU, wherever they were: memory sync's capture, a read)
|
||||
double wb_held = 0; // (of them, the part after the writer's recording was done: the timeline held, not the GPU)
|
||||
double qs_bg = 0; // (the queue submits made while a background import ran)
|
||||
double gpu_waits = 0; // ((Triforce, 11-22, SYNC-27) every thread's time in calls that block on the GPU: gpu_wait_kinds.h)
|
||||
};
|
||||
|
||||
/// Outside the replay and the present: the game's own update, its waits for the replay and the GPU among it. As the stall line's
|
||||
inline double outside(const Split& s) { return s.interval - s.replay - s.present; }
|
||||
|
||||
/// The columns, after the trace's first ones (their names; the line's values in this order: line)
|
||||
constexpr const char* kColumns =
|
||||
"replay_ms,present_ms,dispatch_ms,qsubmit_ms,par_wait_ms,outside_ms,game_wait_ms,captures_ms,wb_wait_ms,wb_held_ms,qs_bg_ms,"
|
||||
"gpu_waits_ms";
|
||||
|
||||
/// The values for kColumns, comma-separated, no line end: snprintf's count
|
||||
inline int line(const Split& s, char* out, size_t cap) {
|
||||
return std::snprintf(out, cap, "%.2f,%.2f,%.2f,%.2f,%.2f,%.2f,%.2f,%.2f,%.2f,%.2f,%.2f,%.2f", s.replay, s.present, s.dispatch,
|
||||
s.qsubmit, s.par_wait, outside(s), s.game_wait, s.captures, s.wb_wait, s.wb_held, s.qs_bg, s.gpu_waits);
|
||||
}
|
||||
|
||||
} // namespace mkwvk::frame_trace
|
||||
49
gpu/vk/framegen_idle.h
Normal file
49
gpu/vk/framegen_idle.h
Normal file
@@ -0,0 +1,49 @@
|
||||
// (Triforce, 11-21, the performance audit's PRESENT-18, after Source A's) When frame generation's images are let go.
|
||||
//
|
||||
// The built-in interpolation's frames, lumas, flows and made-frame sets (and, for AMD FSR 4's frame generation, the four
|
||||
// images of the picture without the HUD) are made at the first frame that needs them and were kept until a resize: turned
|
||||
// off, or left for NVIDIA's own frame generation, they stayed in VRAM for the run. Source A measured 150 to 160 MiB given
|
||||
// back at 2560 x 1440 with 2x after its switch away. They are let go after kPresents presents in a row with frame
|
||||
// generation off, about two seconds at 60: a row stepped through on the way to another choice keeps them. A present that
|
||||
// recorded made frames counts as on whatever the row says by then: its set is busy until the display thread has shown
|
||||
// it, and a drain then would wait for a job that is only queued afterwards. So the count is taken after the present's
|
||||
// queue (vk_replay.cpp, after framegen_queue). They are made again at the next frame that needs them, as after a resize.
|
||||
// RX_VK_FG_KEEP=1: kept, as before.
|
||||
//
|
||||
// The rule alone, no Vulkan: tests/test_framegen_idle.py compiles it.
|
||||
#pragma once
|
||||
#include <cstdint>
|
||||
|
||||
namespace framegen_idle {
|
||||
|
||||
inline constexpr uint32_t kPresents = 120;
|
||||
|
||||
/// The presents in a row off, after this one: on now, or used at this present, starts again
|
||||
inline uint32_t next_off(uint32_t off, bool on_now, bool used) { return on_now || used ? 0 : off < UINT32_MAX ? off + 1 : off; }
|
||||
|
||||
/// Whether this present lets them go: there are some, and they have been off long enough
|
||||
inline bool due(uint32_t off, bool have) { return have && off >= kPresents; }
|
||||
|
||||
/// One present, as framegen_idle sees it after the present's queue
|
||||
struct Present {
|
||||
bool on; // frame generation on by the row, read again (framegen_on)
|
||||
bool recorded; // this present recorded made frames (framegen_record's true)
|
||||
bool hudless_wanted; // AMD FSR 4's frame generation is the row's (fsr4fg::frames_wanted)
|
||||
bool have_images; // the shared images are there (fg.w)
|
||||
bool have_hudless; // the images of the picture without the HUD are there
|
||||
};
|
||||
struct Counts {
|
||||
uint32_t off = 0, hudless_off = 0;
|
||||
};
|
||||
struct Release {
|
||||
bool shared, hudless;
|
||||
};
|
||||
/// The two counts after this present, and what it lets go. keep: RX_VK_FG_KEEP, nothing ever
|
||||
inline Release step(Counts& c, const Present& p, bool keep) {
|
||||
c.off = next_off(c.off, p.on, p.recorded);
|
||||
c.hudless_off = next_off(c.hudless_off, p.on && p.hudless_wanted, false);
|
||||
if (keep) return {false, false};
|
||||
return {due(c.off, p.have_images), due(c.hudless_off, p.have_hudless)};
|
||||
}
|
||||
|
||||
} // namespace framegen_idle
|
||||
40
gpu/vk/framegen_refresh.h
Normal file
40
gpu/vk/framegen_refresh.h
Normal file
@@ -0,0 +1,40 @@
|
||||
// (Triforce, 11-21, the upscaler audit's UPS-05) A fixed multiplier's frames, no more than the display can show.
|
||||
//
|
||||
// The Frame generation row's 2x to 8x make that many frames for each of the game's whatever the display's refresh rate
|
||||
// is; only the row's Auto goes by it. The owner's log of 11-20 has 5x on a 165 Hz display with the game near 66 frames a
|
||||
// second: 1,655 frames shown in every 5 s (331 a second, two for each refresh), each present call 2.2 to 3.0 ms where it
|
||||
// takes 0.03 ms without frame generation, and the display thread's line "0 frames shown on time, 1655 late". Half of the
|
||||
// made frames were never on the screen (the swapchain is mailbox: the newest at each refresh), and each was a pass over
|
||||
// the whole picture and a present the display thread could not keep up with.
|
||||
//
|
||||
// With the switch, a fixed multiplier makes no more frames than refreshes fall in one frame of the game: the game's
|
||||
// recent frame time over the display's period, rounded up, a tenth of a refresh forgiven so a game at a whole fraction
|
||||
// of the refresh rate is not given one more for its frames' jitter. Never more than the row asks, never fewer than the
|
||||
// game's own frame. The row's Auto and AMD FSR's own are not touched.
|
||||
//
|
||||
// RX_VK_FRAMEGEN_REFRESH=1: on. Off unless set (the row's number is the frames made, as before): it changes what a row
|
||||
// the player set shows, so it is the owner's to turn on.
|
||||
//
|
||||
// No Vulkan: tests/test_framegen_refresh.py compiles this.
|
||||
#pragma once
|
||||
#include <cstdint>
|
||||
#include <cstdlib>
|
||||
|
||||
namespace framegen_refresh {
|
||||
|
||||
/// RX_VK_FRAMEGEN_REFRESH (value, null when unset): on when it is not 0; unset, off
|
||||
inline bool on(const char* value) { return value && *value && std::atoi(value) != 0; }
|
||||
|
||||
/// The frames shown for each of the game's (the game's own among them) with a fixed multiplier m, the game's frames game_ns
|
||||
/// apart, the display refreshing every period_ns. An unknown frame time or period: m
|
||||
inline int frames(int m, int64_t game_ns, int64_t period_ns) {
|
||||
if (m <= 1 || game_ns <= 0 || period_ns <= 0) return m;
|
||||
// (the refreshes in a frame of the game, rounded up, a tenth of one forgiven: ceil(game / period - 0.1))
|
||||
const int64_t most = (game_ns * 10 + period_ns * 9 - 1) / (period_ns * 10);
|
||||
return most < 1 ? 1 : most < m ? (int)most : m;
|
||||
}
|
||||
|
||||
/// The display's period for a refresh rate in mHz (vk_main.cpp display_refresh_mhz); 0 when it is not known
|
||||
inline int64_t period_ns(int refresh_mhz) { return refresh_mhz > 0 ? (int64_t)(1000000000000LL / refresh_mhz) : 0; }
|
||||
|
||||
} // namespace framegen_refresh
|
||||
255
gpu/vk/framegen_rows.h
Normal file
255
gpu/vk/framegen_rows.h
Normal file
@@ -0,0 +1,255 @@
|
||||
/* (Triforce) The Insert menu's Frame generation row (vk_menu.cpp, opts.framegen) with its last choice, AMD FSR 4, and what
|
||||
AMD FSR 4's frame generation (vk_fsr4.cpp, fsr4fg) decides without a GPU: where the choice is offered, which of the
|
||||
versions AMD's library lists is taken, when a frame is a camera cut, when a context still fits and when it is let go,
|
||||
and the order it is taken down in. Plain functions, no Vulkan, no menu: tests/test_framegen_fsr4.py checks them. */
|
||||
#pragma once
|
||||
|
||||
#include <cmath>
|
||||
#include <cstdint>
|
||||
#include <cstdio>
|
||||
|
||||
namespace mkwvk::framegen_rows {
|
||||
|
||||
/// opts.framegen (saved as frame_generation): 0 Off, 1 Auto, 2 to 8 the multiplier, as before (an options file from before
|
||||
/// keeps its meaning), then 9: AMD FSR frame generation, one frame made by AMD's library between each two of the game's
|
||||
/// ((Triforce, FG-1) whichever version AMD's library has for the card: FSR 4's on an RX 9000, FSR 3.1's elsewhere. The
|
||||
/// name FSR4 stays: it is the value's, saved in options files as 9)
|
||||
enum { OFF = 0, AUTO = 1, kMaxMult = 8, FSR4 = 9, kRowMax = FSR4 };
|
||||
|
||||
/// ((Triforce, FG-9) RX_VK_FRAMEGEN_MULT=<0..9>, over the options file for a lab run: the row's whole range, 9 AMD's frame
|
||||
/// generation among it (it ran 2 to 8 alone before, and a 9 became the renderer's own 8x). Not a number: 2, as before
|
||||
inline int env_row(const char* e) {
|
||||
if (!e || !*e) return 2;
|
||||
int v = 0;
|
||||
if (std::sscanf(e, "%d", &v) != 1) return 2;
|
||||
return v < OFF ? (int)OFF : v > kRowMax ? (int)kRowMax : v;
|
||||
}
|
||||
|
||||
/// ((Triforce, FG-1) The row's value 9 by the version AMD's library makes the frames with: "AMD FSR (3.1.6)", "AMD FSR
|
||||
/// (4.0.2)"; "AMD FSR" before it is known. out holds n bytes
|
||||
inline const char* row_label(char* out, size_t n, const char* version) {
|
||||
if (!out || !n) return "";
|
||||
unsigned v[3] = {0, 0, 0};
|
||||
const bool known = version && std::sscanf(version, "%u.%u.%u", &v[0], &v[1], &v[2]) == 3;
|
||||
if (known) std::snprintf(out, n, "AMD FSR (%u.%u.%u)", v[0], v[1], v[2]);
|
||||
else std::snprintf(out, n, "AMD FSR");
|
||||
return out;
|
||||
}
|
||||
|
||||
/// Why AMD FSR 4 cannot make the frames here (the row says it); FG_OK when it can. FG_NOT_YET: not asked, the device is not
|
||||
/// made yet
|
||||
enum Why { FG_OK, FG_NOT_BUILT, FG_ENV_OFF, FG_NOT_YET, FG_NO_LIBRARY, FG_NO_FILE, FG_NO_MODEL, FG_FAILED, FG_WHY_COUNT };
|
||||
|
||||
/// ((Triforce) Why, in words, for the log line a run with the row at AMD FSR gets when AMD's library cannot make its frames)
|
||||
constexpr const char* why_name(Why w) {
|
||||
return w == FG_OK ? "it can" : w == FG_NOT_BUILT ? "not in this build (HOOKSHOT_FSR, Windows)"
|
||||
: w == FG_ENV_OFF ? "switched off (RX_VK_FSR4_FG=0 or RX_VK_FSR4=0)" : w == FG_NOT_YET ? "not asked yet (no device)"
|
||||
: w == FG_NO_LIBRARY ? "AMD's loader is not beside the game" : w == FG_NO_FILE ? "amd_fidelityfx_framegeneration_vk.dll is not beside it"
|
||||
: w == FG_NO_MODEL ? "AMD's library lists no frame generation it can take for this GPU"
|
||||
: w == FG_FAILED ? "it failed earlier in this run" : "?";
|
||||
}
|
||||
|
||||
/// The row's last value on this PC: AMD FSR 4 where it is offered, 8x elsewhere (the row as it was)
|
||||
constexpr int row_hi(bool fsr4_offered) { return fsr4_offered ? (int)FSR4 : (int)kMaxMult; }
|
||||
/// What the row shows, and what runs, for a value: itself, or 2x for a saved AMD FSR 4 on a PC where it is not offered
|
||||
/// (one made frame a frame either way, there by the renderer's own interpolation). The saved value stays for a PC where it is
|
||||
constexpr int shown(int v, bool fsr4_offered) { return v == FSR4 && !fsr4_offered ? 2 : v; }
|
||||
/// The frames shown for each of the game's: 1 at Off, auto_most with Auto (the display's refresh over 30 fps), the
|
||||
/// multiplier, and 2 with AMD FSR 4 (its library makes the one frame between two)
|
||||
constexpr int multiplier(int v, int auto_most) {
|
||||
return v == FSR4 ? 2 : v == AUTO ? auto_most : v < 1 ? 1 : v > kMaxMult ? (int)kMaxMult : v;
|
||||
}
|
||||
/// Whether AMD's library is asked for the made frame: the row at AMD FSR 4, where it can, and not beside NVIDIA DLSS frame
|
||||
/// generation, whose frames go in at the present itself (dlssg: vk_dlssg.cpp says it runs in this session)
|
||||
constexpr bool fsr4_asked(int v, Why why, bool dlssg) { return v == FSR4 && why == FG_OK && !dlssg; }
|
||||
|
||||
/// A version as AMD's library names it ("4.0.2"; one the graphics driver brings carries a star, "4.0.2*"): false when the
|
||||
/// name does not start with three numbers
|
||||
inline bool parse_version(const char* name, unsigned v[3]) {
|
||||
v[0] = v[1] = v[2] = 0;
|
||||
return name && std::sscanf(name, "%u.%u.%u", &v[0], &v[1], &v[2]) == 3;
|
||||
}
|
||||
/// Of the frame generation versions AMD's library lists for this GPU, the one contexts are made with: the newest, the first
|
||||
/// of equal ones, so FSR 4's (4.x) wherever it is listed (RX 9000), else FSR 3.1's (3.1.x: RX 6000 and 7000, whose driver
|
||||
/// lists 3.1.6 alone, NX2navi 5bdf583 on the owner's RX 7900 XT); -1 when it lists none it can read. ((Triforce, FG-1)
|
||||
