[Mlir-commits] [mlir] [mlir][gpu] Add patterns to break down subgroup reduce (PR #76271)

[Jakub Kuderski]

================
@@ -39,10 +40,35 @@ struct TestGpuRewritePass
     (void)applyPatternsAndFoldGreedily(getOperation(), std::move(patterns));
   }
 };
+
+struct TestGpuSubgroupReduceLoweringPass
+    : public PassWrapper<TestGpuSubgroupReduceLoweringPass,
+                         OperationPass<ModuleOp>> {
+  MLIR_DEFINE_EXPLICIT_INTERNAL_INLINE_TYPE_ID(
+      TestGpuSubgroupReduceLoweringPass)
+
+  void getDependentDialects(DialectRegistry &registry) const override {
+    registry.insert<arith::ArithDialect, func::FuncDialect, index::IndexDialect,
+                    memref::MemRefDialect, vector::VectorDialect>();
----------------
kuhar wrote:

These are the 'dependent dialects', which excludes the input dialect:
```c++
  /// A pass is expected to register the dialects it will create entities for
  /// (Operations, Types, Attributes), other than dialect that exists in the
  /// input. For example, a pass that converts from Linalg to Affine would
  /// register the Affine dialect but does not need to register Linalg.
```

https://github.com/llvm/llvm-project/pull/76271