Commit Graph

385 Commits

Author SHA1 Message Date
chrg6
7966181115 Load the menu's zstd dictionaries once behind a once-guard, so a second thread asking during the load waits instead of finding none.
The loader set the count to 0 before it read the file; the menu-art thread and the present path both ask in the first frames, and the one that lost took 0 for "no dictionaries" and kept the fallback fonts for the whole run.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 23:27:51 -05:00
chrg6
3e67a74cfc Ask again for a VRAM mirror whose device-local allocation failed, as for one refused for the budget, and log refusals four times per reason.
A pool refused for a failed allocation was never asked again and was read over the bus until the game finalized it. The shared quota of twelve log lines could be used up by budget refusals before the first allocation failure.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 23:27:15 -05:00
chrg6
ccc31ad562 Forget a texture's content hash when what reached its image is not its storage: zeros for storage released mid-upload, or nothing after a failed GPU ASTC decode.
With the hash kept, the same bytes written there again were skipped as unchanged and the image stayed wrong. A texture whose decode failed is still not uploaded again until its storage is written; that part stays open.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 23:26:45 -05:00
chrg6
4ffaf6ea96 Hash a texture's storage only while it is held mapped, at the upload and at the unchanged-bytes check.
Both hashes read guest memory the game may unmap on its own thread; the unchanged-bytes check had no readability test at all.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 23:26:33 -05:00
chrg6
7ccb958c47 Let a save Commit that is under way finish before the crash report or a fatal error ends the process, for at most two seconds.
The game's Commit returns once the commit is queued, so the game may already have reported the save as done while its thread still copied; ending then left an unfinished copy that the next start rolled back.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 23:26:01 -05:00
chrg6
5ddff078c3 Count the save Commits under way instead of flagging one, so a second Commit entered beside the first keeps save writes out of its copy.
With a flag the first commit thread's end cleared it while the second still waited for the transaction lock; a write to the save then passed the wait and landed in the working copy the second thread was copying.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 23:25:39 -05:00
chrg6
d15b4c95aa Count a frame's submissions atomically: several guest threads submit at once from the pause menu's first frame on.
With a lost count the frame's first submission could go without its wait for the GPU, which is there against frames drawn with the next frame's data.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 23:24:49 -05:00
chrg6
61e1ca8bad Drop a zstd stream's slot by its stream state under one hold of the lock, not by a pointer found earlier.
Between the lookup and the drop another thread could take the spent slot for its own stream; the drop then cleared that thread's slot, and a third thread could take it as free and reset the decoder context still in use.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 23:24:34 -05:00
chrg6
9be72ae9c0 When the system-event table is full, reuse the slot of an event the guest has destroyed instead of ending the process.
Entries were never removed, and every reload of a save binds two events at new addresses.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 23:24:12 -05:00
chrg6
4a69a38b48 When the jmp_buf table is full, give the new buffer the slot of the one filled longest ago instead of ending the process.
The table never removes entries and a session keeps meeting new buffer addresses: 12 to 15 more with every reload of a save. A longjmp to a replaced buffer ends the process as one to an unknown buffer always did.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 23:24:12 -05:00
chrg6
abd2e2c392 Compute the menu grain's lattice hash in unsigned arithmetic: the signed products overflowed, which is undefined.
The bits are the same; the undefined-behaviour sanitizer stopped the start at this line.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 23:23:24 -05:00
chrg6
fb44809074 Name an area the game asks for and this game version lacks, in the log and in the crash report.
A save that names such an area (a child save with HyruleCastle_Child on 1.1.1) makes the game read through a null area object right after the missing files; the report now says which area it was and that the save cannot be loaded.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 23:23:12 -05:00
chrg6
e887a1d18b Delete the boot prebuild's running marker at every orderly exit, so closing the game during boot is not taken for a driver crash.
The marker was deleted only when the boot's builds ended. A window closed before that left it, the next start read it as a start that died in the driver's compiler, and every later start on that GPU and driver skipped the prebuild and built each pipeline at its first draw.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 23:22:27 -05:00
chrg6
b933e3d996 List a texture pool slot without an entry once, so the walk over those slots no longer stalls every 8th frame in the Market.
A slot that resolved kept its place in the list until the next walk, and when it stopped resolving before that walk it was listed again. The list reached 160,000 entries for the 32,768 slots after a minute in the adult Market, and each walk over it held the texture lock for 15 to 30 ms.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 23:21:39 -05:00
chrg6
6b68516fb1 Count the page generations handed out in the perf trace's size lines, so a run shows how fast a session uses up the 32-bit generation counter.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-10 22:48:27 -05:00
chrg6
d2afdfa0f4 Under the perf trace, name the guest threads that submit and present on each queue, and log a submit or present that starts while another runs.
The synchronous replay's front end takes no lock of its own and relies on one such call at a time; a run now shows whether that holds.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-10 22:35:29 -05:00
chrg6
78223225f4 Mark DLSS and DLSS frame generation feature creates as work under the perf trace, with the device idle wait of releasing the kept features before them.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-10 22:34:54 -05:00
chrg6
ab4bd8f8f9 Give a slow frame's line under the perf trace its drain wait and replay time, so a stall inside the replay or behind the present thread is no longer read as unnamed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-10 22:34:30 -05:00
chrg6
ded13958de Log the sizes of the caches, pools and tables under the perf trace, every 600 presents and with each memory snapshot.
Four "mem: sizes" lines give the entries and bytes of what no other line counted: the guest heap, physical maps, kernel region, lazily written ranges and the fixed kernel, setjmp, save-file, RomFS and zstd tables; pool records, mirrors, deferred destructions, upload and staging buffers, the device address table and the texture, sampler and slot caches; image views, descriptor spares, the completion queue, uniform snapshot arenas, recording copies and pins; and pipelines, layouts, stage libraries, job queues, the shader memory cache and the translator processes. Each subsystem reads its own counters under its lock; settings now count the replaced values they keep. Segment and scene lines say "-" for the world clock before it starts.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-10 19:51:36 -05:00
chrg6
ecd34fb513 Under the perf trace, report frame times per route segment and take a memory snapshot after each area load.
An input script line "<tick> MARK <name>" (or "@<ms> MARK <name>") ends the running segment with one line of its frame count, mean, 1% low and worst frame, the work marks on any thread, process memory and VRAM, then the 600-frame memory lines, and starts a segment of that name. RomFS opens of an area's static placement file give a "scene:" line, and once no area has loaded for 10 s the memory snapshot is taken with a "settled" line.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-10 19:40:02 -05:00
chrg6
66aedffcdd Name the work behind each frame on any thread under the perf trace: RomFS reads, zstd decodes, pipeline creates, guest memory maps, pool initialize and finalize, and device memory allocations.
A slow frame's line and the 60-frame window's line now say how long each kind took on any thread, how often, how many bytes and the longest single one, so a hitch outside the render thread's own waits has a named cause. While NX2NAVI_TRACE=perf is off each site pays one branch.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-10 19:27:58 -05:00
chrg6
28f78d4045 List the mesh draw switch among the settings, and count the given back pieces under the lock.
Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 16:46:05 -05:00
chrg6
892859cad0 Fill the records of a run of indirect mesh draws in one gap, so their pass is not cut at each of them.
An indirect mesh draw reads its workgroup count through a ring entry that a buffer copy fills, and a copy cannot be
recorded inside a pass: every such draw ended its pass, 45 times a frame in a village. The first draw of a run now
fills the entries of the mesh draws that follow it with only state, draws, counter reports and render conditions
between them, and those draws stay in the pass (mesh_ahead, default 1; 0 is the old way, ab switches in A/B blocks).
An entry is taken only by the draw it was filled for, at the same record address, while no other op was replayed
since. Counter reports count as staying in the pass because the replay closes their query inside it and writes the
value from the completion thread; the game reports one after most of these draws, and without that only 38% of the
draws were served.