/// Before, a list of FSR 3.1 alone gave -1 and the row never offered AMD's frame generation on those cards. The row names
|
||||
/// the version it runs, row_label.) only_fsr4 (RX_VK_FSR4_FG=4): 4.x or nothing, as it was
|
||||
inline int pick_version(const char* const* names, int n, bool only_fsr4) {
|
||||
int best = -1;
|
||||
unsigned bv[3] = {0, 0, 0};
|
||||
for (int i = 0; i < n; i++) {
|
||||
unsigned v[3];
|
||||
if (!parse_version(names[i], v) || (only_fsr4 && v[0] != 4)) continue;
|
||||
const bool newer = v[0] != bv[0] ? v[0] > bv[0] : v[1] != bv[1] ? v[1] > bv[1] : v[2] > bv[2];
|
||||
if (best < 0 || newer) best = i, bv[0] = v[0], bv[1] = v[1], bv[2] = v[2];
|
||||
}
|
||||
return best;
|
||||
}
|
||||
|
||||
/// The camera in the world, from the view matrix's three rows as the game has them (row major, view = R * world + t: the
|
||||
/// rows are its right, up and back axes, the last column the translation). The position is -R^T t
|
||||
struct Pose {
|
||||
float pos[3]{}, right[3]{}, up[3]{}, fwd[3]{};
|
||||
};
|
||||
inline Pose pose_of(const float v[12]) {
|
||||
Pose p;
|
||||
for (int i = 0; i < 3; i++) {
|
||||
p.right[i] = v[i], p.up[i] = v[4 + i], p.fwd[i] = -v[8 + i];
|
||||
p.pos[i] = -(v[i] * v[3] + v[4 + i] * v[7] + v[8 + i] * v[11]);
|
||||
}
|
||||
return p;
|
||||
}
|
||||
/// What counts as a camera cut: the turn (with a move) in degrees for each 60th of a second, and the least the time between
|
||||
/// the two frames is scaled to (in 60ths of a second; four at the most)
|
||||
struct CutRule {
|
||||
float turn_degrees;
|
||||
float min_scale;
|
||||
};
|
||||
/// Frame generation's (vk_fsr4.cpp fsr4fg): NX2navi's first thresholds (6e87c92), 10 degrees, never less than one 60th
|
||||
constexpr CutRule kFrameGenCut{10.0f, 1.0f};
|
||||
/// The upscalers' (FSR 3.1: vk_fsr3_inputs.cpp; DLSS and FSR 4: vk_temporal.cpp's chain): NX2navi's 2495391, 30 degrees.
|
||||
/// A quick swing of the camera around Link turns it 11 to 18 degrees and moves it 0.9 to 1.4 m a frame: at 10 degrees each
|
||||
/// frame of the swing started a new history, the scene at its render size for those frames, blocky at the smaller modes.
|
||||
/// ((Triforce) never less than one 60th, as frame generation's: NX2navi scaled from a quarter of one, but the time here is
|
||||
/// the CPU's between two recordings of the upscale, not the game's step, and the replay records frames back to back
|
||||
/// after a stall (a shader compile, an area load): a few ms apart, which at a quarter took a 1.25 m move, or a 7.5 degree
|
||||
/// turn with 0.125 m, for a cut, an ordinary swing among them. Above 60 fps a cut is the same move as at 60: an area load
|
||||
/// (5 to 180 m in a frame) or a jump of the camera still is one)
|
||||
constexpr CutRule kUpscalerCut{30.0f, 1.0f};
|
||||
|
||||
/// A camera cut between two frames frame_ms apart: the camera moved over 5 units, or turned over the rule's degrees while
|
||||
/// it moved over half a unit, for each 60th of a second (a slower frame may move in proportion, up to four times, a faster
|
||||
/// one down to the rule's least; the game's world is in metres, its near plane at 0.1). A made frame between the two
|
||||
/// pictures of a cut would blend two views that have nothing to do with each other: such a frame starts a new history and
|
||||
/// none is made; an upscaler's history from before a cut is no use after it. The thresholds are NX2navi's, from its camera
|
||||
/// traces of the same game (a median of 0.05 m and 0.08 degrees a frame in play, area loads of 5 to 180 m in one frame)
|
||||
inline bool camera_cut(const Pose& before, const Pose& now, float frame_ms, const CutRule& rule = kFrameGenCut) {
|
||||
const float least = rule.min_scale > 0.0f ? rule.min_scale : 1.0f;
|
||||
const float k = frame_ms / 16.667f;
|
||||
const float scale = !(frame_ms > 0.0f) ? 1.0f : k < least ? least : k > 4.0f ? 4.0f : k;
|
||||
float move2 = 0.0f, dot = 0.0f, la = 0.0f, lb = 0.0f;
|
||||
for (int i = 0; i < 3; i++) {
|
||||
const float d = now.pos[i] - before.pos[i];
|
||||
move2 += d * d, dot += before.fwd[i] * now.fwd[i], la += before.fwd[i] * before.fwd[i], lb += now.fwd[i] * now.fwd[i];
|
||||
}
|
||||
if (!std::isfinite(move2) || !std::isfinite(dot) || !(la > 0.0f) || !(lb > 0.0f)) return true; // (no camera to compare)
|
||||
const float move = std::sqrt(move2);
|
||||
const float cosine = dot / std::sqrt(la * lb);
|
||||
const float turn = std::acos(cosine < -1.0f ? -1.0f : cosine > 1.0f ? 1.0f : cosine); // (radians)
|
||||
const float turn_limit = rule.turn_degrees * 0.017453293f; // (radians)
|
||||
return move > 5.0f * scale || (turn > turn_limit * scale && move > 0.5f * scale);
|
||||
}
|
||||
|
||||
/// What a present with the row at AMD FSR 4 shows between the game's last frame and this one
|
||||
enum Made {
|
||||
MADE, ///< AMD's library made the frame
|
||||
OWN, ///< it did not take this present (no scene's depth and motion: a menu, a movie; not available; failed): the
|
||||
///< renderer's own interpolation makes the frame, as at 2x
|
||||
SKIP, ///< a new history (the first frame, a gap, a camera cut): AMD's library was given the frame and no frame is made
|
||||
/// ((Triforce, FG-2) the scene's depth and motion are there but AMD's library is not given the frame: no picture
|
||||
/// without the HUD this frame. The game's frame is shown alone: never a made frame of AMD's with the HUD moved in it
|
||||
/// (NX2navi 5bdf583's rule), and not the renderer's own in its place. (HDR Native and an RGBA16F scene, which hold for
|
||||
/// every present, are OWN: held_made)
|
||||
GAME
|
||||
};
|
||||
/// Whether a present's answer shows the game's frame alone, nothing made or blended before it (SKIP, GAME)
|
||||
constexpr bool game_frame_alone(Made m) { return m == SKIP || m == GAME; }
|
||||
|
||||
/// ((Triforce, FG-2) Why a present with the scene's depth and motion was not given to AMD's library, for its log.
|
||||
/// HELD_HDR_NATIVE: the game's own HDR (present_native_pq), whose picture without the HUD only the game could encode.
|
||||
/// HELD_HDR_SCENE: a run whose scene is RenoDX's RGBA16F (vk_temporal.cpp temporal_scene_format), whose upscale output is
|
||||
/// no picture without the HUD this renderer can encode as the presented one is. HELD_NO_HUDLESS: this frame has none
|
||||
enum Held { HELD_NONE, HELD_NO_HUDLESS, HELD_HDR_NATIVE, HELD_HDR_SCENE, HELD_COUNT };
|
||||
/// The reason, from what the present has: the game's own HDR first, then the run's RGBA16F scene (neither can have a
|
||||
/// picture without the HUD), then the picture without the HUD of this frame
|
||||
constexpr Held held_for(bool hdr_native, bool hdr_scene, bool have_hudless) {
|
||||
return hdr_native ? HELD_HDR_NATIVE : hdr_scene ? HELD_HDR_SCENE : !have_hudless ? HELD_NO_HUDLESS : HELD_NONE;
|
||||
}
|
||||
/// ((Triforce) Whether a reason holds for every present while it lasts (HDR Native, the RGBA16F scene), not for one frame
|
||||
/// that is missing its picture without the HUD. Then AMD's library is not asked at all (fsr4fg::frames_wanted: no inputs
|
||||
/// are made for it) and the renderer's own interpolation makes the one frame, as at 2x (OWN), which is what a saved 9 ran
|
||||
/// as on an RX 6000 or 7000 before AMD's FSR 3.1 frame generation was offered there. Only a frame missing its picture
|
||||
/// without the HUD shows the game's frame alone (GAME)
|
||||
constexpr bool held_whole_run(Held h) { return h == HELD_HDR_NATIVE || h == HELD_HDR_SCENE; }
|
||||
/// What a held present shows: the renderer's own frame for a reason that lasts, the game's frame alone for a missing picture
|
||||
constexpr Made held_made(Held h) { return held_whole_run(h) ? OWN : GAME; }
|
||||
|
||||
/// ((Triforce, FG-2) The picture without the HUD is taken at the gameplay upscale's output (its size ow x oh), the present
|
||||
/// is the window texture (sw x sh) cropped to x, y, w, h (vk_present_rect.h present_source_rect). The part of the output that
|
||||
/// is the present's crop, scaled when the two sizes differ (the handheld size's scaled crop, NX2navi 81cdd67), in whole
|
||||
/// pixels inside the output; all zero when there is nothing to take
|
||||
struct SourceRect {
|
||||
int32_t x0 = 0, y0 = 0, x1 = 0, y1 = 0;
|
||||
};
|
||||
inline SourceRect hudless_source(uint32_t ow, uint32_t oh, uint32_t sw, uint32_t sh, uint32_t x, uint32_t y, uint32_t w,
|
||||
uint32_t h) {
|
||||
SourceRect r;
|
||||
if (!ow || !oh || !sw || !sh || !w || !h || x >= sw || y >= sh) return r;
|
||||
const uint64_t rx = x + w > sw ? sw : (uint64_t)x + w, ry = y + h > sh ? sh : (uint64_t)y + h;
|
||||
r.x0 = (int32_t)((uint64_t)x * ow / sw), r.y0 = (int32_t)((uint64_t)y * oh / sh);
|
||||
r.x1 = (int32_t)(rx * ow / sw), r.y1 = (int32_t)(ry * oh / sh);
|
||||
if (r.x1 <= r.x0 || r.y1 <= r.y0) r = SourceRect{};
|
||||
return r;
|
||||
}
|
||||
|
||||
/// A frame generation context as it was made: the presented picture's size, the largest scene it takes, and whether the
|
||||
/// scene's far plane is at infinity (a context is made for one kind of depth)
|
||||
struct ContextKey {
|
||||
uint32_t width = 0, height = 0, max_render_width = 0, max_render_height = 0;
|
||||
bool infinite_depth = false;
|
||||
};
|
||||
/// Whether a context made as c takes a frame of w x h with a scene of rw x rh: else it is made again
|
||||
constexpr bool context_fits(const ContextKey& c, uint32_t w, uint32_t h, uint32_t rw, uint32_t rh, bool infinite_depth) {
|
||||
return c.width == w && c.height == h && rw <= c.max_render_width && rh <= c.max_render_height &&
|
||||
c.infinite_depth == infinite_depth;
|
||||
}
|
||||
/// The context for such a frame: the scene is never larger than the picture in play, and a larger one still gets a context
|
||||
constexpr ContextKey context_for(uint32_t w, uint32_t h, uint32_t rw, uint32_t rh, bool infinite_depth) {
|
||||
return ContextKey{w, h, rw > w ? rw : w, rh > h ? rh : h, infinite_depth};
|
||||
}
|
||||
|
||||
/// The presents a context is kept after its last frame: 120 once the row has left AMD FSR 4 (stepping through the row's
|
||||
/// choices passes it: nothing is made and let go again within two seconds), 2 after a failure. Never at the present of
|
||||
/// its last frame: that frame's commands are recorded and not yet on the queue
|
||||
constexpr uint64_t kIdlePresents = 120, kFailedPresents = 2;
|
||||
constexpr bool release_due(bool have, bool failed, uint64_t presents, uint64_t used_at) {
|
||||
return have && presents >= used_at && presents - used_at >= (failed ? kFailedPresents : kIdlePresents);
|
||||
}
|
||||
|
||||
/// Taking AMD's frame generation down, in the order that is safe, as one function the real code (vk_fsr4.cpp) and
|
||||
/// tests/test_framegen_fsr4.py both run. ops.stop: no call into AMD's library starts after it (the lock a call holds is
|
||||
/// taken, so one in progress has returned). At the process's exit (HOS_OPT_EXITING, host/hos_relaunch.c) that is all:
|
||||
/// there is no swapchain or thread of AMD's to stop (the renderer presents the made frames itself), and what the context
|
||||
/// holds goes with the process, as the upscaler's does; AMD's library is not called again, so nothing of it runs under
|
||||
/// ExitProcess. In play (the row left AMD FSR 4, another size, a failure): the GPU is past every command buffer that
|
||||
/// used the context (ops.wait_gpu), then AMD's context goes (ops.destroy_context) before the image it wrote into
|
||||
/// (ops.destroy_images), never the other way round
|
||||
template <class Ops> void teardown(Ops& ops, bool exiting) {
|
||||
ops.stop();
|
||||
if (exiting) return;
|
||||
ops.wait_gpu();
|
||||
ops.destroy_context();
|
||||
ops.destroy_images();
|
||||
}
|
||||
|
||||
// ---- (Triforce, 11-20) What a present writes into a set of frames: never past the frames the set was made with ----
|
||||
// The owner's "upscaling crash": 0xc0000005 reading 0x590 in NVIDIA's driver, inside a vkCmdCopyImage the recorder thread
|
||||
// ran, in three runs of 11-20, each with frame interpolation on and the Frame generation row stepped up in the options
|
||||
// (Auto in the first, 3x at the next start, 5x at the one after). The sets' frames are made for the row's multiplier when
|
||||
// a present begins (vk_framegen.cpp ensure_size); the present then waits for the display thread to let a set go, and
|
||||
// while frames are made the display thread is the one that runs the menu (present_image, menu_tick), so a step of the row
|
||||
// lands in that wait. The multiplier read again after it was more than the frames made, and the game's frame was copied
|
||||
// into an image that was never made (VK_NULL_HANDLE). A step down only left frames unused. The row's new multiplier takes
|
||||
// effect at the next present, which makes the sets anew.
