This is an archive of the discontinued LLVM Phabricator instance.

[AMDGPU] Make generating cache invalidating instructions optional
ClosedPublic

Authored by piotr on Jul 23 2020, 10:28 AM.

Details

Summary

D78800 skipped generating cache invalidating instrucions altogether
on AMDPAL. However, this is sometimes too restrictive - we want a
more flexible option to be able to toggle this behaviour on and off
while we work towards developing a correct implementation of the
alternative memory model.

Diff Detail

Event Timeline

piotr created this revision.Jul 23 2020, 10:28 AM
nhaehnle added inline comments.Jul 24 2020, 6:44 AM
llvm/lib/Target/AMDGPU/SIMemoryLegalizer.cpp
50–52

amdgcn-skip-cache-invalidations?

LGTM with the option name fixed, and please adjust the variable name also (amdgcn-skip-cache-invalidations --> AmdgcnSkipCacheInvalidations).

piotr updated this revision to Diff 280466.Jul 24 2020, 8:08 AM

Renamed the option and variable.

This revision was not accepted when it landed; it landed in state Needs Review.Jul 27 2020, 12:53 AM
This revision was automatically updated to reflect the committed changes.