https://github.com/arsenm commented:
I think this is OK for the __opencl_* builtins, but I do not think it is appropriate to make the behavior of these builtins modal. CUDA does provide atomic functions which do not have the undefined behavior in the private case, which you ideally would be implemented with these builtins https://github.com/llvm/llvm-project/pull/230044 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