|
||||
/// The multiplier a present records with: the row's, and no more than the frames its sets were made with
|
||||
constexpr int set_multiplier(int row_multiplier, int made_with) {
|
||||
const int most = made_with < 1 ? 1 : made_with;
|
||||
return row_multiplier < 1 ? 1 : row_multiplier > most ? most : row_multiplier;
|
||||
}
|
||||
/// The game's own frame's place in a set, after `made` made ones: the set's last frame at most
|
||||
constexpr int game_slot(int made, int made_with) {
|
||||
const int last = (made_with < 1 ? 1 : made_with) - 1;
|
||||
return made < 0 ? 0 : made > last ? last : made;
|
||||
}
|
||||
|
||||
} // namespace mkwvk::framegen_rows
|
||||
@@ -56,4 +56,44 @@ inline float view_distance(const Camera& c, float device_depth) {
|
||||
return c.device_to_view_depth[1] / (device_depth - c.device_to_view_depth[0]);
|
||||
}
|
||||
|
||||
/// The jitter sequence's length for FSR 3.1, AMD's own rule (ffxFsr3GetJitterPhaseCount, ffx_upscale.h): 8 times the
|
||||
/// width's scale squared, truncated, with no upper bound: 18 at Quality (1.5x), 23 at Balanced, 32 at Performance, 71 or 72
|
||||
/// at Ultra Performance (3x), 8 for the Native AA. AMD FSR 4 on the frame chain takes it too ((Triforce, FSR4-10)
|
||||
/// vk_temporal.cpp chain); NVIDIA's rule (dlss_math.h dlss_jitter_phases, by area, rounded, 64 at most) stays DLSS's
|
||||
inline unsigned jitter_phases(unsigned render_width, unsigned display_width) {
|
||||
if (!render_width || !display_width) return 8;
|
||||
const float ratio = (float)display_width / (float)render_width;
|
||||
const int n = (int)(8.0f * ratio * ratio);
|
||||
return n < 1 ? 1u : (unsigned)n;
|
||||
}
|
||||
|
||||
/// The size FSR 3.1's render-size images are made at (the SDK's maxRenderSize), and whether they are made again, for a frame
|
||||
/// of rw x rh into uw x uh with images of have_w x have_h for an output of have_uw x have_uh (0 x 0: none yet). They are kept
|
||||
/// for any frame that fits them, so the game's dynamic resolution (the Render resolution row's Auto) changes only the
|
||||
/// render size each dispatch is given and the history goes on, as the SDK's FFX_FSR3UPSCALER_ENABLE_DYNAMIC_RESOLUTION
|
||||
/// does: its passes clamp what they read to the frame's render size (ClampUv, ClampLoad). Made again only for another
|
||||
/// output, or for a frame larger than them: at the output's size then (a size that grew is the dynamic resolution's, and
|
||||
/// the images are not made again for its next step), and at once with dynamic (the row's Auto). A fixed share is made at
|
||||
/// its own size. Never smaller than the frame, nor larger than the output unless the frame is
|
||||
struct ImagePlan {
|
||||
unsigned width = 0, height = 0;
|
||||
bool remake = false;
|
||||
};
|
||||
inline ImagePlan plan_images(unsigned have_w, unsigned have_h, unsigned have_uw, unsigned have_uh, unsigned rw, unsigned rh,
|
||||
unsigned uw, unsigned uh, bool dynamic) {
|
||||
ImagePlan p;
|
||||
rw = rw ? rw : 1, rh = rh ? rh : 1, uw = uw ? uw : 1, uh = uh ? uh : 1;
|
||||
const bool same_output = have_w && have_h && have_uw == uw && have_uh == uh;
|
||||
if (same_output && rw <= have_w && rh <= have_h) {
|
||||
p.width = have_w, p.height = have_h;
|
||||
return p;
|
||||
}
|
||||
// (a larger frame for the same output: the dynamic resolution's step up)
|
||||
const bool full = dynamic || same_output;
|
||||
p.width = full && uw > rw ? uw : rw;
|
||||
p.height = full && uh > rh ? uh : rh;
|
||||
p.remake = true;
|
||||
return p;
|
||||
}
|
||||
|
||||
} // namespace mkwvk::fsr3_math
|
||||
|
||||
148
gpu/vk/game_grid.h
Normal file
148
gpu/vk/game_grid.h
Normal file
@@ -0,0 +1,148 @@
|
||||
// (Triforce, 11-19) Where the game's 2D layouts are in the window: the 3840 x 2160 grid its menus and lists are laid out on
|
||||
// (the units host/oot_menu.c tells the Mods entry's list in), as the window shows it. The lab's "Title in a 16:9 window (1600x900,
|
||||
// 21:9 picture letterboxed): Mods drawn below the picture, larger, far from the list": the entry was drawn on a 16:9 column of
|
||||
// the window's whole height, which is the grid only when the picture fills the window's height.
|
||||
// - The picture: in the window between black bars, as the present puts it (vk_present_rect.h present_rect, with the
|
||||
// picture's shape: Aspect ratio's 16:10, 21:9 or 32:9, else its pixels').
|
||||
// - The grid in it: the game's layouts keep their 16:9 shape in the picture's middle on a wider picture (host/oot_wide_hud.h:
|
||||
// every layout but the kept backgrounds and fades is squeezed by 16:9 over the picture's ratio, host/hos_main.c
|
||||
// hos_wide_squeeze), its height the picture's; on a narrower one (16:10) the grid is the whole picture, narrower than 16:9
|
||||
// (the layouts squeezed across, not cut). HUD position moves panes of the HUD's own layouts only (oot_hud_out), not
|
||||
// the lists'.
|
||||
// No ImGui and no Vulkan: tests/test_mods_grid.py compiles it alone.
|
||||
#pragma once
|
||||
#include <algorithm>
|
||||
#include <cmath>
|
||||
#include <cstdint>
|
||||
#include "vk_present_rect.h"
|
||||
|
||||
namespace mkwvk::game_grid {
|
||||
|
||||
/// The grid's rectangle in the window's pixels: its left and top, its width and height (2160 units down, 3840 across)
|
||||
struct Place {
|
||||
float x = 0, y = 0, w = 0, h = 0;
|
||||
bool picture = false; // ((11-19) from a picture the present showed; false: the window's column, none known yet)
|
||||
/// ((Triforce, 11-20) the picture itself in the window (the present's rectangle, whole pixels), with picture: the game
|
||||
/// draws nothing outside it)
|
||||
float pic_x = 0, pic_y = 0, pic_w = 0, pic_h = 0;
|
||||
};
|
||||
|
||||
/// The game's picture (picture_w x picture_h, 0 not known yet) shown in a window_w x window_h window, aspect the picture's
|
||||
/// shape (0: its pixels'). With no picture known, the window's middle 16:9 column of its whole height, as before
|
||||
inline Place of(uint32_t picture_w, uint32_t picture_h, uint32_t window_w, uint32_t window_h, double aspect) {
|
||||
Place g;
|
||||
if (!window_w || !window_h) return g;
|
||||
if (!picture_w || !picture_h) {
|
||||
g.h = (float)window_h;
|
||||
g.w = std::min((float)window_w, std::floor((float)window_h * 16.0f / 9.0f + 0.5f));
|
||||
g.x = std::floor(((float)window_w - g.w) * 0.5f);
|
||||
return g;
|
||||
}
|
||||
const PresentRect p = present_rect(picture_w, picture_h, nullptr, window_w, window_h, PresentAspect{aspect});
|
||||
const double ratio = (double)p.w / (double)p.h, squeeze = std::min(1.0, (16.0 / 9.0) / ratio);
|
||||
g.h = (float)p.h, g.y = (float)p.y, g.picture = true;
|
||||
g.pic_x = (float)p.x, g.pic_y = (float)p.y, g.pic_w = (float)p.w, g.pic_h = (float)p.h;
|
||||
g.w = (float)((double)p.w * squeeze);
|
||||
g.x = (float)p.x + ((float)p.w - g.w) * 0.5f;
|
||||
return g;
|
||||
}
|
||||
|
||||
/// A drawing made on a 16:9 column of the grid's height (its own pixels, from 0: w = h x 16 / 9 wide, unit h / 720) put on
|
||||
/// the grid: x across times kx (1 on a 16:9 grid; under 1 where the layouts are squeezed across), then moved by ox, oy
|
||||
struct Map {
|
||||
float ox = 0, oy = 0, kx = 1;
|
||||
float x(float px) const { return ox + px * kx; }
|
||||
float y(float py) const { return oy + py; }
|
||||
/// (the window's pixel back in the drawing's: the mouse)
|
||||
float back_x(float wx) const { return (wx - ox) / kx; }
|
||||
float back_y(float wy) const { return wy - oy; }
|
||||
};
|
||||
inline float column_w(const Place& g) { return std::floor(g.h * 16.0f / 9.0f + 0.5f); }
|
||||
/// ((Triforce, 11-19) on a 16:9 grid (a wider picture's middle, or a 16:9 one) the column is put on whole pixels, not
|
||||
/// stretched: the text drawn on it is on whole pixels already (ImGui's glyphs), and a part of a pixel across or a scale a
|
||||
/// little off 1 (the grid's width from the picture's, the column's rounded) left it between pixels, soft. Only a grid
|
||||
/// narrower than 16:9 (16:10) is squeezed across, as the game squeezes its layouts there)
|
||||
inline Map map_of(const Place& g) {
|
||||
Map m;
|
||||
const float cw = column_w(g);
|
||||
if (cw > 0 && g.w >= cw - 1.0f) {
|
||||
m.ox = std::round(g.x + (g.w - cw) * 0.5f), m.oy = std::round(g.y), m.kx = 1.0f;
|
||||
return m;
|
||||
}
|
||||
m.ox = std::round(g.x), m.oy = std::round(g.y), m.kx = cw > 0 ? g.w / cw : 1.0f;
|
||||
return m;
|
||||
}
|
||||
|
||||
/// A page's 16:9 column (its own pixels from 0: w x h, unit h / 720) and where it goes in the window (map)
|
||||
struct Column {
|
||||
Map map;
|
||||
float w = 0, h = 0;
|
||||
/// ((Triforce, 11-20) over the game's panes: the picture's right edge in the column's own pixels (past w where the
|
||||
/// picture is wider than the grid); 0 otherwise
|
||||
float game_x1 = 0;
|
||||
};
|
||||
/// ((Triforce, 11-19) the column a page of ours is drawn on in a window_w x window_h window. on_game: over the game's own
|
||||
/// panes (its Options screen under our tabs and their header, the Mods entry's list): on the game's grid, as those panes
|
||||
/// are, so that our plates cover them; the lab's "16:9 window, 21:9 picture letterboxed: the game's Default Settings and
|
||||
/// Back and the left edge of its help box show through beside our panel" was our tabs on the window's column of its whole
|
||||
/// height while the game's panes were on the picture's grid. Else (our options over play), and with no picture known yet,
|
||||
/// the window's middle 16:9 column of its whole height, as before (narrower windows: the whole window, not squeezed)
|
||||
inline Column column_of(const Place& grid, float window_w, float window_h, bool on_game) {
|
||||
Column c;
|
||||
if (on_game && grid.picture) {
|
||||
c.map = map_of(grid), c.w = column_w(grid), c.h = grid.h;
|
||||
c.game_x1 = c.map.back_x(grid.pic_x + grid.pic_w);
|
||||
return c;
|
||||
}
|
||||
c.w = std::min(window_w, std::floor(window_h * 16.0f / 9.0f + 0.5f)), c.h = window_h;
|
||||
c.map.ox = std::floor((window_w - c.w) * 0.5f);
|
||||
return c;
|
||||
}
|
||||
|
||||
/// ((Triforce, 11-20) A rectangle of the window, in its pixels
|
||||
struct Rect {
|
||||
float x0 = 0, y0 = 0, x1 = 0, y1 = 0;
|
||||
};
|
||||
/// ((Triforce, 11-20) A dim over a whole window_w x window_h window cut where the game's picture ends (g, from of): out[0]
|
||||
/// the picture, where a dim keeps its colour, then the bars beside, above or below it (up to four, none overlapping, all
|
||||
/// of them and the picture the window exactly), where a dim must be black. The lab's check 3 of 26670c0: our dialogs' dim
|
||||
/// (rgba(8, 5, 2, ...), lighter than black) turned the black bars a dark brown, luminance about 9. Black at the same alpha
|
||||
/// never lightens them; over the picture the dim stays as it was. With no picture known: the whole window, one part.
