[Mlir-commits] [mlir] [MLIR][AMDGPU] Add a wrapper for global LDS load intrinsics in AMDGPU (PR #133498)

[Krzysztof Drewniak]

================
@@ -765,4 +771,37 @@ def AMDGPU_WMMAOp :
   let hasVerifier = 1;
 }
 
+def GlobalLoadMemRefType : MemRefOf<[GlobalLoadTypes]>;
+
+def AMDGPU_GlobalLoadLDSOp :
+    AMDGPU_Op<"global_load", [SameVariadicOperandSize]>,
+    Arguments<(ins
+                   Arg<GlobalLoadMemRefType, "buffer to read from", [MemRead]>:$src,
+                   Variadic<I32>:$srcIndices,
+                   Arg<GlobalLoadMemRefType, "buffer to write to", [MemWrite]>:$dst,
+                   Variadic<I32>:$dstIndices
----------------
krzysz00 wrote:

Same - `index` here, it'll get truncated down

https://github.com/llvm/llvm-project/pull/133498