[Mlir-commits] [mlir] [mlir][ArmNeon] Updates LowerContractionToSMMLAPattern with vecmat unroll patterns (PR #86005)

[Kojo Acquah]

================
@@ -150,16 +150,30 @@ class LowerContractionToSMMLAPattern
       Value tiledAcc =
           extractOperand(op.getAcc(), accPermutationMap, accOffsets);
 
+      // With vecmat, tiled LHS and ACC will contain only one of 2 necessary
+      // rows along dimM. Broadcast both to the full width
+      if (isVecmat) {
+        auto lhsBroadcastType = VectorType::get(
+            {2, 8}, tiledLhs.getType().cast<ShapedType>().getElementType());
+        tiledLhs = rewriter.create<vector::BroadcastOp>(loc, lhsBroadcastType,
+                                                        tiledLhs);
+        auto accBroadcastType = VectorType::get(
+            {2, 2}, tiledAcc.getType().cast<ShapedType>().getElementType());
+        tiledAcc = rewriter.create<vector::BroadcastOp>(loc, accBroadcastType,
+                                                        tiledAcc);
----------------
KoolJBlack wrote:

@dcaballe I couldn't find an equivalent vector op for `tensor::empty` so using a broadcast here. Half the matmul is not needed so the data can be anything. If there is something simpler, let me know. 

https://github.com/llvm/llvm-project/pull/86005