|
||||
/// Returns how many
|
||||
inline int dim_parts(const Place& g, float window_w, float window_h, Rect out[5]) {
|
||||
out[0] = {0, 0, window_w, window_h};
|
||||
if (!g.picture) return 1;
|
||||
const float x0 = std::clamp(g.pic_x, 0.0f, window_w), x1 = std::clamp(g.pic_x + g.pic_w, x0, window_w);
|
||||
const float y0 = std::clamp(g.pic_y, 0.0f, window_h), y1 = std::clamp(g.pic_y + g.pic_h, y0, window_h);
|
||||
out[0] = {x0, y0, x1, y1};
|
||||
int n = 1;
|
||||
if (y0 > 0) out[n++] = {0, 0, window_w, y0}; // (above)
|
||||
if (y1 < window_h) out[n++] = {0, y1, window_w, window_h}; // (below)
|
||||
if (x0 > 0) out[n++] = {0, y0, x0, y1}; // (left, between those)
|
||||
if (x1 < window_w) out[n++] = {x1, y0, window_w, y1}; // (right)
|
||||
return n;
|
||||
}
|
||||
|
||||
/// ((Triforce, 11-20) The game's guides band at the foot of its Options screen (SystemOverlay_00's N_Footer_00: W_GuideBg_00
|
||||
/// and W_GuideBg_01, right-aligned 2110 and 2174 units right of the middle, 160 high from 19 under the foot) goes on past the
|
||||
/// 16:9 column's right edge (1920): on a picture wider than the grid its end shows beside the column. The lab's check 3 of
|
||||
/// 26670c0 (the game's own tab: 16x9_10, 21x9_10, fs_10): the band seen to 238 to 240 units past the column's edge (the
|
||||
/// window pane's last texels clear), from 126 units over the foot, and our guides' plate ending at the column's edge left
|
||||
/// 75 px of it in the 1600 x 900 window and 155 px in fullscreen 3440 x 1440. Where the band ends, in units right of the middle
|
||||
constexpr float kGuideBandRight = 2160;
|
||||
/// ((Triforce, 11-20, a review of 1104352) The Options screen chosen from the pause menu (host/oot_menu.c tells it as 2): its
|
||||
/// footer may be the pause menu's own (MenuOverlay_00's band panes end 2110 units right of the middle; in the pause menu's
|
||||
/// fs_30_sys_on_title_screen.png its band fades out by about 2100), not SystemOverlay_00's to 2160. No lab shot of that screen
|
||||
/// yet: our plate ends at the smaller of the two there, so that it never runs about 50 units (33 px in fullscreen 3440 x 1440)
|
||||
/// past the game's band as a dark sliver over the bars' black
|
||||
constexpr float kPauseGuideBandRight = 2110;
|
||||
/// (where our guides' plate ends over the game's panes on column c, in its own pixels from 0: as far as the game's band
|
||||
/// (band_right units right of the middle: kGuideBandRight, the title's Options; kPauseGuideBandRight, the pause menu's), up to
|
||||
/// the picture's right edge (the game draws nothing past it); the column's right edge on a 16:9 or narrower picture (16:10:
|
||||
/// the grid is the whole picture), over play, or with no picture known)
|
||||
inline float guide_band_x1(const Column& c, float band_right = kGuideBandRight) {
|
||||
const float band = c.w * 0.5f + band_right * c.h / 2160.0f;
|
||||
return std::max(c.w, std::min(band, c.game_x1));
|
||||
}
|
||||
|
||||
} // namespace mkwvk::game_grid
|
||||
132
gpu/vk/game_language.h
Normal file
132
gpu/vk/game_language.h
Normal file
@@ -0,0 +1,132 @@
|
||||
/* The Insert menu's Game language row (vk_menu.cpp), as plain code: what language.txt in the game folder chooses and offers,
|
||||
and the file's text with another language chosen. No Vulkan, no menu: tests/test_game_language.py checks it.
|
||||
|
||||
The game's language is the host's to read, at the start (host/nn_applet.c desired_language): HOS_LANGUAGE, else the
|
||||
file's first line that is not blank or a comment and whose first word is one of the game's codes, in any case; English
|
||||
(US) when there is none. Setup's update keeps a folder's choice by the same line, and takes it only when the line is the
|
||||
code alone, spelled as the game spells it (nx2/make_game.py folder_language). The file lists the codes in its comments
|
||||
("# pt-BR Portuguese (Brazil)"), as the host and Setup write it.
|
||||
|
||||
The row offers the languages the file itself names, as its choice or in that list, that the game has; it writes the
|
||||
choice as the code alone on the choice's line, and no code the file does not name. */
|
||||
#pragma once
|
||||
|
||||
#include <cstddef>
|
||||
#include <string>
|
||||
#include <string_view>
|
||||
|
||||
namespace mkwvk::game_language {
|
||||
|
||||
struct Language {
|
||||
const char* code;
|
||||
const char* name;
|
||||
};
|
||||
/// The game's languages: host/nn_applet.c's table and nx2/make_game.py's LANGUAGES, in their order (the test holds the
|
||||
/// three together)
|
||||
constexpr Language kLanguages[] = {
|
||||
{"en-US", "English (US)"}, {"en-GB", "English (UK)"}, {"fr", "French (France)"}, {"fr-CA", "French (Canada)"},
|
||||
{"de", "German"}, {"it", "Italian"}, {"es", "Spanish (Spain)"}, {"es-419", "Spanish (Latin America)"},
|
||||
{"pt-BR", "Portuguese (Brazil)"}, {"nl", "Dutch"}, {"pl", "Polish"}, {"ru", "Russian"},
|
||||
{"ja", "Japanese"}, {"ko", "Korean"}, {"zh-Hans", "Chinese (Simplified)"}, {"zh-Hant", "Chinese (Traditional)"},
|
||||
};
|
||||
constexpr int kCount = sizeof kLanguages / sizeof kLanguages[0];
|
||||
constexpr int kDefault = 0; // (English (US): the host's answer when the file chooses nothing)
|
||||
inline unsigned bit(int language) { return 1u << language; }
|
||||
|
||||
/// A code's place in kLanguages, in any case (the host compares so: "pt-br" is pt-BR); -1 when the game has no such code
|
||||
inline int find(std::string_view code) {
|
||||
for (int i = 0; i < kCount; i++) {
|
||||
const std::string_view k = kLanguages[i].code;
|
||||
if (k.size() != code.size()) continue;
|
||||
bool same = true;
|
||||
for (size_t c = 0; c < k.size() && same; c++) {
|
||||
const auto lower = [](char x) { return x >= 'A' && x <= 'Z' ? (char)(x - 'A' + 'a') : x; };
|
||||
same = lower(k[c]) == lower(code[c]);
|
||||
}
|
||||
if (same) return i;
|
||||
}
|
||||
return -1;
|
||||
}
|
||||
|
||||
/// What a language.txt says
|
||||
struct File {
|
||||
int chosen = -1; ///< the language its choice line names (kLanguages' index); -1: it has no such line
|
||||
unsigned offered = 0; ///< a bit a language the file names: its choice, and each "# <code> ..." line
|
||||
size_t at = 0, len = 0; ///< the choice line's text to replace: after a byte order mark, without its line ending
|
||||
bool crlf = false; ///< its lines end with CR LF
|
||||
};
|
||||
|
||||
/// The first word of s (after spaces and tabs), up to a space, a tab or a line's end
|
||||
inline std::string_view word(std::string_view s) {
|
||||
size_t a = 0;
|
||||
while (a < s.size() && (s[a] == ' ' || s[a] == '\t')) a++;
|
||||
size_t b = a;
|
||||
while (b < s.size() && s[b] != ' ' && s[b] != '\t' && s[b] != '\r' && s[b] != '\n') b++;
|
||||
return s.substr(a, b - a);
|
||||
}
|
||||
|
||||
/// The file as the host reads it, line by line: what a line longer than its 79 bytes holds past them is not a line's start,
|
||||
/// a byte order mark may open any line, and the choice is the first line that is neither blank nor a comment and names a
|
||||
/// language of the game (a line that names none is passed over, as there)
|
||||
inline File read(std::string_view text) {
|
||||
File f;
|
||||
f.crlf = text.find("\r\n") != std::string_view::npos;
|
||||
for (size_t start = 0; start < text.size();) {
|
||||
size_t end = text.find('\n', start);
|
||||
if (end == std::string_view::npos) end = text.size();
|
||||
size_t stop = end; // (the line without its ending)
|
||||
if (stop > start && text[stop - 1] == '\r') stop--;
|
||||
std::string_view line = text.substr(start, stop - start);
|
||||
size_t lead = 0; // (bytes of the line before what the host looks at)
|
||||
if (line.size() >= 3 && (unsigned char)line[0] == 0xEF && (unsigned char)line[1] == 0xBB && (unsigned char)line[2] == 0xBF) lead = 3;
|
||||
// (the host's buffer: 79 bytes of the line, cut at a zero byte. It keeps 31 bytes of the word, more than any code has: a
|
||||
// longer word names no language there or here)
|
||||
std::string_view seen = line.substr(0, 79);
|
||||
if (const size_t zero = seen.find('\0'); zero != std::string_view::npos) seen = seen.substr(0, zero);
|
||||
seen = seen.substr(lead < seen.size() ? lead : seen.size());
|
||||
size_t blank = 0;
|
||||
while (blank < seen.size() && (seen[blank] == ' ' || seen[blank] == '\t')) blank++;
|
||||
const std::string_view rest = seen.substr(blank);
|
||||
if (!rest.empty() && rest[0] == '#') { // (a comment: a language it lists, "# <code> <name>")
|
||||
const int listed = find(word(rest.substr(1)));
|
||||
if (listed >= 0) f.offered |= bit(listed);
|
||||
} else if (!rest.empty() && f.chosen < 0) {
|
||||
const int named = find(word(rest));
|
||||
if (named >= 0) f.chosen = named, f.offered |= bit(named), f.at = start + lead, f.len = stop - (start + lead);
|
||||
}
|
||||
start = end + 1;
|
||||
}
|
||||
return f;
|
||||
}
|
||||
|
||||
/// The file's text with `language` chosen: its choice line replaced by the code alone (as Setup's update reads it), every
|
||||
/// other byte as it was; with no choice line, the code on a line of its own at the end. The text as it is when the file
|
||||
/// does not name that language: never a code it does not list
|
||||
inline std::string write(std::string_view text, int language) {
|
||||
const File f = read(text);
|
||||
if (language < 0 || language >= kCount || !(f.offered & bit(language))) return std::string(text);
|
||||
const std::string_view code = kLanguages[language].code;
|
||||
std::string out(text);
|
||||
if (f.chosen >= 0) {
|
||||
out.replace(f.at, f.len, code);
|
||||
return out;
|
||||
}
|
||||
const char* const eol = f.crlf ? "\r\n" : "\n";
|
||||
if (!out.empty() && out.back() != '\n') out += eol;
|
||||
out += code;
|
||||
out += eol;
|
||||
return out;
|
||||
}
|
||||
|
||||
/// The languages a file offers, in kLanguages' order, into out (kCount of room): how many. None unless the language in
|
||||
/// effect (the file's choice, else the host's English (US)) is among them: the row then has nothing sound to show
|
||||
inline int offered(const File& f, int* out) {
|
||||
const int now = f.chosen >= 0 ? f.chosen : kDefault;
|
||||
if (!(f.offered & bit(now))) return 0;
|
||||
int n = 0;
|
||||
for (int i = 0; i < kCount; i++)
|
||||
if (f.offered & bit(i)) out[n++] = i;
|
||||
return n;
|
||||
}
|
||||
|
||||
} // namespace mkwvk::game_language
|
||||
73
gpu/vk/gate_wait.h
Normal file
73
gpu/vk/gate_wait.h
Normal file
@@ -0,0 +1,73 @@
|
||||
// (Triforce) A submission whose command buffer begins with a wait for an event the game signals from the CPU: the submit waits for
|
||||
// the signal (RX_VK_GATE_WAIT_MS at most) before its capture and its replay, after NX2navi's gate_wait_ms (runtime/vk_exec.c
|
||||
// vk_exec_gate_wait and gate_first_wait).
