[Mlir-commits] [mlir] [MLIR][XeGPU] Add offsets to load/store/prefetch (PR #150576)

Adam Siemieniuk llvmlistbot at llvm.org
Mon Jul 28 09:58:40 PDT 2025


================
@@ -189,6 +189,7 @@ def XeGPU_TensorDesc: XeGPUTypeDef<"TensorDesc", "tensor_desc",
   let genVerifyDecl = 1;
 }
 
+def XeGPU_TensorDescOrMemRef : AnyTypeOf<[XeGPU_TensorDesc,Non0RankedMemRefOf<[XeGPU_ScalarType]>, UI64]>;
----------------
adam-smnk wrote:

To rephrase a bit, the overall direction sounds good.
However, looking from an outsider's perspective, it's currently hard to derive all these decisions and intensions from the existing upstream docs. With some updates, it should be much easier to understand and follow the new/current design.

https://github.com/llvm/llvm-project/pull/150576


More information about the Mlir-commits mailing list