[Mlir-commits] [mlir] [mlir][AMDGPU] Enable emulating vector buffer_atomic_fadd for bf16 on gfx942 (PR #129029)

[Jakub Kuderski]

================
@@ -189,6 +189,14 @@ void mlir::amdgpu::populateAmdgpuEmulateAtomicsPatterns(
     } else {
       target.addIllegalOp<RawBufferAtomicFmaxOp>();
     }
+    // gfx950 has bf16 atomics
+    if (chipset < Chipset(9, 5, 0)) {
----------------
kuhar wrote:

This part of MLIR is underdeveloped, we don't have a good target description mechanism. This can come from some options object provided by the caller for the time being, but we shouldn't block this PR based on the lack of target feature mechanism.

https://github.com/llvm/llvm-project/pull/129029