|
||||
//
|
||||
// OoT's queues order some of their work with events the CPU signals once it has written what the work reads. A submission that
|
||||
// meets such a wait unmet is taken (its capture of the game's memory at the submit) and replayed up to the wait, where it stops; once
|
||||
// the CPU signals, the order thread (vk_replay.cpp sync_watch, a millisecond at a time) or the next submit or present lets it go on,
|
||||
// with a second capture (resume_queue, RX_VK_RESUME_CAPTURE). NX2navi found the signal 1 to 2 ms after the submit call in about
|
||||
// every other frame in Hyrule Field, and made the submit call wait for it, before its copies, as the console's GPU waits at the
|
||||
// wait itself: the capture is then taken after the game has written the frame's data, and the frame is replayed in the submit call
|
||||
// and on the submitting thread (54.5 fps instead of 53 there, the longest frame 33 ms instead of 46). Triforce meets it far less
|
||||
// often (the "captures at a CPU event's wait" of the order line: 100 to 170 in a lab run of 26,000 frames at the Market), and a gate
|
||||
// that runs out holds the submitting thread for the whole bound, so it is off unless asked for, for the lab to measure.
|
||||
// The gate waits only where the replay would stop: the first op of the submission's first command buffer that is not state the
|
||||
// replay only stores (bound programs, buffers, textures, targets, viewports...) is a wait for an event, the wait is not met in the
|
||||
// replay's order now, and the CPU has signalled that event before (vk_replay.cpp cpu_events). It takes no lock while it waits (the
|
||||
// thread that will signal may need the order lock first), and looks at the event's word in the game's memory, which the CPU's signal
|
||||
// writes. It ends at the signal, or at the bound, and the submission then goes as before. After kOuts waits that ran out, once they
|
||||
// are more than a tenth of the waits, it is off for the run (the signal is then one that comes only after the submitting thread goes
|
||||
// on, and every wait would be a stall).
|
||||
// RX_VK_GATE_WAIT_MS unset or 0: off, as before; 1 to 100: the bound in milliseconds (NX2navi's default is 10)
|
||||
// Apart from Vulkan and the stream's op numbers so that tests/test_gate_wait.py can run the rules and the gate over authored streams.
|
||||
#pragma once
|
||||
|
||||
#include <cstdint>
|
||||
#include <cstdlib>
|
||||
#include <cstring>
|
||||
#if defined(__x86_64__) || defined(_M_X64) || defined(__i386__) || defined(_M_IX86)
|
||||
#include <immintrin.h>
|
||||
#endif
|
||||
|
||||
namespace mkwvk::gate_wait {
|
||||
|
||||
constexpr int kMaxOps = 256; // (the ops looked at before the wait, at most: NX2navi's)
|
||||
constexpr int64_t kOuts = 4; // (waits that ran out before the gate may go off for the run)
|
||||
|
||||
/// RX_VK_GATE_WAIT_MS: the bound in milliseconds, 0 (off) to 100; unset or not a number: 0
|
||||
inline int ms(const char* env) {
|
||||
if (!env || !*env) return 0;
|
||||
const int v = std::atoi(env);
|
||||
return v < 0 ? 0 : v > 100 ? 100 : v;
|
||||
}
|
||||
|
||||
/// The command buffer's first op that is not state (records {uint32 op, uint32 size}, sizes 8-aligned, from p): the wait it begins
|
||||
/// with, or null. kind(op): 1 state the replay only stores (looked past), 2 a wait for an event, 0 anything else (a draw, a copy, a
|
||||
/// signal, a barrier, the end: the gate does not wait)
|
||||
template <class K>
|
||||
const uint8_t* first_wait(const uint8_t* p, K kind) {
|
||||
for (int k = 0; k < kMaxOps; k++) {
|
||||
uint32_t h[2];
|
||||
std::memcpy(h, p, sizeof h);
|
||||
if (h[1] < 8 || (h[1] & 7)) return nullptr;
|
||||
const int what = kind(h[0]);
|
||||
if (what == 2) return p;
|
||||
if (what != 1) return nullptr;
|
||||
p += h[1];
|
||||
}
|
||||
return nullptr;
|
||||
}
|
||||
|
||||
/// Whether the gate goes off for the run, after `gates` waits of which `outs` ran out
|
||||
inline bool off_after(int64_t gates, int64_t outs) {
|
||||
return outs >= kOuts && outs * 10 > gates;
|
||||
}
|
||||
|
||||
/// A pause of the host core while it looks (the waiting thread holds no virtual core: the signaller runs)
|
||||
inline void pause() {
|
||||
#if defined(__x86_64__) || defined(_M_X64) || defined(__i386__) || defined(_M_IX86)
|
||||
_mm_pause();
|
||||
#endif
|
||||
}
|
||||
|
||||
} // namespace mkwvk::gate_wait
|
||||
311
gpu/vk/gpl_rules.h
Normal file
311
gpu/vk/gpl_rules.h
Normal file
@@ -0,0 +1,311 @@
|
||||
// (Triforce) The rules of the fast pipeline link (VK_EXT_graphics_pipeline_library; vk_draw.cpp graphics_pipeline,
|
||||
// warm_libraries, prepare_run), apart from Vulkan so that tests/test_gpl_rules.py runs them. Off unless the hidden Debug
|
||||
// tab's Fast pipeline link row or RX_VK_GPL=1 switches it on: nobody has run the game with it.
|
||||
//
|
||||
// What "the driver step" was, and why the link stayed off "until the driver step is safe". A draw can only be linked from
|
||||
// its program's two shader libraries if the driver has them compiled; compiled in the draw they stall it for a compile
|
||||
// (0.2 to 1.1 s measured on NX2navi), so a draw asks for them under VK_PIPELINE_CREATE_FAIL_ON_PIPELINE_COMPILE_REQUIRED
|
||||
// and goes without when they are cold. They are warm only if the start's "driver's part" made them: the step between the
|
||||
// translation and the pipelines (the start-up screen's "Compiling shaders", vk_draw.cpp warm_libraries), where every
|
||||
// kept program variant's two libraries are compiled once, on every processor, into the pipeline cache. That step was not
|
||||
// safe to leave on:
|
||||
// - it hands every shader of the game to the driver's compiler, and a driver's compiler faults on some of them (amdvlk64
|
||||
// on a mesh program, NVIDIA 615.78 on one vertex library at the same point of every preparation, NX2navi's issue 30;
|
||||
// the lab's start of 11-13 that hung with the driver's own lock held after such a fault);
|
||||
// - what it kept of a start that died in it was one empty file, <prepared>.warming: "this step began and did not end".
|
||||
// The next start then left the whole step out and wrote it down as done for that driver, for good. One shader the
|
||||
// driver cannot compile, a start closed from the task manager, a power cut: each left every library cold, and with
|
||||
// cold libraries no draw is linked, while each new pipeline still has two libraries built behind it on the workers
|
||||
// its whole build waits for. The link then costs and gives nothing;
|
||||
// - a fault the guard catches in that step starts the game again (driver_fault_restart), into exactly that.
|
||||
// NX2navi's equivalent step is its library warm (runtime/vk_pipeline.c lw_worker, lb_warm_prepared). What makes it safe
|
||||
// there, and here now: while a worker has a variant in the driver's hands, a journal on disk names it. A start that finds
|
||||
// a journal left over puts the variants it names on a skip list for this driver, which the step leaves out from then on
|
||||
// (each is built at its first draw, as without libraries), and goes on with the step: what the last start compiled comes
|
||||
// back from the driver's own cache in a moment. Only after kTries starts in a row that began the step and did not end it
|
||||
// is the step left out, as it was after the first. NX2navi's journal and skip list are not its owner's lines alone, so
|
||||
// nothing of them is copied: this is written from what they do.
|
||||
// ((Triforce, 11-18) A try is a start that died with the driver compiling for the step: the journal is opened before the
|
||||
// step, and its "began" word says whether that can be so. The stage libraries' pass (warm_libraries) writes it (and
|
||||
// flushes it) just before its first build_graphics, and it stays: their compiles run back to back. The program pass's
|
||||
// compute warms come first and are spread over the translation, which calls no compiler of the driver's: for them the
|
||||
// word is written just before a guarded_compute when none is under way (Journal::Compiling) and written back to 0 when the
|
||||
// last of them returned. A start that died or was stopped with the word at 0 used no try (the next one finds the count it
|
||||
// found itself): before the step, or in the translation between or after the compute warms, the 5-40 minutes after an
|
||||
// update. One with the word written counts as before, a kill in a hung driver call included. A journal of an older build,
|
||||
// without the word, counts as before)
|
||||
//
|
||||
// And the rules of the link itself, which were in graphics_pipeline's lines:
|
||||
// - a draw is linked only from libraries that are all there; a cold or failed library, or a link the driver refuses,
|
||||
// leaves the entry not ready and the whole build to finish it: what a draw does with an entry that is not ready is
|
||||
// what it does without libraries, nothing more is drawn with it;
|
||||
// - the worker's whole (optimised) build always follows a link, and takes its place when it is made; if it fails, the
|
||||
// linked pipeline stays, as before;
|
||||
// - the linked pipeline a whole build replaced is given back to the driver (it was kept for the run: NX2navi measured
|
||||
// 491 MiB of driver memory a session for its replaced links), but only kRetirePresents presents and kRetireMs later:
|
||||
// a draw recorded with it may still be on its way to the GPU. RX_VK_GPL_KEEP_LINKS=1: kept, as before.
|
||||
#pragma once
|
||||
|
||||
#include <atomic>
|
||||
#include <cstdint>
|
||||
#include <cstdio>
|
||||
#include <deque>
|
||||
#include <mutex>
|
||||
#include <string>
|
||||
#include <vector>
|
||||
|
||||
namespace gpl_rules {
|
||||
|
||||
// ---- an entry: linked first, then the whole build over it
|
||||
|
||||
/// The draw's thread, for a new entry: link is what the libraries linked to, "none" when a library was cold or failed or
|
||||
/// the driver refused the link. True: the entry is ready with it. False: nothing changed, the whole build makes it ready
|
||||
template <class Entry, class Pipeline>
|
||||
bool linked(Entry& e, Pipeline link) {
|
||||
if (link == Pipeline{}) return false;
|
||||
e.fast.store(link, std::memory_order_relaxed);
|
||||
e.pipeline.store(link, std::memory_order_relaxed);
|
||||
e.ready.store(true, std::memory_order_release);
|
||||
return true;
|
||||
}
|
||||
|
||||
/// The worker, the whole build over: whole is "none" when it failed. The entry is ready from here on: with the whole
|
||||
/// build, or with the link it had (the build failed), or with nothing (it failed and there was no link: the draws are left
|
||||
/// out, as before). Returns the link the whole build replaced, for retire(); "none" when there is nothing to give back
|
||||
template <class Entry, class Pipeline>
|
||||
Pipeline built(Entry& e, Pipeline whole) {
|
||||
Pipeline replaced{};
|
||||
if (!(whole == Pipeline{})) {
|
||||
e.pipeline.store(whole, std::memory_order_release);
|
||||
replaced = e.fast.exchange(Pipeline{}, std::memory_order_acq_rel);
|
||||
}
|
||||
e.ready.store(true, std::memory_order_release);
|
||||
return replaced;
|
||||
}
|
||||
|
||||
// ---- the replaced links, given back later
|
||||
|
||||
inline constexpr uint64_t kRetirePresents = 600;
|
||||
inline constexpr uint64_t kRetireMs = 10000;
|
||||
|
||||
/// The links whole builds replaced, until both waits are over (any thread adds, the present's thread takes)
|
||||
template <class Pipeline>
|
||||
class Retired {
|
||||
public:
|
||||
void add(Pipeline p, uint64_t present, uint64_t ms) {
|
||||
std::lock_guard lk(mu_);
|
||||
q_.push_back({p, present, ms});
|
||||
}
|
||||
/// Those replaced kRetirePresents presents and kRetireMs before now, the oldest first: the caller destroys them
|
||||
std::vector<Pipeline> due(uint64_t present, uint64_t ms) {
|
||||
std::vector<Pipeline> out;
|
||||
std::lock_guard lk(mu_);
|
||||
while (!q_.empty() && present >= q_.front().present + kRetirePresents && ms >= q_.front().ms + kRetireMs) {
|
||||
out.push_back(q_.front().p);
|
||||
q_.pop_front();
|
||||
}
|
||||
return out;
|
||||
}
|
||||
size_t waiting() {
|
||||
std::lock_guard lk(mu_);
|
||||
return q_.size();
|
||||
}
|
||||
|
||||
private:
|
||||
struct One {
|
||||
Pipeline p;
|
||||
uint64_t present, ms;
|
||||
};
|
||||
std::mutex mu_;
|
||||
std::deque<One> q_;
|
||||
};
|
||||
|
||||
// ---- the driver's part of the preparation: its journal and its skip list
|
||||
|
||||
/// A program variant (Program::key, Program::variant)
|
||||
struct Id {
|
||||
uint64_t key = 0, variant = 0;
|
||||
bool operator==(const Id& o) const { return key == o.key && variant == o.variant; }
|
||||
};
|
||||
|
||||
inline constexpr uint32_t kJournalMagic = 0x4A574654; // ("TFWJ")
|
||||
inline constexpr uint32_t kSkipMagic = 0x53574654; // ("TFWS")
|
||||
inline constexpr int kSlots = 64; // (the workers of the preparation at most: prepare_threads.h)
|
||||
inline constexpr uint32_t kTries = 3; // (starts in a row that may begin the step and not end it)
|
||||
inline constexpr uint32_t kBeganMark = 0x42574654; // ("TFWB": the driver may be compiling for the step)
|
||||
inline constexpr long kBeganAt = 16 + 16 * kSlots; // (the word after the slots; a file without it is an older build's)
|
||||
|
||||
/// What a start found of the last one's journal
|
||||
struct Left {
|
||||
bool found = false; // (a journal was there: the last start began the step and did not end it)
|
||||
bool began = true; // (and the driver may have been compiling for it: false, it used no try)
|
||||
uint32_t tries = 0; // (the starts in a row that did so, the last one counted; 0: none, or another driver's)
|
||||
std::vector<Id> in_hand; // (the variants its workers had in the driver's hands when it ended)
|
||||
};
|
||||
|
||||
/// The journal at path as the last start left it. stamp: what the driver's part is made for now (vk_draw.cpp warmed_stamp).
