https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103985
--- Comment #5 from Dr. Thomas Orgis <thomas.or...@uni-hamburg.de> --- OK, I tested with gcc-12-20220109. I repeated the test job, got 17102 successful runs and no crash. So one can hope that this particular one has been fixed at some point. All good? Is the specific change easily identified for backporting to a bugfix release? I wonder if it is really a regression. I did not try so many runs with earlier GCC versions and cannot really keep the cluster busy just for that. That keeps my performance worry open: Is it really the case that the write to a character variable needs a (global?) lock? PS: I had trouble building netcdf 4.6.3 with this snapshot … rather basic test failures. I did not investigate this further and just went along as the crash is not in the file storage parts. I hope this is just an artifact of my setup and not some very basic actual breakage.