Page MenuHomePhabricator

Actually support volatile memcpys in NVPTX lowering
ClosedPublic

Authored by eliben on Jul 9 2015, 4:27 PM.

Diff Detail

Repository
rL LLVM

Event Timeline

eliben updated this revision to Diff 29410.Jul 9 2015, 4:27 PM
eliben retitled this revision from to Actually support volatile memcpys in NVPTX lowering.
eliben updated this object.
eliben added reviewers: jingyue, jholewinski.
eliben set the repository for this revision to rL LLVM.
jingyue accepted this revision.Jul 9 2015, 5:31 PM
jingyue edited edge metadata.

LGTM

More of a feature request, llvm.memcpy now takes alignments as arguments. We should emit larger loads/stores (e.g. st.u32 or even vector store) when alignments permit. Worth a TODO?

This revision is now accepted and ready to land.Jul 9 2015, 5:31 PM

LGTM

More of a feature request, llvm.memcpy now takes alignments as arguments. We should emit larger loads/stores (e.g. st.u32 or even vector store) when alignments permit. Worth a TODO?

Yes, will do.

This revision was automatically updated to reflect the committed changes.