|
||||
/// A file of another stamp is another driver's or other shaders': found, with nothing to count and no variant to blame. A
|
||||
/// file without the header is an older build's empty marker: one start, no variant named
|
||||
inline Left journal_left(const std::string& path, uint64_t stamp) {
|
||||
Left l;
|
||||
FILE* f = fopen(path.c_str(), "rb");
|
||||
if (!f) return l;
|
||||
l.found = true;
|
||||
uint32_t head[2]{};
|
||||
uint64_t was = 0;
|
||||
if (fread(head, 4, 2, f) != 2 || fread(&was, 8, 1, f) != 1 || head[0] != kJournalMagic) {
|
||||
l.tries = 1;
|
||||
} else if (was == stamp) {
|
||||
l.tries = head[1] ? head[1] : 1;
|
||||
int slots = 0;
|
||||
for (; slots < kSlots; slots++) {
|
||||
uint64_t r[2];
|
||||
if (fread(r, 8, 2, f) != 2) break;
|
||||
if (r[0] || r[1]) l.in_hand.push_back({r[0], r[1]});
|
||||
}
|
||||
// ((Triforce) the began word: 0, the start ended with no compile of the step under way, and it is not counted.
|
||||
// Without the word (an older build's journal, or one cut short) or with anything else in it, it counts, as before)
|
||||
uint32_t began = kBeganMark;
|
||||
if (slots == kSlots && fread(&began, 4, 1, f) == 1 && began == 0 && l.in_hand.empty()) l.began = false, l.tries--;
|
||||
}
|
||||
fclose(f);
|
||||
return l;
|
||||
}
|
||||
|
||||
/// Whether a start that found l goes on with the step (true) or leaves it out for this driver, as every start did before
|
||||
inline bool goes_on(const Left& l) { return !l.found || l.tries < kTries; }
|
||||
|
||||
/// The journal of a step under way. Every mark is on its way to the disk before the driver is called
|
||||
class Journal {
|
||||
public:
|
||||
Journal() = default;
|
||||
Journal(const Journal&) = delete;
|
||||
Journal& operator=(const Journal&) = delete;
|
||||
~Journal() { close(); }
|
||||
/// tries: the starts in a row that began the step, this one counted (it counts only while the word is on the disk: began(), Compiling)
|
||||
bool open(const std::string& path, uint64_t stamp, uint32_t tries) {
|
||||
std::lock_guard lk(mu_);
|
||||
if (f_) fclose(f_);
|
||||
began_.store(false, std::memory_order_relaxed);
|
||||
compiling_ = 0;
|
||||
f_ = fopen(path.c_str(), "wb");
|
||||
if (!f_) return false;
|
||||
const uint32_t head[2] = {kJournalMagic, tries};
|
||||
const uint64_t zero[2 * kSlots]{};
|
||||
const uint32_t not_yet = 0;
|
||||
const bool ok = fwrite(head, 4, 2, f_) == 2 && fwrite(&stamp, 8, 1, f_) == 1 &&
|
||||
fwrite(zero, 8, (size_t)(2 * kSlots), f_) == (size_t)(2 * kSlots) && fwrite(¬_yet, 4, 1, f_) == 1 &&
|
||||
fflush(f_) == 0;
|
||||
if (!ok) fclose(f_), f_ = nullptr;
|
||||
return ok;
|
||||
}
|
||||
/// ((Triforce) Any worker of the stage libraries' pass, just before a compile call: the start counts as a try from
|
||||
/// here on, for good. The first call writes the word; the others wait for it on the lock, then go by the flag. No
|
||||
/// journal: nothing. (A failed write, over bytes the file has already, is tried once: the marks go on as before it)
|
||||
void began() {
|
||||
if (began_.load(std::memory_order_acquire)) return;
|
||||
std::lock_guard lk(mu_);
|
||||
if (!f_ || began_.load(std::memory_order_relaxed)) return;
|
||||
word(kBeganMark);
|
||||
began_.store(true, std::memory_order_release);
|
||||
}
|
||||
bool has_begun() const { return began_.load(std::memory_order_acquire); }
|
||||
/// ((Triforce) One compile of the step that is not one of a run of them (the program pass's compute warms), held for as
|
||||
/// long as the driver has it: the first of those under way writes the word, the last to return writes 0 over it, so a
|
||||
/// start that dies between them, in the translation, uses no try. Once began() wrote it, the word stays. No journal:
|
||||
/// nothing. (A 0 that cannot be written leaves the word as it was: the start counts, as before)
|
||||
class Compiling {
|
||||
public:
|
||||
explicit Compiling(Journal& j) : j_(j.enter() ? &j : nullptr) {}
|
||||
Compiling(const Compiling&) = delete;
|
||||
Compiling& operator=(const Compiling&) = delete;
|
||||
~Compiling() {
|
||||
if (j_) j_->leave();
|
||||
}
|
||||
|
||||
private:
|
||||
Journal* j_;
|
||||
};
|
||||
/// Whether the word is written now (by began(), or a Compiling under way): what a start that died here would leave
|
||||
bool counts() {
|
||||
std::lock_guard lk(mu_);
|
||||
return f_ && (began_.load(std::memory_order_relaxed) || compiling_ > 0);
|
||||
}
|
||||
/// Worker slot's variant, about to go to the driver ({0, 0}: it came back). A slot past the last has no line
|
||||
void mark(int slot, Id id) {
|
||||
if (slot < 0 || slot >= kSlots) return;
|
||||
std::lock_guard lk(mu_);
|
||||
if (!f_) return;
|
||||
const uint64_t r[2] = {id.key, id.variant};
|
||||
if (fseek(f_, 16 + 16 * slot, SEEK_SET) == 0 && fwrite(r, 8, 2, f_) == 2) fflush(f_);
|
||||
}
|
||||
void close() {
|
||||
std::lock_guard lk(mu_);
|
||||
if (f_) fclose(f_), f_ = nullptr;
|
||||
}
|
||||
|
||||
private:
|
||||
void word(uint32_t v) { // (under mu_)
|
||||
if (fseek(f_, kBeganAt, SEEK_SET) == 0 && fwrite(&v, 4, 1, f_) == 1) fflush(f_);
|
||||
}
|
||||
bool enter() {
|
||||
std::lock_guard lk(mu_);
|
||||
if (!f_ || began_.load(std::memory_order_relaxed)) return false;
|
||||
if (compiling_++ == 0) word(kBeganMark);
|
||||
return true;
|
||||
}
|
||||
void leave() {
|
||||
std::lock_guard lk(mu_);
|
||||
if (!f_ || compiling_ == 0) return;
|
||||
if (--compiling_ == 0 && !began_.load(std::memory_order_relaxed)) word(0);
|
||||
}
|
||||
std::mutex mu_;
|
||||
FILE* f_ = nullptr;
|
||||
std::atomic<bool> began_{false};
|
||||
unsigned compiling_ = 0; // (the Compiling under way; mu_)
|
||||
};
|
||||
|
||||
/// The skip list at path for this stamp (another stamp's, or no file: empty)
|
||||
inline std::vector<Id> skip_read(const std::string& path, uint64_t stamp) {
|
||||
std::vector<Id> ids;
|
||||
FILE* f = fopen(path.c_str(), "rb");
|
||||
if (!f) return ids;
|
||||
uint32_t head[2]{};
|
||||
uint64_t was = 0;
|
||||
if (fread(head, 4, 2, f) == 2 && fread(&was, 8, 1, f) == 1 && head[0] == kSkipMagic && was == stamp) {
|
||||
uint64_t r[2];
|
||||
while (ids.size() < 65536 && fread(r, 8, 2, f) == 2) ids.push_back({r[0], r[1]});
|
||||
}
|
||||
fclose(f);
|
||||
return ids;
|
||||
}
|
||||
|
||||
/// The skip list at path with more added (each variant once), written whole. Returns the list, as skip_read would give it
|
||||
inline std::vector<Id> skip_add(const std::string& path, uint64_t stamp, const std::vector<Id>& more) {
|
||||
std::vector<Id> ids = skip_read(path, stamp);
|
||||
bool grew = false;
|
||||
for (const Id& id : more) {
|
||||
bool have = false;
|
||||
for (const Id& o : ids) have = have || o == id;
|
||||
if (!have) ids.push_back(id), grew = true;
|
||||
}
|
||||
if (!grew) return ids;
|
||||
const std::string tmp = path + ".tmp";
|
||||
FILE* f = fopen(tmp.c_str(), "wb");
|
||||
if (!f) return ids;
|
||||
const uint32_t head[2] = {kSkipMagic, 0};
|
||||
bool ok = fwrite(head, 4, 2, f) == 2 && fwrite(&stamp, 8, 1, f) == 1;
|
||||
for (const Id& id : ids) {
|
||||
const uint64_t r[2] = {id.key, id.variant};
|
||||
ok = ok && fwrite(r, 8, 2, f) == 2;
|
||||
}
|
||||
ok = fclose(f) == 0 && ok;
|
||||
if (ok) {
|
||||
remove(path.c_str()); // (rename does not replace on Windows)
|
||||
ok = rename(tmp.c_str(), path.c_str()) == 0;
|
||||
}
|
||||
if (!ok) remove(tmp.c_str());
|
||||
return ids;
|
||||
}
|
||||
|
||||
} // namespace gpl_rules
|
||||
@@ -2,10 +2,68 @@
|
||||
there. */
|
||||
#pragma once
|
||||
|
||||
#include <cctype>
|
||||
#include <string>
|
||||
#include <vector>
|
||||
|
||||
namespace mkwvk::gpu_choice {
|
||||
|
||||
/// The row's choice as a position in the Vulkan 1.3 devices (0-based), or -1 for Auto. menu: opts.gpu_device (0 Auto,
|
||||
/// n the n-th usable device); usable: how many there are now. A choice past them (a GPU taken out since) is Auto.
|
||||
inline int menu_device(int menu, int usable) { return menu > 0 && menu <= usable ? menu - 1 : -1; }
|
||||
|
||||
/// ((Triforce) 11-20) A device's name as the choice compares it: lower case, every run of white space one space, none at
|
||||
/// either end. Windows and Vulkan name the same card alike but not always to the space ("AMD Radeon RX 7800 XT" from one
|
||||
/// AMD driver, "AMD Radeon RX 7800 XT" from the other)
|
||||
inline std::string name_key(const std::string& s) {
|
||||
std::string out;
|
||||
bool space = false;
|
||||
for (unsigned char c : s) {
|
||||
if (std::isspace(c) || c < 0x20) { space = !out.empty(); continue; }
|
||||
if (space) out += ' ', space = false;
|
||||
out += (char)std::tolower(c);
|
||||
}
|
||||
return out;
|
||||
}
|
||||
|
||||
/// ((Triforce) 11-20, tickets "iGPU used on laptop with dGPU to compile shaders", "Launcher choosing the wrong GPU when
|
||||
/// compiling shaders" and two more) The device a choice means: by its name when one was saved (triforce_options.ini
|
||||
/// gpu_device_name, which the launcher and the menu write with the number), else by the number as before (menu_device).
