This is an archive of the discontinued LLVM Phabricator instance.

[MLIR][GPU][lld] Use LLD bundled in ROCm, removing workaround

Authored by krzysz00 on Feb 10 2022, 10:55 AM.



Having clarified that executing the SerializeToHsaco pass can
depend on a ROCm installation, switch from calling lld as a library to
using the copy of lld guaranteed to be included in a ROCm install.

This removes the workaround introduced in D119277

Diff Detail

Event Timeline

krzysz00 created this revision.Feb 10 2022, 10:55 AM
krzysz00 requested review of this revision.Feb 10 2022, 10:55 AM
whchung accepted this revision.Feb 10 2022, 11:01 AM

LGTM. This approach is used in TensorFlow XLA for ROCm, where XLA build may not have lld configured, and thus must rely on the external lld binary in a ROCm installation.

This revision is now accepted and ready to land.Feb 10 2022, 11:01 AM