Measured in the village: 42 of the 45 draws a frame are served by an entry filled ahead; gaps with these copies
41 to 3 a frame (0.22 to 0.04 ms of GPU time); passes and dispatches 248 to 208 a frame (302 to 254 in Hyrule
Field). With the piece release of the commit before, one run each: village 142.6 fps and GPU busy 3.87 ms (138.0
and 135.4 fps, 4.15 ms without this change), Hyrule Field 136.8 fps and 4.34 ms (131.5 and 131.3, 4.70 ms). In A/B
blocks of one run with the narrower rule (counter reports still ending a run) the GPU's busy time went from 4.81 to
4.64 ms.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 16:44:02 -05:00
chrg6
1241efa458 Give back only the touched piece of a lazily written range, and say in the log why ranges are written back.
With the lazy write-back on, a village and Hyrule Field still merged 20 to 29 MiB a frame back into guest memory at
the end of every submission, inside the GPU work the game's fence waits for. The new log line (mem: lazy write-back
at frame N: written back and not armed since the last line) names the reason by bytes: none of it lay in frozen
regions or on pages that could not be armed; it lay in regions held after a touch. The game's CPU reads a few words
at the start of buffers its shaders fill (12 bytes into ranges of 6.3 and 15.2 MiB among others), about once a
second. Each read gave back the whole armed range and held every 2 MiB region it reached for 1 s, doubling to 32 s,
and nothing was armed in a held region.