|
||||
/// The launcher numbered the cards in Windows' order before the game had ever run (no triforce_gpus.txt yet), and Vulkan's
|
||||
/// is often the other way round on a laptop: its "2: RTX 5060" was the game's 2, the 860M, and the build prepared every
|
||||
/// shader there. A name that matches several cards (two of the same) takes the numbered one when its name is that name,
|
||||
/// else the first. No name, or none matching (the card taken out): the number decides. Auto (menu 0) is Auto whatever name
|
||||
/// the file still has. names: the usable devices' names in the game's order. Returns a position in them (0-based) or -1 for Auto
|
||||
inline int pick(const std::string& want, const std::vector<std::string>& names, int menu) {
|
||||
if (menu <= 0) return -1; // (Auto: a name left in the file chooses nothing)
|
||||
const std::string key = name_key(want);
|
||||
if (!key.empty()) {
|
||||
const int numbered = menu_device(menu, (int)names.size());
|
||||
if (numbered >= 0 && name_key(names[(size_t)numbered]) == key) return numbered;
|
||||
for (size_t i = 0; i < names.size(); i++)
|
||||
if (name_key(names[i]) == key) return (int)i;
|
||||
}
|
||||
return menu_device(menu, (int)names.size());
|
||||
}
|
||||
|
||||
/// ((Triforce) 11-22, tickets "igpu targeted for shader cache", "Launcher \"GPU device\" setting is non-functional" (RX 6600
|
||||
/// with a Ryzen 5600G's graphics) and on 0.6.0 "still picking IGPU for building shaders") The choice that pick takes by its
|
||||
/// number alone (no saved name, or a name no card has here) when that number is the processor's graphics while a card of its
|
||||
/// own is listed: not taken, Auto instead. Such a number comes from an options file written before 11-20, when the launcher
|
||||
/// numbered the cards in Windows' order (the display class's driver keys: the processor graphics' driver first on a 5600G)
|
||||
/// and saved no name with it; the update to 0.6.0 still prepared its shaders there and the game started there, as no name
|
||||
/// came with the number until the row was chosen again. A choice saved with its name (the launcher and the menu since 11-20)
|
||||
/// is taken as it is, the processor graphics too. types: the usable devices' VkPhysicalDeviceType (1 integrated, 2 discrete),
|
||||
/// in the game's order, as names
|
||||
inline bool unnamed_integrated(const std::string& want, const std::vector<std::string>& names, const std::vector<int>& types,
|
||||
int menu) {
|
||||
const int k = pick(want, names, menu);
|
||||
if (k < 0 || (size_t)k >= types.size() || types[(size_t)k] != 1) return false;
|
||||
const std::string key = name_key(want);
|
||||
if (!key.empty() && name_key(names[(size_t)k]) == key) return false; // (chosen by its name)
|
||||
for (const int t : types)
|
||||
if (t == 2) return true;
|
||||
return false;
|
||||
}
|
||||
|
||||
} // namespace mkwvk::gpu_choice
|
||||
|
||||
57
gpu/vk/gpu_clock.h
Normal file
57
gpu/vk/gpu_clock.h
Normal file
@@ -0,0 +1,57 @@
|
||||
// (Triforce) Whether the renderer holds an AMD GPU at amdgpu's stable power state while the game runs, and which one (Linux:
|
||||
// vk_gpu_clock.cpp; RX_VK_GPU_CLOCK). Plain code with no system calls: tests/test_gpu_clock.py runs it on made-up devices.
|
||||
//
|
||||
// The Steam Deck's automatic GPU clock drops while the frame waits on the CPU or a readback (200 MHz for 27% of a Kokiri run), so
|
||||
// the GPU work after the wait runs slow. amdgpu lets any process that opens the GPU's render node ask for a stable power state
|
||||
// through a context of its own (DRM_IOCTL_AMDGPU_CTX, AMDGPU_CTX_OP_SET_STABLE_PSTATE, Linux 5.18): no root and no sysfs write.
|
||||
// The kernel gives the level back when that context goes, at the process's end too, a crash or kill -9 included. Measured on a
|
||||
// Deck OLED (SteamOS 3.9.5, plugged in): Kokiri handheld 11.1 -> 16.2 fps, 1% low 6.8 -> 12.1, GPU time 58 -> 32 ms at the same
|
||||
// 15 W; Hyrule Field +19% for 4 W more. It costs battery where the APU was under its power cap.
|
||||
//
|
||||
// - Unset: the peak state, on the Deck's APU only (its Vulkan name, or SteamDeck=1 from Steam on an integrated AMD GPU), and
|
||||
// only while the level is amdgpu's own (power_dpm_force_performance_level "auto"): a level the player set is left alone.
|
||||
// - RX_VK_GPU_CLOCK=0: nothing opened. =peak (or 1), =standard: that state on any AMD GPU, whatever the level.
|
||||
#pragma once
|
||||
|
||||
#include <cstdint>
|
||||
#include <cstring>
|
||||
|
||||
namespace mkwvk::gpu_clock {
|
||||
|
||||
enum Want { kNone, kPeak, kStandard };
|
||||
|
||||
// (uapi/drm/amdgpu_drm.h, written out: no libdrm and no kernel headers needed to build)
|
||||
constexpr uint32_t kCtxOpAlloc = 1, kCtxOpSetStablePstate = 6;
|
||||
constexpr uint32_t kPstateStandard = 1, kPstatePeak = 4;
|
||||
constexpr uint32_t kCtxArgBytes = 16; // (union drm_amdgpu_ctx)
|
||||
/// DRM_IOCTL_AMDGPU_CTX: DRM_IOWR(DRM_COMMAND_BASE + DRM_AMDGPU_CTX, union drm_amdgpu_ctx) = _IOWR('d', 0x40 + 0x02, 16 bytes)
|
||||
constexpr unsigned long kCtxIoctl = (3ul << 30) | ((unsigned long)kCtxArgBytes << 16) | ((unsigned long)'d' << 8) | 0x42ul;
|
||||
|
||||
/// The Deck's APU: "AMD Custom GPU 0405" (LCD) or "0932" (OLED) in its Vulkan name; or SteamDeck=1 (Steam sets it there) on an
|
||||
/// integrated GPU
|
||||
inline bool deck_apu(const char* device_name, bool integrated, const char* steam_deck_env) {
|
||||
if (device_name && (std::strstr(device_name, "Custom GPU 0405") || std::strstr(device_name, "Custom GPU 0932"))) return true;
|
||||
return integrated && steam_deck_env && !std::strcmp(steam_deck_env, "1");
|
||||
}
|
||||
|
||||
struct Choice {
|
||||
Want want;
|
||||
const char* why;
|
||||
};
|
||||
|
||||
/// switch_value: RX_VK_GPU_CLOCK (nullptr unset); level: power_dpm_force_performance_level's text, first word (nullptr: unread)
|
||||
inline Choice choose(const char* switch_value, uint32_t vendor_id, bool integrated, const char* device_name,
|
||||
const char* steam_deck_env, const char* level) {
|
||||
if (vendor_id != 0x1002) return {kNone, "not an AMD GPU"};
|
||||
if (switch_value && *switch_value) {
|
||||
if (!std::strcmp(switch_value, "0")) return {kNone, "RX_VK_GPU_CLOCK=0"};
|
||||
if (!std::strcmp(switch_value, "peak") || !std::strcmp(switch_value, "1")) return {kPeak, "RX_VK_GPU_CLOCK=peak"};
|
||||
if (!std::strcmp(switch_value, "standard")) return {kStandard, "RX_VK_GPU_CLOCK=standard"};
|
||||
}
|
||||
if (!deck_apu(device_name, integrated, steam_deck_env)) return {kNone, "not a Steam Deck (RX_VK_GPU_CLOCK=peak: any AMD GPU)"};
|
||||
if (!level) return {kNone, "the GPU's performance level could not be read"};
|
||||
if (std::strcmp(level, "auto")) return {kNone, "the GPU's performance level is set by hand"};
|
||||
return {kPeak, "the Steam Deck's GPU, its level amdgpu's own"};
|
||||
}
|
||||
|
||||
} // namespace mkwvk::gpu_clock
|
||||
86
gpu/vk/gpu_wait_kinds.h
Normal file
86
gpu/vk/gpu_wait_kinds.h
Normal file
@@ -0,0 +1,86 @@
|
||||
// (Triforce, 11-22, the performance audit's SYNC-27, after Source A's) The host's time in calls that block on the GPU,
|
||||
// counted by kind where the Vulkan calls are wrapped (vk_work.h).
|
||||
//
|
||||
// The log's "GPU-bound wait" counts only the waits for a recording slot: it reads about 0 while threads wait on the GPU
|
||||
// elsewhere (a fence of the game's, a write-back, the present's frame fence, a device idle, a query read that waits).
|
||||
// Those were timed apart or not at all. Here every blocking wait is timed in one place and summed by kind:
|
||||
// fences vkWaitForFences with a timeout above 0
|
||||
// timelines vkWaitSemaphores with a timeout above 0
|
||||
// idle vkQueueWaitIdle, vkDeviceWaitIdle
|
||||
// queries vkGetQueryPoolResults with VK_QUERY_RESULT_WAIT_BIT
|
||||
// A poll (timeout 0, no wait bit) counts nothing. A wait made inside another counts once, as the outer one (a depth of
|
||||
// the thread's). A wait inside an Exclude scope counts nothing: pacing (the driver's low-latency sleep), which is not a
|
||||
// wait for the GPU's work. The totals only grow; each reader keeps its own last values (Reader), so the log's line and
|
||||
// the frame trace's column each see every nanosecond.
|
||||
//
|
||||
// The cost is two clock reads a blocking wait. RX_VK_GPU_WAITS=0, or the lab knob nogpuwaits in its phases: the calls
|
||||
// straight through, nothing counted. No Vulkan here: tests/test_gpu_waits.py compiles it.
|
||||
#pragma once
|
||||
#include <atomic>
|
||||
#include <chrono>
|
||||
#include <cstdint>
|
||||
#include <cstdlib>
|
||||
|
||||
#include "lab_ab.h"
|
||||
|
||||
namespace gpu_wait_kinds {
|
||||
|
||||
enum Kind : int { FENCE = 0, TIMELINE = 1, IDLE = 2, QUERY = 3, KINDS = 4 };
|
||||
|
||||
struct Totals {
|
||||
std::atomic<int64_t> ns[KINDS]{}, n[KINDS]{};
|
||||
};
|
||||
inline Totals totals; // (every thread's, since the start: they only grow)
|
||||
inline thread_local int depth = 0; // (this thread is inside a counted wait)
|
||||
inline thread_local int excluded = 0; // (this thread is inside an Exclude scope)
|
||||
|
||||
/// Counted at all (RX_VK_GPU_WAITS, read once; on unless 0), and not in the lab knob's phases
|
||||
inline bool on() {
|
||||
static const bool env = [] {
|
||||
const char* e = getenv("RX_VK_GPU_WAITS");
|
||||
return !(e && *e && atoi(e) == 0);
|
||||
}();
|
||||
return env && !mkwvk::lab_ab::on("nogpuwaits");
|
||||
}
|
||||
|
||||
/// The waits made on this thread while one of these lives count nothing
|
||||
struct Exclude {
|
||||
Exclude() { excluded++; }
|
||||
~Exclude() { excluded--; }
|
||||
Exclude(const Exclude&) = delete;
|
||||
Exclude& operator=(const Exclude&) = delete;
|
||||
};
|
||||
|
||||
/// call(), and its time under `kind` when it blocks (blocks: a timeout above 0, the wait bit), unless this thread is
|
||||
/// inside a counted wait or an Exclude scope, or the count is off. What call() returns
|
||||
template <class Call>
|
||||
inline auto timed(Kind kind, bool blocks, Call&& call) -> decltype(call()) {
|
||||
if (!blocks || depth || excluded || !on()) return call();
|
||||
depth++;
|
||||
const auto t0 = std::chrono::steady_clock::now();
|
||||
const auto result = call();
|
||||
const int64_t ns = std::chrono::duration_cast<std::chrono::nanoseconds>(std::chrono::steady_clock::now() - t0).count();
|
||||
depth--;
|
||||
totals.ns[kind].fetch_add(ns, std::memory_order_relaxed);
|
||||
totals.n[kind].fetch_add(1, std::memory_order_relaxed);
|
||||
return result;
|
||||
}
|
||||
|
||||
/// One reader's last values: take() gives what was counted since its last take (ns and waits by kind, either may be
|
||||
/// null) and returns the nanoseconds of all kinds together
|
||||
struct Reader {
|
||||
int64_t last_ns[KINDS]{}, last_n[KINDS]{};
|
||||
int64_t take(int64_t* ns_by_kind, int64_t* n_by_kind) {
|
||||
int64_t all = 0;
|
||||
for (int k = 0; k < KINDS; k++) {
|
||||
const int64_t ns = totals.ns[k].load(std::memory_order_relaxed), n = totals.n[k].load(std::memory_order_relaxed);
|
||||
if (ns_by_kind) ns_by_kind[k] = ns - last_ns[k];
|
||||
if (n_by_kind) n_by_kind[k] = n - last_n[k];
|
||||
all += ns - last_ns[k];
|
||||
last_ns[k] = ns, last_n[k] = n;
|
||||
}
|
||||
return all;
|
||||
}
|
||||
};
|
||||
|
||||
} // namespace gpu_wait_kinds
|
||||
112
gpu/vk/guest_holes.h
Normal file
112
gpu/vk/guest_holes.h
Normal file
@@ -0,0 +1,112 @@
|
||||
// (Triforce) The all-zero upload check's reads (vk_texture.cpp upload_all_zero): which bytes of a storage it reads, split between its
|
||||
// threads. tests/test_guest_holes.py.
