[Mlir-commits] [mlir] [MLIR][AMDGPU] Add support for fp8 ops on gfx12 (PR #106388)

[Jakub Kuderski]

================
@@ -601,6 +604,10 @@ static std::optional<StringRef> wmmaOpToIntrinsic(WMMAOp wmma,
     return ROCDL::wmma_bf16_16x16x16_bf16::getOperationName();
   } else if (elemSourceType.isInteger(8) && elemDestType.isInteger(32)) {
     return ROCDL::wmma_i32_16x16x16_iu8::getOperationName();
+  } else if (elemSourceType.isFloat8E4M3FN() && elemDestType.isF32()) {
+    return ROCDL::wmma_f32_16x16x16_fp8::getOperationName();
+  } else if (elemSourceType.isFloat8E5M2() && elemDestType.isF32()) {
+    return ROCDL::wmma_f32_16x16x16_bf8::getOperationName();
----------------
kuhar wrote:

nit: no else after return, here and above: https://llvm.org/docs/CodingStandards.html#don-t-use-else-after-a-return

https://github.com/llvm/llvm-project/pull/106388