[Mlir-commits] [mlir] [mlir][arith] Expand the scaling ops with the value of the scale (PR #217892)

[Krzysztof Drewniak]

================
@@ -555,8 +661,8 @@ func.func @scaling_extf_vector_to_f32_using_f16_scales_fastmath(%arg0: vector<4x
 }
 
 // SCHECK-LABEL: @scaling_extf_vector_to_f32_using_f16_scales_fastmath
-// SCHECK: %[[TRUNCF_SCALE:.+]] = arith.truncf %arg1 fastmath<fast> : vector<4xf16> to vector<4xf8E8M0FNU>
-// SCHECK: %[[EXT_SCALE:.+]] = arith.extf %[[TRUNCF_SCALE]] fastmath<fast> : vector<4xf8E8M0FNU> to vector<4xf32>
+// SCHECK-NOT: f8E8M0FNU
+// SCHECK: %[[EXT_SCALE:.+]] = arith.extf %arg1 fastmath<fast> : vector<4xf16> to vector<4xf32>
----------------
krzysz00 wrote:

Nit: let's not hard-code argument names but instead `SCHECK-SAME` under the `SCHECK-LABEL`  to match them

https://github.com/llvm/llvm-project/pull/217892