This is an archive of the discontinued LLVM Phabricator instance.

[X86] Fix SmallVector sizes in constant pool shuffle decoding to avoid heap allocation
ClosedPublic

Authored by craig.topper on Feb 26 2017, 6:28 PM.

Details

Summary

Some of the vectors are under sized to avoid heap allocation. In one case the vector was oversized.

Not sure if 64 uint64_ts is maybe too much stack space.

Diff Detail

Repository
rL LLVM

Event Timeline

craig.topper created this revision.Feb 26 2017, 6:28 PM
RKSimon accepted this revision.Feb 27 2017, 6:38 AM

LGTM - I wonder if there's any useful (i.e. not unnecessarily noisy) that we could get the compiler to report cases where SmallVector has exceeded its 'stack' size?

This revision is now accepted and ready to land.Feb 27 2017, 6:38 AM
This revision was automatically updated to reflect the committed changes.