[Mlir-commits] [mlir] [MLIR][Transform] FuseOp: accept transform params, add use_forall argument (PR #161883)

[Rolf Morel]

================
@@ -395,31 +395,72 @@ def EliminateLinalgOpAnchoredEmptyTensorsOp
 //===----------------------------------------------------------------------===//
 
 def FuseOp : Op<Transform_Dialect, "structured.fuse",
-    [FunctionalStyleTransformOpTrait, MemoryEffectsOpInterface,
-     DeclareOpInterfaceMethods<TransformOpInterface>,
-     ReportTrackingListenerFailuresOpTrait]> {
+       [AttrSizedOperandSegments,
+        DeclareOpInterfaceMethods<MemoryEffectsOpInterface>,
+        TransformOpInterface, ReportTrackingListenerFailuresOpTrait]> {
   let description = [{
     Tiles the operations pointed to by the target handle and fuses their
     producers greedily using the options provided as attributes.
 
     If `apply_cleanup` is true then slice canonicalization is applied between
-    fusion steps.
+    fusion steps. If `use_forall` is true then tiling method generates a
+    `scf.forall` loop instead of `scf.for` loops.
   }];
 
   let arguments =
     (ins TransformHandleTypeInterface:$target,
-         DefaultValuedAttr<I64ArrayAttr, "{}">:$tile_sizes,
-         DefaultValuedAttr<I64ArrayAttr, "{}">:$tile_interchange,
-         DefaultValuedAttr<BoolAttr, "false">:$apply_cleanup);
+        Variadic<TransformAnyParamTypeOrAnyHandle> : $tile_sizes,
+        Variadic<TransformAnyParamTypeOrAnyHandle> : $tile_interchange,
----------------
rolfmorel wrote:

Does the op work in case `$tile_interchange` is a handle to a payload op/value? The type `...OrAnyHandle` allows this.

Here I am not so sure that the underlying transforms support the loop order only being resolved at binary execution time.

https://github.com/llvm/llvm-project/pull/161883