Use AVX2 gather for samplerjit #15275
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
From looking at assembly, just trying to keep it small.
Planning to expand this to more paths.
This should be rare, but a game could easily shove a CLUT4 texture at the end of VRAM, and then accessing the last index would segfault.
We only use it for DXT now, so let's not keep the dead code around.
|
That's a great result! Really cool. Think I've underestimated the gather instruction. |
2 tasks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Although VPGATHERDD is high latency, it's still much better than manually assembling the values, and especially than the old way with CALL.
This had some nice gains in several areas on an AVX2 supporting PC. VC3 in game is now at ~46 FPS, LBP at ~56 FPS, and KHBBS is ~74 FPS. The VC3 logo screen improved a lot though, up to ~99 FPS from ~93.
Didn't heavily test the non-AVX2 paths, though.
-[Unknown]