A touch now gives back the 64 KiB around its address, fetched alone, and holds only that piece; the rest of the
range stays armed (wb_lazy_chunk, default 1; 0 is the old rule). More than four touches in one region within a
second are taken as a reader going through the range: that range goes whole and its regions are held as before,
and the region's touches go that way at once while it keeps being touched within a minute. Each of the first eight
touches of a log window is logged under the perf trace with its thread, address, range and what was given back.

Measured on one machine (NVIDIA, uncapped, frame generation off, scripted walks, medians over the windows of 46 s
of world clock):
- village: 121.8, 124.1 and 124.4 fps before; 138.0 and 135.4 with this change; GPU busy 4.8 to 4.15 ms a frame
- Hyrule Field: 101.1 and 113.1 fps before; 131.5 and 131.3 after; GPU busy 6.5 to 4.7 ms
- merged a frame: 20 to 29 MiB before, 1.7 to 2.0 MiB after
- faults on armed pages in 1,000 frames: 14 and 16 before, 17 to 18 after in the village; 19 and 23 before, 17 to
  21 after in Hyrule Field; time spent fetching the same (4 to 5 ms in 1,000 frames)
A run with synthetic reads of four armed ranges per log window (wb_lazy_touch = 4, 769 faults) ended without an
error. The picture of every presented frame of a village walk and a field walk (thumbnails, 10,131 and 9,531
frames) had no black frame. The village walk has single frames whose brightness jumps: 14 with the change and 28
without it in about 10,000 frames each; the two looked at, one of each, are jumps of the camera.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 16:44:02 -05:00
chrg6
f17f4f9bcc Decode the game's zstd files on the host by default on 1.1.1 as well.
That build kept the game's own decoder until its sites had been through a check run. The check has now run over a village and a field: 3,072 streams and 31.7 GiB of output compared byte for byte, none different. The game's decoder took about a fifth of the running time of the game's threads in a village and about a second of each area load. zstd_native = 0 keeps the game's decoder.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 15:33:28 -05:00
chrg6
fc71691e91 Merge branch 'claude/amd-fight-crash-ef03a9' into test-all-fixes 2026-11-10 15:22:07 -05:00
chrg6
5151a64dbc Merge branch 'claude/pipeline-list' into test-all-fixes 2026-11-10 15:22:06 -05:00
chrg6
62e8578aa2 Turn the lazy write-back off by default on every card of AMD's own Windows driver, and let the device probe print the fatal summary line.
The default was off only on AMD's driver branch for RDNA 1 and 2, on the grounds that builds with the default on had been measured on an RX 7900 XT. Those measurements were standing in Hyrule Field and walking through a village; the report this is about is of a game that ended in fights, on an AMD card the player did not name, and fights are where the lazy path is busiest. Nobody has played one on an AMD card under our eyes on either branch, so the exception is now AMD's own driver as a whole (wb_lazy_choice.h, by VkDriverId), until a card of each branch has played fights with wb_lazy = 1 and a log to show. The cost is the open-area frame rate of 0.7.0 on every AMD card under Windows; the RX 7900 XT figures in the draft notes come from builds that had it on by default, which the notes now say. wb_lazy = 1 still turns it on.

