[PATCH] D57748: AMDGPU: Add inverse ballot intrinsic

[Matt Arsenault via Phabricator via llvm-commits]

arsenm added inline comments.


================
Comment at: test/CodeGen/AMDGPU/llvm.amdgcn.inverse.ballot.ll:18-20
+define amdgpu_ps float @vgpr(i32 %v0, i32 %v1) {
+  %tmp = insertelement <2 x i32> undef, i32 %v0, i32 0
+  %tmp2 = insertelement <2 x i32> %tmp, i32 %v1, i32 1
----------------
cwabbott wrote:
> arsenm wrote:
> > You can just use an i64 argument here instead of this build vector and cast thing
> No, that crashes:
> 
> 
> ```
> Formal argument #0 has unhandled type i64
> UNREACHABLE executed at ../lib/CodeGen/CallingConvLower.cpp:98!
> ```
> 
> I just copied the amdgpu_ps calling convention from somewhere else, is there a better one to use?
I know I've written the patch to fix this before, but I guess I never committed it.

You can just use a default calling convention function for this just as well


Repository:
  rL LLVM

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D57748/new/

https://reviews.llvm.org/D57748