jhuber6 wrote: > @jhuber6 do you think can we use `native` instead? I think it would be a > somewhat better option here. If we have to choose a GPU variant by default, > we may as well choose the actual GPU, rather than a conditional choice > between generic SPIR-V or an old GPU, which has the disadvantage of not being > the right choice for anyone, and depending on another moving part > (amd-llvm-spirv)
OpenMP already defaults to `native` if you don't specify anything. Having the actual toolchain variant on commandline arguments makes me a little uncertain. Long term I'd prefer people do something like `--offload-targets=spirv64-amd-amdhsa` or something and that just defaults to the proper toolchain. https://github.com/llvm/llvm-project/pull/139281 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
