[PATCH] D80142: [mlir][gpu][rocdl] Introduce GPUToROCm conversion passes.

[Stephan Herhut via Phabricator via llvm-commits]

herhut added a comment.

Cool, nice to see ROCm support coming!



================
Comment at: mlir/lib/Conversion/GPUToROCm/ConvertKernelFuncToHsaco.cpp:50
+/// The function body is erased.
+class GpuKernelToHsacoPass
+    : public PassWrapper<GpuKernelToHsacoPass,
----------------
This is a lot of common code. Could we have a base class, e.g., `GpuKernelToLLVMGeneratedBlobPass', that factors out all the common logic but has extensions points for the LLVM target to use, etc.?


================
Comment at: mlir/lib/Conversion/GPUToROCm/ConvertKernelFuncToHsaco.cpp:146
+
+  auto ptx = translateModuleToLLVM(llvmModule, *targetMachine);
+
----------------
`ptx` -> `hsa`?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D80142/new/

https://reviews.llvm.org/D80142