================
@@ -824,9 +831,15 @@ SIAtomicAddrSpace
SIMemOpAccess::toSIAtomicAddrSpace(unsigned AS) const {
return SIAtomicAddrSpace::OTHER;
}
+// TODO: Consider moving single-wave workgroup->wavefront scope relaxation to
an
----------------
Pierre-vh wrote:
I mean it doesn't lose any semantics exactly, it's more about intent I guess.
But I think it's more of a personal preference than a logical argument. I see
this demotion more as a relaxation of the code sequences than an actual
syncscope change, but I think the majority sees it the other way.
If we want to do the lowering in the IR, then I won't oppose.
https://github.com/llvm/llvm-project/pull/224065
_______________________________________________
llvm-branch-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
