Make the wait on the GPU's signal the default again after a frame's first submission.

The fence's wait cost about 4.5 % of the frames a second, and the player who reported the flicker still sees it and black frames with it. wb_split = 0 keeps the fence's wait for whoever wants to compare.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
This commit is contained in:
chrg6
2026-11-11 05:18:37 -05:00
parent b55d2eb3e1
commit f38428fa4d
3 changed files with 5 additions and 4 deletions

View File

@@ -74,8 +74,9 @@ static int hold_max = -1;
// ---- the executed timeline
// The wait after a frame's first submission (nvn_api_core.c nvnQueueSubmitCommands) is there so that the GPU has run
// that submission's passes before the game rewrites what they read. It is the fence's wait (NX2NAVI_WB_SPLIT=0, the
// default). NX2NAVI_WB_SPLIT=1 makes it a wait on exec_tl, a second timeline that reaches a submission's value when
// that submission's passes before the game rewrites what they read. With NX2NAVI_WB_SPLIT=0 it is the fence's wait.
// The default is 1 again: the fence's wait cost about 4.5 % of the frames a second and the flicker it was put in
// against was still reported with it. NX2NAVI_WB_SPLIT=1 makes it a wait on exec_tl, a second timeline that reaches a submission's value when
// the GPU has executed its own commands: the replay ends the part before the write-backs (vk_exec.c,
// vk_sync_split_exec; vk_memory.c vk_mem_mirror_end: what the shaders wrote into VRAM mirrors goes back to guest
// memory), that part signals exec_tl, and the write-backs follow in a part of their own, which signals the timeline