[Mlir-commits] [mlir] [MLIR][XeGPU]Enhance Pack/Unpack for XeGPUUnroll (PR #163459)

Igor Zamyatin llvmlistbot at llvm.org
Thu Oct 16 14:24:53 PDT 2025


================
@@ -145,8 +145,21 @@ XeGPUBlockingPass::getTileShape(const T &operandOrResult) const {
   xegpu::DistributeLayoutAttr layout =
       xegpu::getDistributeLayoutAttr(operandOrResult);
   if (layout && layout.isForSubgroup()) {
-    if (!layout.getEffectiveInstDataAsInt().empty())
-      return layout.getEffectiveInstDataAsInt();
+    if (!layout.getEffectiveInstDataAsInt().empty()) {
+      SmallVector<int64_t> instData = layout.getEffectiveInstDataAsInt();
+      // Remove leading unit dimensions from inst_data
+      // Skip it for xegpu nd ops since it will be 2D
----------------
Garra1980 wrote:

can you please elaborate a bit regarding unit dimensions, some short example perhaps

https://github.com/llvm/llvm-project/pull/163459


More information about the Mlir-commits mailing list