[Mlir-commits] [mlir] [mlir][XeGPU][VectorToXeGPU] Propagate vector layouts to xegpu ops (PR #163071)

[Dmitry Chigarev]

================
@@ -616,16 +636,34 @@ struct GatherLowering : public OpRewritePattern<vector::GatherOp> {
         computeOffsets(rewriter, gatherOp, meta.first, meta.second);
     Value flatMemref = memrefToIndexPtr(gatherOp, rewriter);
 
+    auto layoutRes = xegpu::getDistributeLayoutAttr(gatherOp.getResult());
+    auto layoutIndices =
+        xegpu::getDistributeLayoutAttr(gatherOp.getIndicesMutable());
----------------
dchigarev wrote:

otherwise it won't get `layout_operand_i` assigned to this operand

https://github.com/llvm/llvm-project/pull/163071