x86jit: Perform vector transfers instead of flushing to memory #18234
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
It's faster, this performs better.
|
It does seem like doing this on arm64 may have finally made a dent in arm64 FPS, but still not sure since a breeze can make it vary so much. -[Unknown] |
|
Indeed seems plausible that keeping stuff in registers is very important on ARM64: x86 are optimized for fast-ish spilling since they have to do it so much due to low register count, while ARM64 doesn't really have that problem so I guess isn't as much... |
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Outdated suggestions cannot be applied.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.
This attempts to optimally retain values in registers even when going from and to Vec4. For example, a
lv.q; vdot.t; sv.qpreviously caused the Vec4 to spill and each component to be loaded back, then to spill again and load as a vec4 for the store.With this change, the values are kept in registers are reassembled if possible. This results in a 1-2% FPS improvement in LittleBigPlanet, for example. Could use more testing in other games, though. Remains to be seen if it'll manage to make any difference on arm64, but not done yet...
At this point LBP is around ~13% faster FPS than the old jit.
-[Unknown]