--probe-vkdevice prints the line a fatal error or a crash adds to the log, so its wording can be read without a failing run. On an RTX 5090: "fatal: gpu NVIDIA GeForce RTX 5090, driver NVIDIA "621.01" version 621.1.0.0; memory mode mirror, lazy write-back on (the default on Windows (wb_lazy = 0 turns it off)); texture detail not chosen yet (no texture made)", and with the reporting player's nx2navi.ini beside the exe "lazy write-back off (wb_lazy = 0 is set)", after a start whose log names async_compute, vram_safe_mode and exception_handler as no settings of this version.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 15:20:47 -05:00
chrg6
caf50ecf1c Turn the lazy write-back off by default on AMD's driver for RX 5000 and 6000 cards, and say in the log which settings a run had.
A player on an AMD card reported crashes in fights and that a line added to nx2navi.ini cured them. That line, async_compute = 1, is no setting of the game, and neither are two more in the same file; the one line of the file that changes how memory reaches the GPU is wb_lazy = 0. A second report, an RX 6600 on 0.7.1, lost its device 3.5 minutes into play with the lazy write-back on. Neither proves it the cause: the player changed several lines at once and sent no log, the RX 6600's last pass mark is inside a draw pass and not a fetch, cards with 8 GB lost their device the same way on NVIDIA before the lazy write-back existed, and reading the fault handler, the fetch, the arming and the buffers' lifetimes found no failure path. What depends on the driver and cannot be checked without the card: pages of guest memory are inaccessible while a host-pointer import of them is bound and written by the GPU, and Vulkan calls run inside the access-fault handler.

The default is now decided once the driver is known (wb_lazy_choice.h, vk_memory.c vk_mem_lazy_decide): on under Windows as before, except on AMD's driver branch for RDNA 1 and 2, which none of us has run it on; AMD's driver for RDNA 3 and newer keeps it on, since builds with that default were measured on an RX 7900 XT. What the player sets wins everywhere, and one line at start says the choice and its reason. On those cards open areas run at the 0.7.0 frame rate again (on an RTX 5090 that was 65 fps against 90 in Hyrule Field; not measured on AMD). tests/wb_lazy_check.c checks the choice by platform, driver and setting without a GPU.

For reports of this kind the log now holds what was missing. At start: every setting nx2navi.ini sets, and in one line the keys that are no settings of this version and do nothing; the settings menu shows those in its bottom row. At a fatal error or a crash: one line with the card, its driver, the memory mode, the lazy write-back with its reason, armed ranges, fetches so far and a fetch in progress, and the texture detail (dispatch.c rt_fatal_note_add, vk_device.c fatal_note), since a report is often the end of a log alone.

--probe-cfg loads the player's file: the three lines are named, and the menu's save keeps them and every other line it does not change byte for byte. Nothing here ran on an AMD card, and no fatal error was provoked to read the new line.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 15:14:25 -05:00
chrg6
7b35fe7504 Let the limit on released textures follow the card's memory as well as the PC's.
The limit was a thirty-second of the installed memory, 1 GiB on a PC with 32 GB whatever its card. The game lets go of the textures of what leaves the view and takes them back when it returns: walking through a village the released ones passed 1 GiB within two or three minutes, and from then on older ones were dropped and decoded again on their way back. The limit is now also a quarter of the card's budget where the PC has an eighth of its memory to match, up to 8 GiB; the video memory and commit pressure rules still drop released textures first.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 14:57:51 -05:00
chrg6
87551c7618 An update that carries the shaders over builds the kit's current pipeline list, and a GPU without mesh shaders skips the list's mesh pipelines.
The carry path of an update does not run the shader preparation: it runs --warm-shaders, which builds the listed pipelines from the seed-pipeline-keys.bin the game folder already has. make_game.py now writes that file from the kit's list first, so a newer list reaches an updated install. Its shader variants are not translated on this path; a record whose variant the folder has neither prepared nor met is left out and counted.

