[Mlir-commits] [mlir] [MLIR][AMDGPU] Adding Vector to AMDGPU conversion lowering (PR #131803)

[Jakub Kuderski]

================
@@ -1333,6 +1333,16 @@ def ConvertVectorToArmSMEPass : Pass<"convert-vector-to-arm-sme"> {
   let dependentDialects = ["arm_sme::ArmSMEDialect", "arm_sve::ArmSVEDialect"];
 }
 
+//===----------------------------------------------------------------------===//
+// VectorToAMDGPU
+//===----------------------------------------------------------------------===//
+
+def ConvertVectorToAMDGPUPass : Pass<"convert-vector-to-amdgpu"> {
+  let summary = "Lower the operations from the vector dialect into the AMDGPU "
+                "dialect";
+  let dependentDialects = ["amdgpu::AMDGPUDialect", "vector::VectorDialect"];
----------------
kuhar wrote:

Do we need amdgpu? This doesn't emit any amdgpu ops.

https://github.com/llvm/llvm-project/pull/131803