ivanradanov wrote: I see. I was doing this ``` clang -foffload-via-llvm --cuda-path=/usr/local/cuda input.o -o a.out ```
which only passes `--cuda-path` to `clang-linker-wrapper` directly: ``` "clang-linker-wrapper" ... "--cuda-path=/usr/local/cuda" ... ``` I suppose under the current infra, all the offloading compilation arguments need to be present when linking as well, so as to invoke the appropriate toolchains, and the above usage is not recommended? https://github.com/llvm/llvm-project/pull/149107 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
