[Mlir-commits] [mlir] [MLIR][AMDGPU] Add amdgpu.global_transpose_load op for RDNA4 global memory transpose loads (PR #195287)

[Krzysztof Drewniak]

================
@@ -1449,6 +1449,45 @@ def AMDGPU_TransposeLoadOp :
   let hasVerifier = 1;
 }
 
+def AMDGPU_GlobalTransposeLoadOp :
+    AMDGPU_Op<"global_transpose_load", [SameVariadicOperandSize]>,
+    Arguments<(ins Arg<AnyMemRef, "buffer to transpose load from", [MemRead]>:$src,
+                      Variadic<Index>:$srcIndices)>,
+    Results<(outs AnyTypeOf<[AnyVectorOfNonZeroRank]>:$result)> {
----------------
krzysz00 wrote:

1. I don't think we need the `AnyTypeOf`, and I think we should constrain this to a 1-D vector and *maybe* list the permitted element types too

https://github.com/llvm/llvm-project/pull/195287