[Mlir-commits] [mlir] [mlir][x86] Support for `f8` AMX tiled dot-product. (PR #194786)

[Adam Siemieniuk]

================
@@ -72,6 +72,40 @@ func.func @amx_tile_mulf_f16(
   return
 }
 
+// CHECK-LABEL: define void @amx_tile_mulf_f8
+func.func @amx_tile_mulf_f8E4M3FN(
----------------
adam-smnk wrote:

nit: let's drop the suffix `E4M3FN`

https://github.com/llvm/llvm-project/pull/194786