[PATCH] D35146: AMDGPU : Widen extending scalar loads to 32-bits

[Matt Arsenault via Phabricator via llvm-commits]

arsenm added inline comments.


================
Comment at: test/CodeGen/AMDGPU/widen_extending_scalar_loads.ll:48
+}
+
+attributes #0 = { nounwind }
----------------
arsenm wrote:
> Needs tests with half and <2 x half>, i1, and maybe another exotic size. I'm pretty sure this will assert for half  as is now. Also need tests with various alignments and volatile
Also would be good to have a test specifically loading from the dispatch packet like happens in the workitem ID calculation


Repository:
  rL LLVM

https://reviews.llvm.org/D35146