|
||||
//
|
||||
// On Linux, where /dev/udmabuf is usable, the guest window is a memfd's (host/hos_main.c). A page the game never wrote is a hole
|
||||
// there: it reads as zero, but a read through the shared mapping makes the kernel allocate and zero it. The check read every byte
|
||||
// of a 32 MB+ storage, so the never-written 288 MB shadow map (12288 x 12288 D16) became RAM at each Hyrule Field arrival, and
|
||||
// took 200-300 ms of the replay thread on a Steam Deck. fd_data_extents asks the memfd where its data is (lseek SEEK_DATA and
|
||||
// SEEK_HOLE): a hole is zero by definition, and SEEK_DATA may call a hole data, never the other way round (a page in swap is
|
||||
// data), so reading only the data says the same. Anywhere else, or on any failure, one extent of the whole storage.
|
||||
#pragma once
|
||||
|
||||
#include <algorithm>
|
||||
#include <atomic>
|
||||
#include <cstdint>
|
||||
#include <cstring>
|
||||
#include <thread>
|
||||
#include <vector>
|
||||
#if defined(__linux__) || defined(GUEST_HOLES_TEST) // (the test: its own lseek)
|
||||
#include <cerrno>
|
||||
#include <unistd.h>
|
||||
#endif
|
||||
|
||||
namespace mkwvk::guest_holes {
|
||||
|
||||
struct Extent {
|
||||
uint64_t lo, hi; // (bytes from the storage's start)
|
||||
};
|
||||
|
||||
inline uint64_t bytes(const std::vector<Extent>& ex) {
|
||||
uint64_t n = 0;
|
||||
for (const Extent& e : ex) n += e.hi - e.lo;
|
||||
return n;
|
||||
}
|
||||
|
||||
/// Thread t's part of n: the extents' bytes, not their range, split into n runs of whole 4 KB (the last takes the rest); every
|
||||
/// byte of the extents in exactly one part
|
||||
inline std::vector<Extent> share(const std::vector<Extent>& ex, uint32_t t, uint32_t n) {
|
||||
std::vector<Extent> out;
|
||||
const uint64_t total = bytes(ex);
|
||||
if (!n || t >= n || !total) return out;
|
||||
const uint64_t per = ((total / n) + 4095) & ~4095ull;
|
||||
const uint64_t a = std::min<uint64_t>(total, (uint64_t)t * per), b = t + 1 == n ? total : std::min<uint64_t>(total, a + per);
|
||||
uint64_t at = 0; // (bytes of the extents before e)
|
||||
for (const Extent& e : ex) {
|
||||
const uint64_t len = e.hi - e.lo, lo = std::max(a, at), hi = std::min(b, at + len);
|
||||
if (lo < hi) out.push_back({e.lo + (lo - at), e.lo + (hi - at)});
|
||||
at += len;
|
||||
if (at >= b) break;
|
||||
}
|
||||
return out;
|
||||
}
|
||||
|
||||
/// Whether any byte of p's extents is not 0, read on nthreads threads (this one among them), each stopping once one found it
|
||||
inline bool any_set(const uint8_t* p, const std::vector<Extent>& ex, uint32_t nthreads) {
|
||||
std::atomic<bool> found{false};
|
||||
if (ex.empty()) return false;
|
||||
nthreads = std::max(nthreads, 1u);
|
||||
auto part = [&](uint32_t t) {
|
||||
for (const Extent& e : share(ex, t, nthreads))
|
||||
for (uint64_t at = e.lo; at < e.hi && !found.load(std::memory_order_relaxed); at += 65536) {
|
||||
const uint64_t end = std::min<uint64_t>(e.hi, at + 65536);
|
||||
uint64_t acc = 0;
|
||||
uint64_t k = at;
|
||||
for (; k + 8 <= end; k += 8) {
|
||||
uint64_t w;
|
||||
std::memcpy(&w, p + k, 8);
|
||||
acc |= w;
|
||||
}
|
||||
for (; k < end; k++) acc |= p[k];
|
||||
if (acc) found.store(true, std::memory_order_relaxed);
|
||||
}
|
||||
};
|
||||
if (nthreads > 1) {
|
||||
std::vector<std::thread> th;
|
||||
for (uint32_t t = 1; t < nthreads; t++) th.emplace_back(part, t);
|
||||
part(0);
|
||||
for (auto& x : th) x.join();
|
||||
} else {
|
||||
part(0);
|
||||
}
|
||||
return found.load();
|
||||
}
|
||||
|
||||
#if defined(__linux__) || defined(GUEST_HOLES_TEST) // (the test: its own lseek)
|
||||
/// The data of [off, off + len) in fd, as extents from off (none: all holes). At most max extents: past them, the rest as one.
|
||||
/// False when the file cannot say (a seek failed): the caller reads it all
|
||||
inline bool fd_data_extents(int fd, uint64_t off, uint64_t len, std::vector<Extent>* out, size_t max = 1024) {
|
||||
out->clear();
|
||||
const uint64_t end = off + len;
|
||||
uint64_t pos = off;
|
||||
while (pos < end) {
|
||||
if (out->size() + 1 >= max) { // (the last one: the rest, whole)
|
||||
out->push_back({pos - off, len});
|
||||
return true;
|
||||
}
|
||||
const off_t d = lseek(fd, (off_t)pos, SEEK_DATA);
|
||||
if (d < 0) {
|
||||
if (errno == ENXIO) return true; // (no data from pos to the file's end)
|
||||
return out->clear(), false;
|
||||
}
|
||||
if ((uint64_t)d >= end) return true;
|
||||
const off_t h = lseek(fd, d, SEEK_HOLE);
|
||||
if (h < 0 || (uint64_t)h <= (uint64_t)d) return out->clear(), false;
|
||||
const uint64_t hi = std::min<uint64_t>(end, (uint64_t)h);
|
||||
out->push_back({(uint64_t)d - off, hi - off});
|
||||
pos = hi;
|
||||
}
|
||||
return true;
|
||||
}
|
||||
#endif
|
||||
|
||||
} // namespace mkwvk::guest_holes
|
||||
49
gpu/vk/hang_watch.h
Normal file
49
gpu/vk/hang_watch.h
Normal file
@@ -0,0 +1,49 @@
|
||||
// The hang watch's judgement, apart from its thread, the clock and the stacks (vk_hangdump.cpp). After NX2navi's hang
|
||||
// watch (its runtime/hang_watch.c hw_step): a report when no frame is shown for some seconds, once a stall, a few a run.
|
||||
// tests/test_hang_watch.py
|
||||
#pragma once
|
||||
|
||||
#include <cstdint>
|
||||
|
||||
namespace mkwvk::hangwatch {
|
||||
|
||||
constexpr int kReports = 3; // (at most this many a run: a game that stalls again and again says so three times)
|
||||
|
||||
// ((Triforce) The stacks asked for from outside: a file of this name beside the report's, "triforce_hang.txt.now" in the game
|
||||
// folder, is looked for every second; the watch takes it away and writes where every thread is, into the log and the
|
||||
// report's file, whether or not a frame is shown. For a game that goes on showing frames and gets nowhere: the owner's
|
||||
// save that did not load on 2026-11-20 showed its loading screen at 240 frames a second for minutes, one game thread
|
||||
// (BaseProcCreateAndDeleteThread) at a whole core, and the stall's own report never came: frames were shown. Asked for
|
||||
// with a file because that needs nothing done to the game's process from outside. At most kAsked a run)
|
||||
constexpr const char* kAskSuffix = ".now";
|
||||
constexpr int kAsked = 16;
|
||||
/// Whether an ask found now is served: the first kAsked of a run (`served`: how many were, counted up here)
|
||||
inline bool ask(int& served) {
|
||||
if (served >= kAsked) return false;
|
||||
served++;
|
||||
return true;
|
||||
}
|
||||
|
||||
struct Watch {
|
||||
int64_t presents = -1; // the presents counted at the last look
|
||||
int still = 0; // judged seconds running without one
|
||||
bool told = false; // this stall has its report
|
||||
int reports = 0;
|
||||
};
|
||||
|
||||
/// One second's look. presents: the frames shown so far. judged: whether a second without one counts (the game's window
|
||||
/// in front and not minimized: the game may stop showing frames behind another window, and a watch asked for by name
|
||||
/// judges always). True when a report is due now: `limit` judged seconds running without a present, once per stall (a
|
||||
/// new one needs a present first), at most kReports a run.
|
||||
inline bool step(Watch& w, int64_t presents, bool judged, int limit) {
|
||||
if (presents != w.presents || !judged) {
|
||||
w.presents = presents, w.still = 0, w.told = false;
|
||||
return false;
|
||||
}
|
||||
if (++w.still < limit || w.told || w.reports >= kReports) return false;
|
||||
w.told = true;
|
||||
w.reports++;
|
||||
return true;
|
||||
}
|
||||
|
||||
} // namespace mkwvk::hangwatch
|
||||
60
gpu/vk/helper_spin.h
Normal file
60
gpu/vk/helper_spin.h
Normal file
@@ -0,0 +1,60 @@
|
||||
// (Triforce, the third performance audit's P6, after NX2navi's replay workers: runtime/vk_exec.c rw_main's idle spin and the
|
||||
// submitting thread's spin for its chunks, epicmeme546, 3010a637, 7534193b, 7eee4817, ffa599ef) A bounded spin before a sleep
|
||||
// on a condition variable, for the parallel replay (vk_replay.cpp): a helper thread with no segment to record spins for the
|
||||
// next before it sleeps (par_worker_main), and the replay thread spins for a segment a helper records before it sleeps for it
|
||||
// (replay_handle). What each waits for is published where the spin can read it without the lock (ParPool::queued, Job::done),
|
||||
// and the condition variable's wait with its predicate follows the spin as before: the spin only makes the wait shorter when
|
||||
// what it waits for comes within it, it never decides anything. The audit's trace: the helpers began a segment 0.262 ms after
|
||||
// they were given it, and the replay thread waited 1.74 ms a frame for them (r22-trace). tests/test_helper_spin.py
|
||||
//
|
||||
// RX_VK_HELPER_SPIN_US=<us>: the spin's bound (20 since 11-21; 100 before, NX2navi's; at most 1000000); 0: no spin, straight to the wait as before
|
||||
#pragma once
|
||||
#include <chrono>
|
||||
#include <cstdint>
|
||||
#include <cstdlib>
|
||||
#if defined(__x86_64__) || defined(_M_X64) || defined(__i386__) || defined(_M_IX86)
|
||||
#include <immintrin.h>
|
||||
#endif
|
||||
|
||||
namespace helper_spin {
|
||||
|
||||
/// The bound in nanoseconds that RX_VK_HELPER_SPIN_US's text gives (null: unset)
|
||||
inline int64_t bound_ns(const char* text) {
|
||||
if (!text || !*text) return 20000; // ((Triforce, 11-21) 20 us, was 100 (NX2navi's): with the game's threads the frame's limit, a
|
||||
// helper spinning 100 us took its SMT sibling from them. Lab A/B 79.3 -> 80.0, 1% low +2.0)
|
||||
const long long us = std::atoll(text);
|
||||
return us <= 0 ? 0 : us >= 1000000 ? 1000000000LL : us * 1000;
|
||||
}
|
||||
/// RX_VK_HELPER_SPIN_US, read once
|
||||
inline int64_t spin_ns() {
|
||||
static const int64_t v = bound_ns(std::getenv("RX_VK_HELPER_SPIN_US"));
|
||||
return v;
|
||||
}
|
||||
inline void pause() {
|
||||
#if defined(__x86_64__) || defined(_M_X64) || defined(__i386__) || defined(_M_IX86)
|
||||
_mm_pause();
|
||||
#endif
|
||||
}
|
||||
/// Spins until done() holds or ns have passed; whether done() held. ns <= 0: done() once, no spin. *spent (when given): the
|
||||
/// time spun
|
||||
template <class Done>
|
||||
bool until(Done&& done, int64_t ns, int64_t* spent = nullptr) {
|
||||
if (spent) *spent = 0;
|
||||
if (done()) return true;
|
||||
if (ns <= 0) return false;
|
||||
const auto t0 = std::chrono::steady_clock::now();
|
||||
const auto end = t0 + std::chrono::nanoseconds(ns);
|
||||
bool held = false;
|
||||
for (unsigned i = 1;; i++) {
|
||||
pause();
|
||||
if ((held = done())) break;
|
||||
if (!(i & 7) && std::chrono::steady_clock::now() >= end) {
|
||||
held = done();
|
||||
break;
|
||||
}
|
||||
}
|
||||
if (spent) *spent = std::chrono::duration_cast<std::chrono::nanoseconds>(std::chrono::steady_clock::now() - t0).count();
|
||||
return held;
|
||||
}
|
||||
|
||||
} // namespace helper_spin
|
||||
Some files were not shown because too many files have changed in this diff Show More
Reference in New Issue
Block a user