On a GPU that has no mesh shaders every mesh record failed with a log line each. They are counted in the step's line instead.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 14:56:56 -05:00
chrg6
918d399d55 Add the shared pipeline list: mesh pipelines players met are built at the shader preparation, before the first start.
The preparation translates every shader and builds each one's stage library, so a pipeline a draw meets for the first time is linked in 2 to 6 ms. A mesh pipeline has no stage libraries: on an install that never drew it, it compiles whole in play, 0.2 to 1.1 s with its draws left out of the picture (18 of them on a first visit of Kokiri Forest). Whole pipelines were built ahead only from the folder's own pipeline-keys.bin.

tools/pipeline_list.py merge folds the pipeline-keys.bin and shader-answers.bin of game folders that were played into one list file: every key record once, with the answers record of each shader it names. A record names its shaders by input id, which depends on the game's code and the answers alone, and holds the game's pipeline state; what depends on the GPU is derived when the pipeline is built, so a record made on one machine is good on another with the same game version and runtime. Four folders on this PC give 5,432 records, 317 of them mesh pipelines (147 KB of keys).

tools/prepare_shaders.py reads the list (tools/data/pipeline_list.bin, data/pipeline_list.bin in a release folder, or --pipeline-list), adds its shader variants to the seed and writes its records as seed-pipeline-keys.bin. nx2navi.exe --prepare-shaders builds the mesh pipelines among them after the seed is translated and saves the pipeline cache before the stage libraries are built (NX2NAVI_PREPARE_PIPELINES=all builds the vertex pipelines too); --warm-shaders does it again after an update or a new driver, and --prepare-pipelines is that step alone. The pipelines go into the caches and are not kept or added to the folder's own key file, so they cost no start-up time or video memory until a draw meets them. A marker file keeps a driver that crashes on one of them from failing every preparation.

The list is derived from game data and stays out of the repository: check_content.py refuses files recorded from a game folder by their first bytes, and make_kit.py, make_launcher.py and make_release.py carry a list given with --pipeline-list. All three now also carry pipeline_list.py, which prepare_shaders.py imports.

Checked without the game: tests/pipeline_list_test.py, and prepare_shaders.py --no-run on the game files with a list of 317 mesh pipelines (all 266 shader variants found in this game version, 177 new to the seed). The build step in nx2navi.exe compiles and links and has not been run.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 14:52:46 -05:00
chrg6
682d9c5c50 A draw does not wait for the fast link of a mesh pipeline that is known to be compiling.
A draw whose pipeline is missing waits up to 3 ms for its fast link before it is skipped, 6 ms a frame in all. A mesh pipeline has no stage libraries and gets no fast link, so for the 0.2 to 1.1 s of its compile every frame that drew with it lost those milliseconds and skipped the draw all the same. Once such a pipeline has been building for longer than the 8 ms a cached one needs, its draws are skipped at once, as the one-shot mesh draws already are.

Not run in the game yet.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 14:52:25 -05:00
chrg6
e8f5f8e01e Merge branch 'main' of https://github.com/tangerineturtle1/NX2navi into test-all-fixes
# Conflicts:
#	runtime/cfg.c
2026-11-10 14:27:47 -05:00
chrg6
2db1ce835d Merge branch 'post-0.7.1-fixes' into test-all-fixes 2026-11-10 14:22:52 -05:00
chrg6
834b36a6d9 Stamp a retired texture as used when it is taken back.
While retired it got no use stamp, so it counted as idle the moment it came back, and a sweep over the video memory budget could evict it as idle in the same call that took it back: on the cards that are over budget, which are the ones the take-back is for.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 14:22:52 -05:00
chrg6
0e5e40b620 Merge branch 'claude/linux-old-headers' into test-all-fixes 2026-11-10 14:18:29 -05:00
chrg6
61c4c873d5 Merge branch 'claude/pipeline-stutter-110e1b' into test-all-fixes
# Conflicts:
#	docs/building.md
2026-11-10 14:18:29 -05:00
chrg6
c98a918a34 Merge branch 'claude/pr-intake-4-947f9e' into test-all-fixes
# Conflicts:
#	runtime/vk_texture.c
2026-11-10 14:18:21 -05:00
chrg6
be136d8d3b Credit a sampled texture through the slots that follow it, and take back a retired texture that cannot be destroyed.
A mark on a pool slot without an entry of its own went to the slot's own key, so a texture sampled only through a second descriptor at its address counted as unsampled, was retired after two minutes under memory pressure, and a mark on its stand-in could not take it back. Where the pool's syncs kept leaving slots behind, a retired texture was never destroyed either and its slots kept the stand-in: on a card with 6 GB the ground of the field was drawn without its textures and its height. The mark now goes to the textures the slot follows, and a retired texture that is not due after five seconds is taken back.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 14:18:02 -05:00
ordinarybob
4a954a1c6d Refresh the texture pool from a list of its entries instead of walking every slot, for about 7 % more frames a second where the replay thread is the limit.
Each submission's first sync of the bindless texture pool walked every registered slot, 33,000 in Kokiri Forest, to find the few textures the CPU wrote and the few slots that changed. Three parts of that walk change, each with its own switch (pool_fast, pool_presnap, pool_group; on by default, 0 and ab accepted):

