HomePhabricator

[Analysis] simplify code for scaleShuffleMask
Concern Raisedebf83c36e29a

Authored by spatel on Mon, Mar 23, 8:34 AM.

Description

[Analysis] simplify code for scaleShuffleMask

This is NFC-ish. The results should be identical, but perf is hopefully
better with the fast-path for no scaling. Added a unit test for that.

The code is adapted from what used to be the DAGCombiner equivalent
function before D76508 (rG0eeee83d7513).

Details

Auditors
jdoerfert
Committed
spatelMon, Mar 23, 8:47 AM
Parents
rGc57689bef261: [Attributor][NFC] Copy llvm::function_ref, don't use references
Branches
Unknown
Tags
Unknown

Event Timeline

jdoerfert raised a concern with this commit.Mon, Mar 23, 10:40 AM
jdoerfert added a subscriber: jdoerfert.

Could this be the reason for the broken bot? http://lab.llvm.org:8011/builders/clang-ppc64be-linux-lnt/builds/36387
It looks the most likely to me but you say it's "NFC-ish".

This commit now has outstanding concerns.Mon, Mar 23, 10:40 AM

Could this be the reason for the broken bot? http://lab.llvm.org:8011/builders/clang-ppc64be-linux-lnt/builds/36387
It looks the most likely to me but you say it's "NFC-ish".

Hmm...I can't rule it out though it seems unlikely. I'm happy to revert and see if it changes.
I've always been confused about bot logs wrt test-suite. This is saying that a test (MultiSource/Benchmarks/SciMark2-C/scimark2) execution time regressed, but we can't tell anything else from that log, can we?