sarnex wrote: Just building OpenMP is much simpler. The unlucky thing here is that the problem is in OpenMP SPIR-V offload, which is experimental so it is not built by default, and the only GPUs supported are Intel GPUs (which use the Level Zero runtime).
The issue was caught by our buildbot, and I don't think its reasonable for us to expect for you to test on Intel GPUs (even though this specific issue does not require hardware), so IMO you can just keep doing your normal development process. https://github.com/llvm/llvm-project/pull/187051 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
