mirror of
https://github.com/tangerineturtle1/NX2navi.git
synced 2026-11-15 17:09:57 -04:00
7a8bbad47e
The user asked for it: the mode no longer gains frame rate, and the mesh shader fallback covers the cards it was written for. Its switch in the settings menu's Game tab is gone with the rows under it, and so are their 57 text ids and strings. low_end_gpu itself does nothing any more. It is still a known key, so an older nx2navi.ini gives no unknown-key warning, and while it is on the log says once that the mode was removed in 0.10.2. The preset that gave nine keys other defaults is gone, and so is the rule that read the mode's keys only while it was on: each of them (mesh_emulation, tex_clobber, foliage, vram_heap_limit, vram_reserve_mb, upload_batch_mb, mirror_stage, mirror_hot, submit_finish_gpu and the others) is an ordinary setting with the default of its row or of its reader, so a line a player wrote by hand keeps counting and a file with only the switch runs the game as it runs for everyone else. guest_opt in nx2navi.ini counts for the next rebuild without the switch too (tools/make_game.py). The recorder thread read the switch itself: on an AMD card without mesh shaders the mode turned it on. It is off on every AMD card now unless vk_recorder = 1 is set. nx2navi.ini.example lists the former mode's settings as single settings, the 0.10.2 notes have a Removed entry and name the foliage fix of the commit before this one, and the comments that said what the mode sets say so in the past tense. tests/cfg_parse_check has the new reading of the switch and of the keys without it; the test matrix is at 90 ok with the same two not ok as before (installer_test, worktree_test), and a clean build compiles 169 runtime files and links. Not run in the game. Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
60 lines
4.2 KiB
C
60 lines
4.2 KiB
C
// How the game's task and mesh programs are drawn (vk_device.c create_device). Header only, so that
|
|
// tests/mesh_path_check.c checks the decision without a device.
|
|
//
|
|
// native VK_EXT_mesh_shader pipelines
|
|
// fallback the mesh fallback: the programs run as compute passes and an ordinary draw reads what they wrote
|
|
// (vk_exec_draw.c "mesh fallback", vk_mesh_fb.c)
|
|
// emulation the mesh emulation (mesh_emulation = 1 on a device without mesh shaders, vk_exec_draw.c); never on an
|
|
// NVIDIA card (below)
|
|
// none a device without mesh shaders with both off: mesh draws are skipped
|
|
//
|
|
// mesh_fallback = auto (the default) takes the fallback on a device without mesh shaders, and on AMD's Windows driver
|
|
// for RDNA 1 and 2 (vk_caps.amd_sc), which has them on RDNA 2. That driver lost the device ("No fault detected", a
|
|
// timeout) minutes into play on RX 6600 to RX 6950 XT cards in every version from 0.6.0 to 0.9.2 (0.10.x built its mesh
|
|
// pipelines without the driver's optimiser, NX2NAVI_VK_NOOPT, which no reporter has confirmed), with the pipeline
|
|
// library, the replay workers, the asynchronous replay, the lazy write-back and the low-end GPU mode each on and off.
|
|
// Where the pass marks named the pass the GPU was in, it was one that starts with a native mesh draw. The same driver
|
|
// on RDNA 1 has no mesh shaders and draws through this fallback, and AMD's driver for RDNA 3 and 4 (LLPC) draws them
|
|
// natively; neither has such a report. That the native mesh path is the cause is not shown: the fallback is what the
|
|
// marks leave to try, at the cost of a compute pass and a pass break per mesh draw, and of mesh programs that are
|
|
// left out of the picture until their compute passes are built, at their first sight in a run. mesh_fallback = 0
|
|
// draws them natively again; 1 takes the fallback on every device.
|
|
//
|
|
// The mesh emulation is the older of the two ways for a card without mesh shaders, and the low-end GPU mode turned it
|
|
// on up to 0.10.1 (cfg.c lowend_set). On NVIDIA's cards without mesh shaders (GTX 10 and older) it ended the game
|
|
// inside the driver at the first mesh draws of an area: an access violation at nvoglv64.dll+0x10cdffe on a GTX 1050 Ti
|
|
// and on a GTX 1070, both on 0.10.1, and a Titan X (Pascal) report of the same mode. The GTX 1050 Ti ran on once
|
|
// mesh_emulation was set to 0, through the fallback, with the rest of that mode still on. So an NVIDIA card never
|
|
// takes the emulation: with it asked for, the card is drawn for as if it were not.
|
|
#pragma once
|
|
#include <string.h>
|
|
|
|
enum { MESH_PATH_NATIVE, MESH_PATH_FALLBACK, MESH_PATH_EMULATION, MESH_PATH_NONE };
|
|
|
|
// has_mesh: the device has VK_EXT_mesh_shader with task and mesh shaders; amd_sc: AMD's Windows driver for RDNA 1 and
|
|
// 2; nvidia: an NVIDIA card; fallback: the mesh_fallback setting ("auto", "0", "1"; anything else reads as auto);
|
|
// emulation: mesh_emulation ("0" off). *why: the reason, for the boot line.
|
|
static inline int mesh_path_decide(int has_mesh, int amd_sc, int nvidia, const char *fallback, const char *emulation,
|
|
const char **why) {
|
|
const char *w;
|
|
int path;
|
|
if (!has_mesh && strcmp(emulation, "0") && !nvidia)
|
|
path = MESH_PATH_EMULATION, w = "mesh_emulation = 1 on a device without mesh shaders";
|
|
else if (!strcmp(fallback, "1"))
|
|
path = MESH_PATH_FALLBACK, w = "mesh_fallback = 1";
|
|
else if (!strcmp(fallback, "0"))
|
|
path = has_mesh ? MESH_PATH_NATIVE : MESH_PATH_NONE, w = "mesh_fallback = 0";
|
|
else if (!has_mesh)
|
|
path = MESH_PATH_FALLBACK, w = strcmp(emulation, "0") ? "this device has no mesh shaders; mesh_emulation is "
|
|
"not used on an NVIDIA card, where it ended the game "
|
|
"inside the driver"
|
|
: "this device has no mesh shaders";
|
|
else if (amd_sc)
|
|
path = MESH_PATH_FALLBACK, w = "the default on AMD's driver for RDNA 1 and 2, where native mesh pipelines lost "
|
|
"the device in play; mesh_fallback = 0 draws them natively";
|
|
else
|
|
path = MESH_PATH_NATIVE, w = "VK_EXT_mesh_shader";
|
|
if (why) *why = w;
|
|
return path;
|
|
}
|