AMDGPU: Add LLVM IR Intrinsic implementation and a LIT test for v_lerp_u8
Details
Details
Diff Detail
Diff Detail
- Repository
- rL LLVM
Event Timeline
test/CodeGen/AMDGPU/llvm.amdgcn.lerp.ll | ||
---|---|---|
7 | There still is no FUNC check prefix. You don't need separate SI and VI lines because they are identitcal |
test/CodeGen/AMDGPU/llvm.amdgcn.lerp.ll | ||
---|---|---|
2–3 | You should use either the default check prefix or GCN |
test/CodeGen/AMDGPU/llvm.amdgcn.lerp.ll | ||
---|---|---|
7 | This is still using FUNC |
This pattern is redundant with the one for the instruction definition and should be removed