[Mlir-commits] [mlir] [MLIR][NVGPU] Introduce `nvgpu.wargroup.mma.store` Op for Hopper GPUs (PR #65441)

[llvmlistbot at llvm.org]

================
@@ -1299,6 +1301,82 @@ struct NVGPUWarpgroupMmaOpLowering
   }
 };
 
+struct NVGPUWarpgroupMmaStoreOpLowering
+    : public ConvertOpToLLVMPattern<nvgpu::WarpgroupMmaStoreOp> {
+  using ConvertOpToLLVMPattern<
+      nvgpu::WarpgroupMmaStoreOp>::ConvertOpToLLVMPattern;
+
+  void storeFragmentedMatrix(Value wgmmaResult, nvgpu::WarpgroupMmaStoreOp op,
----------------
qcolombet wrote:

Could you add a comment explaining what each of the argument hold?

https://github.com/llvm/llvm-project/pull/65441