fix(ssa): Check arrays used as a terminator argument in the RC tracker #7165
Merged
aakoshh merged 13 commits intoaf/fix-11249from Jan 23, 2025
Merged
fix(ssa): Check arrays used as a terminator argument in the RC tracker #7165aakoshh merged 13 commits intoaf/fix-11249from
aakoshh merged 13 commits intoaf/fix-11249from
Conversation
vezenovm
changed the title
vezenovm
changed the title
Contributor
Changes to Brillig bytecode sizes
🧾 Summary (10% most significant diffs)
Full diff report 👇
|
vezenovm
changed the title
Contributor
Changes to number of Brillig opcodes executed
🧾 Summary (10% most significant diffs)
Full diff report 👇
|
5 tasks
TomAFrench
force-pushed
the
af/fix-11249
branch
from
f1342c9 to
ec002e2
Compare
5 tasks
vezenovm
commented
Jan 23, 2025
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
⚠️ Performance Alert ⚠️
Possible performance regression was detected for benchmark 'Compilation Time'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.20.
| Benchmark suite | Current: 55b75cf | Previous: 2d415ca | Ratio |
|---|---|---|---|
rollup-block-root-single-tx |
110 s |
89.7 s |
1.23 |
This comment was automatically generated by workflow using github-action-benchmark.
CC: @TomAFrench
aakoshh
force-pushed
the
mv/check-ref-containing-array-rc-tracker
branch
from
dd04047 to
b8b5935
Compare
Contributor
Changes to circuit sizes
🧾 Summary (10% most significant diffs)
Full diff report 👇
|
Contributor
|
Looking at the increase in the opcodes executed, one difference between this and mine is that I only did the special treatment during the preprocessing, but when the functions were fully inlined it was removed. Do you reckon something like that for the final pass (maybe piggybacking on the |
Contributor
Author
Ah yes good call out. I think we can do that as a follow-up optimization as we want will want to have some better tests around under what circumstances we do not check the terminator (e.g. are we actually fully inlined?). For now, let's just get this fix in. |
Contributor
|
I tried commenting out preprocessing and the terminator marking separately and together and they did not make a difference in the output of this: |
Contributor
|
It's probably down to the change in the change in rules in which functions to preprocess. |
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.
Description
Problem*
Resolves failures in AztecProtocol/aztec-packages#11294 from the preprocessing pass. #7163 is a bit too conservative with marking that we should keep the inc rcs on any parameters. Doing so happens to also fix the test case, but marking arrays passed to terminators is the more accurate fix for how the RC tracker should be looking at mutated arrays.
Summary*
We want to make sure we do not remove inc rcs on arrays that are potentially mutated elsewhere. To do so, we check whether an array was passed to a store instruction or a call. We used to only run DIE (which removes these RCs) at the end of the SSA optimizations where we were aggressively inlining.
Now that we are starting to process functions in isolation we triggered a bug as outlined here AztecProtocol/aztec-packages#11294 (comment). We need to account for whether an array is used in a terminator argument such as a return as that returned array may be mutated later.
Additional Context
Will need to add a new test making sure we do not remove an inc rc on an array type used in a terminator argument.
Documentation*
Check one:
PR Checklist*
cargo fmton default settings.