This is an archive of the discontinued LLVM Phabricator instance.

[AArch64] LoopVectorizer test case for cost model to handle signed division by a power of 2
ClosedPublic

Authored by suyog on Nov 8 2014, 1:42 AM.

Details

Summary

This is addition to r216371 (SLP and Loop Vectorization) and r218607 where
cost model for signed division by power of 2 was improved for AArch64.

The revision r218607 missed test case for Loop Vectorization.
Adding it here.

Please help in reviewing this patch. I will commit it after the review.

Regards,
Suyog

Diff Detail

Repository
rL LLVM

Event Timeline

suyog updated this revision to Diff 15953.Nov 8 2014, 1:42 AM
suyog retitled this revision from to [AArch64] LoopVectorizer test case for cost model to handle signed division by a power of 2.
suyog updated this object.
suyog edited the test plan for this revision. (Show Details)
suyog added a subscriber: Unknown Object (MLST).
mcrosier accepted this revision.Nov 10 2014, 7:35 AM
mcrosier edited edge metadata.

LGTM with one small nit.

test/Transforms/LoopVectorize/AArch64/sdiv-pow2.ll
9 ↗(On Diff #15953)

I'd prefer a space between the semicolon and the CHECK directive.

This revision is now accepted and ready to land.Nov 10 2014, 7:35 AM
suyog closed this revision.Nov 10 2014, 11:30 PM
suyog updated this revision to Diff 16029.

Closed by commit rL221674 (authored by @suyog).

Thanks Chad. committed in r221674 with space as per your review.