jit: reuse shared-penalty physregs across preds#9696
Open
igogo-x86 wants to merge 2 commits into
Open
Conversation
When a block has several Vregs that share the same penalty vector, the per-Vreg penalty tweaks in color_block_initialize don’t enforce a consistent ordering. We can end up with the right physreg set but in a different order than predecessors, so extra edge copies get inserted. After coloring, reorder shared-penalty groups to match the ordering chosen by already-processed predecessors when they agree, keeping the pattern stable and avoiding those redundant copies. This is guarded to AArch64: without xchg, an inconsistent ordering forces a copy2 that lowers to three 4-byte instructions, whereas x86 would emit a single 2-byte xchg.
Contributor
|
@facebook-github-bot has imported this pull request. If you are a Meta employee, you can view this in D89387615. (Because this pull request was imported automatically, there will not be any future comments.) |
charles-typ
reviewed
Jan 4, 2026
| } | ||
| } | ||
|
|
||
| if (arch() == Arch::ARM) |
Contributor
There was a problem hiding this comment.
@igogo-x86 Thanks for making the PR!
Sharing internal reviews here:
Nitpick: Braces for if statement unless it can all fit on one line.
Contributor
|
@igogo-x86 has updated the pull request. You must reimport the pull request before landing. |
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
When a block has several Vregs that share the same penalty vector, the per-Vreg penalty tweaks in color_block_initialize don’t enforce a consistent ordering. We can end up with the right physreg set but in a different order than predecessors, so extra edge copies get inserted. After coloring, reorder shared-penalty groups to match the ordering chosen by already-processed predecessors when they agree, keeping the pattern stable and avoiding those redundant copies.
This is guarded to AArch64: without xchg, an inconsistent ordering forces a copy2 that lowers to three 4-byte instructions, whereas x86 would emit a single 2-byte xchg.