[SILOptimizer]: slow OSSA lifetime canonicalization mitigation #82313
Conversation
|
@swift-ci please smoke test |
|
@swift-ci please test compiler performance |
|
@swift-ci Please smoke test compiler performance |
1 similar comment
|
@swift-ci Please smoke test compiler performance |
There was a problem hiding this comment.
Thank you for taking a stab at this @jamieQ ! I added a few comments.
OSSA lifetime canonicalization can take a very long time in certain cases in which there are large basic blocks. to mitigate this, add logic to skip walking the liveness boundary for extending liveness to dead ends when there aren't any dead ends in the function. Updates `DeadEndBlocks` with a new `isEmpty` method and cache to determine if there are any dead-end blocks in a given function.
jamieQ
force-pushed
the
copy-prop-ossa-dead-ends-build-time-fix
branch
from
f675cc7 to
1f3f830
Compare
|
@swift-ci please smoke test |
|
@swift-ci Please smoke test compiler performance |
|
When you're ready, @jamieQ, please merge this and cherry-pick it to release/6.2 . |
@nate-chandler happy to get the 6.2 release PR going, but i'll need your (or someone's) help in merging as i do not have commit access. the Linux tests appear to have failed for unrelated reasons, but i'll retry them once more. |
|
@swift-ci Please smoke test Linux platform |
|
@swift-ci Please clean smoke test Linux platform |
|
@swift-ci Please smoke test Linux platform |
OSSA lifetime canonicalization can take a very long time in certain cases in which there are large basic blocks. to mitigate this, add logic to skip walking the liveness boundary for extending liveness to dead ends when there aren't any dead ends in the function.
resolves: #82317