[Mlir-commits] [mlir] [MLIR][NVVM] Update support for conversions to f8x2 and f6x2 types (PR #137781)

[Srinivasa Ravi]

================
@@ -1120,6 +1121,153 @@ def NVVM_CvtToF6x2Op : NVVM_Op<"cvt.to.f6x2"> {
   }];
 }
 
+def CVTFP8E4M3 : I32EnumAttrCase<"E4M3", 0, "e4m3">;
+def CVTFP8E5M2 : I32EnumAttrCase<"E5M2", 1, "e5m2">;
+def CVTFP8UE8M0 : I32EnumAttrCase<"UE8M0", 2, "ue8m0">;
----------------
Wolfram70 wrote:

Yeah, that makes sense. I did it this way here since the other existing type enums were prefixed with the corresponding Op name. But we should probably unify all of them and rename them with an `NVVMFP` prefix instead perhaps.

https://github.com/llvm/llvm-project/pull/137781