- The freshness check goes through a list of the entries the slots resolved to, with a check by 2 MiB region first; slots that resolved to nothing are looked up again when the game's memory map changes and every eighth submission. The use stamps the idle eviction reads are left from a count of the slots that hold each entry.
- The check takes the write-watch snapshots of its regions first, so unchanged entries pass the region test.
- The GPU-written slots of one address bucket with one descriptor form a group, which the dirty walk leaves alone while its first member is unchanged.

Measured on an RTX 5090 at 2560x1440, standing in Kokiri Forest, in-run A/B over five runs with all three switched (three of them with the barrier change switched too): 93.6 to 100.7, 96.4 to 103.1, 95.8 to 102.6, 99.3 to 107.1 and 91.9 to 99.9 frames a second, the replay thread 0.7 to 0.8 ms a frame shorter; a control's blocks differ by 0.4 %. On a walk with NX2NAVI_POOL_GROUP_CHECK=1 no group that was left alone had missed a write, and screenshots across the A and B blocks show the same picture.

From pull request #390. Not tried: the texture eviction under memory pressure (a 9 GiB budget was never reached on the walk), AMD and Intel drivers, Linux. One of the nine runs with it on ended by itself with exit code 0 and nothing in its log two seconds after the window lost the focus; three repeats of that run did not.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 14:09:21 -05:00
ordinarybob
ce206a9239 Order passes and dispatches by the game's own barriers on NVIDIA graphics cards, which takes about a tenth off the GPU's time for a frame.
Every pass begin and every compute dispatch recorded a full barrier, which makes the GPU finish everything before each of them. With barrier_game = 1 (NX2NAVI_BARRIER_GAME) a pass after an earlier pass gets a barrier for its attachments only and a dispatch after other dispatches gets none; a change between drawing and compute keeps the full barrier, and so do the game's own barriers and the replay's copies, clears, fills and uniform updates. On an RTX 5090 at 2560x1440 the GPU's time for a frame fell from 5.4 to 4.9 ms standing in Kokiri Forest (two in-run A/B runs, the control's blocks differ by 0.7 %) and from 5.1 to 4.4 ms on a walk through it. The frame rate did not change there, because the replay thread is the limit on that machine; it should rise where the GPU is.

The default is 1 on NVIDIA graphics cards only, and 0 on every other vendor's: the game placed its barriers for an NVIDIA GPU, Vulkan asks for more of them than the game recorded, and no AMD or Intel driver has run this. 34 screenshots across the A and B blocks of one run show the same picture on NVIDIA. NX2NAVI_TRACE=perf logs the barriers a frame by kind.

From pull request #389, which had it on for every vendor; the default by vendor and the settings text are added here. The validation layer's synchronization check could not be run on this machine.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 14:08:48 -05:00
epicmeme546
da0d0be700 Read occlusion query results in batches, and wait in the submit call for the game's signal that the frame's data is written.
Both are about the time between two frames. Ocarina waits for the previous frame's fence before it lets the next frame's GPU work start: its main thread signals an event from the CPU, and the main queue's commands begin with a wait for it.

The completion thread ran a submission's callbacks before it let the submission's waiters go, and the occlusion counts were among them: one vkGetQueryPoolResults and one vkResetQueryPool per reported query, about a hundred queries a frame in Hyrule Field. On an RX 7900 XT (26.10.1) each call goes into the kernel, 3.9 ms of every frame, so the fence was reached 4 to 6 ms after the GPU had finished with every game thread waiting. Slots come from the ring in order, so one call now reads up to 256 slots from the one asked for, the available results are kept for the segments that follow, and the slots taken are reset in runs (vk_exec_xfer.c oq_take, NX2NAVI_OQ_BATCH). Standing in Hyrule Field at 3440x1440 with FSR: 52.7 fps against 44.6, the longest frame of the measured minute 46 ms against 55; the completion thread spends 0.3 ms a frame in the callback.

