Known issues

Here are some known issues on the NHR systems:

• Floating point exception with Intel MPI 2019.x using one task per node

• Cannot contact external servers/services

• INTEL-MPI version 19 slower than INTEL-MPI version 18

• Known warnings

• MPI executable dies with error: “hfi_userinit: mmap of status page …

• MPI Jobs with more than 40 (96) tasks per node failing

• Slow execution of nccopy from netcdf 4.6.3

• SLURM does not recognize job script

• Too many open files

• Unspecific error messages when reading huge input files