================
@@ -5330,11 +5330,13 @@ static uint32_t
getIdentityValueFor32BitWaveReduction(unsigned Opc) {
case AMDGPU::S_MAX_U32:
return std::numeric_limits<uint32_t>::min();
case AMDGPU::S_MAX_I32:
+ case AMDGPU::V_SUB_F32_e64: // +0.0
----------------
easyonaadit wrote:
Sorry, I think I got the values mixed up. I'll update it.
I didn't take it from a specification as such, just that the atomic optimizer
uses these values for identity.
Its supposed to be `-0.0` for add, and `+0.0` for sub.
https://github.com/llvm/llvm-project/pull/161815
_______________________________________________
llvm-branch-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