When the CPU's signal came a millisecond or two after the main queue's submit call, the submission took its copies, stopped at the wait and was held; the render thread polled the event, presented, and the present replayed the frame under a value the game's fence does not stand for. The next frame then skipped its wait for the GPU and was replayed inside its submit call, and frames of 23 and 15 ms took turns. A submission that begins with a wait for an event the CPU has signalled before now waits for the signal in the submit call, before its copies are taken, 10 ms at most (vk_exec.c vk_exec_gate_wait, NX2NAVI_GATE_WAIT_MS): 54.5 fps and a longest frame of 33 ms in the same place, 58.6 fps with the recorder thread on. With a camera sweep the pictures went out 21.0 ms apart on average with a 99th percentile of 55 ms, against 25.4 and 64. A village went from 48.6 to 53.7 fps and a graveyard from 56.5 to 64.2. 900 consecutive frames of a still scene show no dark or changed frame, with the recorder thread and without.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 13:37:12 -05:00
chrg6
df0db643ed Add a test switch that lets a new depth image take the live one's depth.
A contributor traced the one-frame dark shadowing at the castle town gate to depth targets the game remakes at other sizes about a thousand times a minute: the new image started cleared, so everything near the camera read as in shadow for a frame. With depth_alias = 1 a depth-only image of the same format copies the live one's depth for the region both cover. Their three runs each way showed 6, 4 and 0 dark frames without it and none with it; the cost of the copies and other scenes are not measured yet, so it is off by default.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 13:19:51 -05:00
chrg6
0a502ecf65 Let the pick-up window's number count at least one a frame, so the window closes above 60 fps.
The window that shows a pick-up above the character counts its number up to the new total by a whole step a frame and closes once they are equal. The game rounds that step down, and below one 60 Hz frame per update it was 0: above 60 fps the window stayed until the area was left, and for larger amounts the counting sound was started again every frame. The step is now floored at 1, switchable with pickup_count.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 13:10:34 -05:00
epicmeme546
e38f42a959 Test for NaN on the bits of the number on AMD's driver for RX 5000 and 6000 cards, so that no shader compiler can leave a test out.
The reports of black and white blocks on Navi, on Link next to her and on bright surfaces seen from inside a house come from RX 5000 and 6000 cards, whose Windows driver compiles with "AMD proprietary shader compiler" (vk_caps.amd_sc); an RX 7900 XT (LLPC) draws the same scenes clean. The game's shaders test for NaN with OpIsNan in 48,717 of an install's 54,959 translated modules, 26 tests each on average, and the translator's guard that keeps NaN out of every float colour output, so that one bad texel does not spread through blur and bloom into blocks, is such a test too. SignedZeroInfNanPreserve asks the driver to keep those tests; whether that compiler does is not known.

With that driver, vk_pipeline.c spirv_nan_bits now turns every 32-bit isnan, scalar or vector, into (bits & 0x7fffffff) > 0x7f800000 on the number's bits, and a float select with a constant whose bits are a NaN into a select of the bits (issue #156), when the module is created. NX2NAVI_VK_NAN_BITS=1 does it on every driver, 0 on none. Run over every module of an install's shader cache, the rewrite changed 1,281,675 tests and 57 selects with every output well formed (each id defined once, below the bound, the uint types declared once); --selftest checks it on a small module, and Link's house with it forced on (RX 7900 XT) draws as without it. Not yet tried on an RX 5000 or 6000 card.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-10 13:08:03 -05:00
Chris Toph
5374d04d0f Keep an optional history of GPU address ranges, so the report of a lost graphics device can name what held the faulting address.
After a lost device the driver's fault report gives an address and nothing says what was there. With va_report = 1 (NX2NAVI_VA_REPORT, off by default) the game keeps every address range the driver binds to a buffer or image and every range it unbinds (VK_EXT_device_address_binding_report, runtime/vk_vaddr.c). The report of a lost device then names the ranges that covered each faulting page, or the nearest ones when none did, and vaddr-history.txt beside the log holds the whole history, also when the driver gives no fault report. With 2 the call stack behind each range of 1 MiB or more is kept too. A player sees nothing of this unless the setting is on: without it no extension is enabled and no callback runs.

