[Mlir-commits] [mlir] [MLIR][Linalg] Add aggregate ops decomposition pass and softmax decom… (PR #97582)

[llvmlistbot at llvm.org]

================
@@ -211,31 +213,61 @@ func.func @softmax(%arg0: tensor<2x16x32xf32>, %dst: tensor<2x16x32xf32>) -> ten
 // CHECK-SAME:           %[[ARG0:[a-zA-Z0-9_]+]]: tensor<2x16x32xf32>, %[[DST:[a-zA-Z0-9_]+]]: tensor<2x16x32xf32>) -> tensor<2x16x32xf32> {
 // CHECK-DAG:        %[[D1:.+]] = tensor.empty() : tensor<2x16xf32>
 // CHECK-DAG:        %[[CST:.+]] = arith.constant -3.40282347E+38 : f32
-// CHECK:        %[[D2:.+]] = linalg.fill ins(%[[CST]] : f32) outs(%[[D1]] : tensor<2x16xf32>) -> tensor<2x16xf32>
+// CHECK:        %[[D2:.+]] = linalg.generic {indexing_maps = [#[[$MAP2]], #[[$MAP3]]],
+// CHECK-SAME:      iterator_types = ["parallel", "parallel"]}
+// CHECK-SAME:      ins(%[[CST]] : f32) outs(%[[D1]] : tensor<2x16xf32>) {
+// CHECK-NEXT:      ^bb0(%[[IN:.+]]: f32, %[[OUT:.+]]: f32):
+// CHECK-NEXT:      linalg.yield %[[IN]] : f32
+// CHECK-NEXT:    } -> tensor<2x16xf32>
 // CHECK:        %[[D3:.+]] = linalg.generic {indexing_maps = [#[[$MAP]], #[[$MAP1]]], iterator_types = ["parallel",
 // CHECK-SAME:     "parallel", "reduction"]} ins(%[[ARG0]] : tensor<2x16x32xf32>) outs(%[[D2]] : tensor<2x16xf32>) {
 // CHECK:        ^bb0(%[[IN:.+]]: f32, %[[OUT:.+]]: f32):
 // CHECK:          %[[D8:.+]] = arith.maxnumf %[[IN]], %[[OUT]] : f32
 // CHECK:          linalg.yield %[[D8]] : f32
 // CHECK:        } -> tensor<2x16xf32>
-// CHECK:        %[[D4:.+]] = linalg.generic {indexing_maps = [#[[$MAP]], #[[$MAP1]], #[[$MAP]]], iterator_types =
-// CHECK-SAME:     ["parallel", "parallel", "parallel"]} ins(%[[ARG0]], %[[D3]] : tensor<2x16x32xf32>, tensor<2x16xf32>)
+// CHECK:        %[[BCST:.+]] = linalg.generic {indexing_maps = [#[[$MAP1]], #[[$MAP]]],
----------------
MaheshRavishankar wrote:

Why are the broadcast ops separate?

https://github.com/llvm/llvm-project/pull/97582