HPC Systems¶
Platform-specific setup and known issues for the clusters CosmicGrain has been run on. If something here doesn't apply to your system, it probably belongs on Installation or Compilation instead — this page is only for things that are true on one specific machine.
TACC Stampede¶
Known issue — restart-file portability. Restart files are not portable across different MPI task counts on Stampede; restarting with a different task count than the original run hangs. Use snapshot-based restarts instead when changing task count.
NMSU Discovery Cluster¶
Known issue — persistent hang. A hardware/MPI-stack issue has been identified as the cause of certain hangs on this cluster; .