[PATCH] D24623: [AMDGPU] Implement memory model

[Tony Tye via Phabricator via llvm-commits]

t-tye added inline comments.


================
Comment at: lib/Target/AMDGPU/SIMemoryLegalizer.cpp:277
+      ++MI;
+      Changed |= InsertWaitcntVmcnt0(MI);
+      Changed |= InsertBufferWbinvl1Vol(MI);
----------------
Is this needed only if the the instruction is a VMEM or FLAT instruction, not if a DS? It is only ensuring that the load has completed before doing the VMEM invalidate.


https://reviews.llvm.org/D24623