[Mlir-commits] [mlir] [mlir][ArmSME] Switch to an attribute-based tile allocation scheme (PR #73253)

[Cullen Rhodes]

================
@@ -118,13 +135,13 @@ struct ZeroOpConversion : public ConvertOpToLLVMPattern<arm_sme::ZeroOp> {
     //  * Mask    -> 10001000 = (00010001 << 3)
     //
     // This holds for all tile sizes.
-    auto tileMask = rewriter.create<arith::ShLIOp>(
-        loc, baseMask, castTileIDToI32(tileId, loc, rewriter));
-    rewriter.create<arm_sme::aarch64_sme_zero>(loc, tileMask);
+    int32_t zeroMask = baseMaskForSize << int32_t(tileId.getInt());
----------------
c-rhodes wrote:

nice!

https://github.com/llvm/llvm-project/pull/73253