[Mlir-commits] [mlir] [mlir][ArmSME] Fold transpose into xfer read to enable in-flight transpose (PR #92562)

[Benjamin Maxwell]

================
@@ -356,6 +356,20 @@ struct TransposeOpToArmSMELowering
       return failure();
 
     auto loc = transposeOp.getLoc();
+    Value input = transposeOp.getVector();
+
+    if (auto xferOp = input.getDefiningOp<vector::TransferReadOp>()) {
+      // Fold transpose into transfer_read to enable in-flight transpose when
+      // converting to arm_sme.tile_load.
----------------
MacDue wrote:

The memref is a problem for IREE though, and if we hit somewhere where we needed to lower a `vector.transpose` alone it'd probably be better just to use an extra tile. 

https://github.com/llvm/llvm-project/pull/92562