From pull request #349. Its second part, which stops a frame dump from turning on mesh query statistics, is left out: it changes what a dump writes for everyone who uses one. Built for Windows; no device was lost to try it.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 12:50:56 -05:00
Michael Schienbein
541c4ff84c Start the game again after Restart on every processor it had (Linux).
On Linux each Restart from the menu cost the next game one more processor core on machines with 12 or more logical processors. A process started with posix_spawn begins with the mask of the thread that spawned it, and since the recording thread has a core of its own no thread has every processor: the restarted game began without that core and took the next fast one for its own recording thread. plat_restart (runtime/platform/linux.c) now widens the calling thread's mask to all of the process's processors for the time of the spawn and puts it back. A mask set from outside is kept, helper programs started with plat_spawn are unchanged, and Windows is not touched.

From pull request #435, with its test tests/platform_restart_mask_test.c, which needs Linux. Not compiled or run here: this machine has no Linux compiler.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 12:49:10 -05:00
Michael Schienbein
919aa7b2a8 Let the mouse camera in stick mode reach the game's top turn speed at a camera speed above 1.
With the mouse camera in stick mode and a camera speed above 1, the mouse could never turn faster than a full stick does at camera speed 1, however high the mouse sensitivity was set. The mouse's counts were limited to full tilt first and divided by the camera speed afterwards, so at camera speed 1.5 the most the mouse could send was two thirds of full tilt, which the game's own factor brought back to the old top speed. The division now comes before the limit (runtime/input_map.c input_mouse_stick): a slow movement turns as far as before, and a fast flick reaches full tilt, which is the game's top speed with the camera speed in it. Direct mouse mode, the default, does not use this code. A camera speed of 1 or less behaves as before.

From pull request #424. tests/mouse_stick_check.py is new here: it builds input_map.c alone and checks the deflection at each camera speed without the game.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 12:48:50 -05:00
chrg6
61cd4784f8 Build the Linux kit against pinned Vulkan headers and without #embed, and try the compiler before the long steps.
A player's AppImage build of 0.7.1 stopped in its compile step after eight minutes of RomFS work: four
files did not compile, because the kit builds with the system's compiler and headers and theirs were
older than the runtime needs. vk_pipeline.c and vk_device.c name VK_KHR_pipeline_binary,
VK_KHR_dynamic_rendering_local_read and VK_KHR_shader_maximal_reconvergence, which their Vulkan headers
did not declare, and ui/menu.cpp takes its font in with #embed, which clang has from version 19.

- Vulkan headers: tools/vulkan_pin.py holds the release the other builds use (1.4.363, the archive and
  hash tools/make_release_linux.py already pinned, which now reads them from there). The AppImage carries
  its C headers in toolchain/vulkan-headers and builds against them whatever the system has; a kit or a
  source checkout uses the system's when they are at least that release and fetches the pinned ones
  otherwise (build_game.py vulkan_include). Their folder goes first on the include path of the runtime's
  sources only, so the lifted code's objects keep their keys, and a runtime object compiled before
  against the system's headers is taken over instead of compiled again. AppRun no longer asks for the
  system's Vulkan headers.
- #embed: tools/build_shaders.py also writes the font's bytes as runtime/ui/roboto_ttf.h (ignored by git,
  carried by a kit), and menu.cpp includes that where the compiler has no __has_embed. With #embed
  nothing changes.
- Before its long steps a Linux build that will compile now checks every runtime source that needs no
  header of the install with the compile step's own command (clang -fsyntax-only) and links an empty
  program against the game's libraries (make_game.py preflight): an old compiler, an old header or a
  missing development package stops it in the first seconds with the compiler's own errors. A failed
  compile names every file that failed, not only the three whose errors it prints.

Checked on Windows: tests/launcher_test.py passes with four new tests (117); the header of bytes and
#embed give the same 162,588 bytes, and both branches of menu.cpp compile; the early check passes on the
Windows file set with clang 23 and stops on files made to fail; the 34 local headers of release 363 are
the upstream tag's files by git blob id. Not run: anything on Linux, the pinned archive's download, a
compiler older than 23.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
2026-11-10 12:42:17 -05:00