[Mlir-commits] [mlir] [mlir][vector] Constrain broadcast->shape_cast folding (PR #190230)

[Andrzej Warzyński]

================
@@ -10,7 +10,12 @@ func.func @mmt4d_to_fma(%A: tensor<16x16x8x1xf32>, %B: tensor<16x16x8x1xf32>, %C
 
 
 // CHECK-LABEL:     @mmt4d_to_fma
-// CHECK-COUNT-8:         vector.fma
+// CHECK-NOT:         linalg.mmt4d
+// Lowering may produce vector.fma (outerproduct) or elementwise arith.mulf +
+// vector.reduction depending on vector.contract lowering; both are valid.
----------------
banach-space wrote:

How do we define "valid" ? :) 

Semantically both are correct, yes, but as the name of this test file implies, it's about lowering to `vector.fma` rather than `arith.mulf`. We should preserve that part of the test. 

https://github.com/llvm/llvm-project/pull/190230