This is an archive of the discontinued LLVM Phabricator instance.

Reduce overhead of OMPT
ClosedPublic

Authored by jmellorcrummey on Oct 7 2015, 2:40 AM.

Details

Summary

Avoid computing state needed only by OMPT unless the ompt_enabled flag is set.

Properly handle a corner case in OMPT where team == NULL.

Diff Detail

Repository
rL LLVM

Event Timeline

jmellorcrummey retitled this revision from to Reduce overhead of OMPT.
jmellorcrummey updated this object.
jmellorcrummey added reviewers: jlpeyton, Hahnfeld.
Hahnfeld edited edge metadata.Oct 7 2015, 5:06 AM

LGTM. The change for __kmpc_for_static_fini may be applied to much more functions. I don't know whether this is relevant for the performance or the compiler is smart enough...

jlpeyton accepted this revision.Oct 9 2015, 10:42 AM
jlpeyton edited edge metadata.

LGTM as well.

This revision is now accepted and ready to land.Oct 9 2015, 10:42 AM
This revision was automatically updated to reflect the committed changes.