[Mlir-commits] [mlir] [MLIR][NVVM] Add support for f6x2 conversion (PR #136537)

[Srinivasa Ravi]

================
@@ -1066,6 +1066,56 @@ def NVVM_CvtFloatToTF32Op : NVVM_Op<"cvt.float.to.tf32"> {
   }];
 }
 
+def CVTFP6E2M3 : I32EnumAttrCase<"E2M3", 0, "e2m3">;
----------------
Wolfram70 wrote:

I did this here because I see that for the `wgmma.mma_async` Op, we have types like `WGMMATypeF8E4M3` to represent the operand types. If we have a more general enumeration, should we change the usage in `wgmma.mma_async` too (since we will be having cvt Ops for the FP8 types as well)?

https://github.com/llvm/llvm-project/pull/136537