[PATCH] D63203: [AMDGPU] gfx1010 dpp16 and dpp8

[Matt Arsenault via Phabricator via llvm-commits]

arsenm added inline comments.


================
Comment at: include/llvm/IR/IntrinsicsAMDGPU.td:1457
+  Intrinsic<[llvm_anyint_ty],
+            [LLVMMatchType<0>, llvm_i32_ty],
+            [IntrNoMem, IntrConvergent]>;
----------------
arsenm wrote:
> Needs immarg, and to be marked as a source of divergence
I think it's a source of divergence? It's missing from the existing DPP intrinsics, but that might be a bug?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D63203/new/

https://reviews.llvm.org/D63203