Cast optimization improvement: also handle local.tee #6507
Merged
Conversation
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
Converts the following:
(some.operation
(ref.cast .. (local.tee $ref ..))
(local.get $ref)
)
into:
(some.operation
(local.tee $temp
(ref.cast .. (local.tee $ref ..))
)
(local.get $temp)
)
This removes close to 10% of the casts in a test program (from 38113
casts down to 36119).
I also tried to improve the optimization that moves more refined casts
earlier, but this does not seem very effective (it only eliminated 9
further casts), so I'm not sure it is worth it.
kripken
reviewed
Apr 17, 2024
| // blocks (TODO 2, though RedundantSetElimination does that as well). | ||
| // However, we should consider whether improving those other passes | ||
| // might make more sense (as it would help more than casts, if we could | ||
| // make them operate "backwards" and/or past basic blocks). |
Member
There was a problem hiding this comment.
The first TODO is still relevant, isn't it?
Contributor
Author
There was a problem hiding this comment.
Oh, it is indeed only partially done (with connectAdjacentBlocks set to true).
src/passes/OptimizeCasts.cpp
Outdated
| void handleRefinement(Expression* curr) { | ||
| auto* teeFallthrough = Properties::getFallthrough( | ||
| curr, options, *getModule(), Properties::FallthroughBehavior::NoTeeBrIf); | ||
| if (auto* set = teeFallthrough->dynCast<LocalSet>()) { |
Member
There was a problem hiding this comment.
Suggested change
| if (auto* set = teeFallthrough->dynCast<LocalSet>()) { | |
| if (auto* tee = teeFallthrough->dynCast<LocalSet>()) { |
Also the isTee check below is not needed, as a fallthrough LocalSet must be a tee (because only a tee flows out a value that can fall through to some place).
src/passes/OptimizeCasts.cpp
Outdated
| return; | ||
| } | ||
| updateBestCast(curr, get->index); | ||
| return; |
src/passes/OptimizeCasts.cpp
Outdated
| updateBestCast(curr, set->index); | ||
| } | ||
| } | ||
| auto* fallthrough = Properties::getFallthrough(curr, options, *getModule()); |
Member
There was a problem hiding this comment.
Suggested change
| auto* fallthrough = Properties::getFallthrough(curr, options, *getModule()); | |
| auto* fallthrough = Properties::getFallthrough(teeFallthrough, options, *getModule()); |
This saves a little work: we looked a little the first time, and the second time we don't need to re-do that work, and just continue from there.
kripken
approved these changes
Apr 18, 2024
Member
|
I tested this on a large Java program and it removed 3.2% of casts. Nice work @vouillon ! |
Closed
jonludlam
pushed a commit
to jonludlam/js_of_ocaml
that referenced
this pull request
May 15, 2025
This reverts commit 25315fb4d3231ecd0fdc07c57adec59f3e697dc3. No longer useful with WebAssembly/binaryen#6507
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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.
Converts the following:
into:
This removes close to 10% of the casts in a test program (from 38113 casts down to 36119).
I also tried to improve the optimization that moves more refined casts earlier, but this does not seem very effective (it only eliminated 9 further casts), so I'm not sure it is worth it.