[Mlir-commits] [mlir] [MLIR][XeGPU] Support order attribute and add pattern for vector.transpose in WgToSg Pass (PR #165307)

[Charitha Saumya]

================
@@ -1217,6 +1217,93 @@ struct WgToSgMultiDimReductionOp
   }
 };
 
+// This pattern transforms vector.transpose ops to work at subgroup level.
+struct WgToSgVectorTransposeOp
+    : public OpConversionPattern<vector::TransposeOp> {
+  using OpConversionPattern<vector::TransposeOp>::OpConversionPattern;
+
+  LogicalResult
+  matchAndRewrite(vector::TransposeOp op, OneToNOpAdaptor adaptor,
+                  ConversionPatternRewriter &rewriter) const override {
+    VectorType resultType = dyn_cast<VectorType>(op.getResult().getType());
+    if (!resultType)
+      return failure();
----------------
charithaintc wrote:

not needed. use `getResultVectorType`. result is always guranteed to be vector type. 

https://github.com/llvm/llvm-project/pull/165307