RSS Feed Migrating from lemieux or ben to bigben
The following chart helps to illustrate the differences between lemieux, ben and bigben. Please read the bigben document for details on using bigben.
| Lemieux/ Ben | Bigben | |
|---|---|---|
| Local disk on compute nodes | $LOCAL | N/A |
| Compilers | HP, Gnu, Java | PG, Gnu, UPC |
| MPI | use -lmpi, -lelan | use wrappers cc, ftn, CC |
| OpenMP | Yes | No |
| SHMEM | Yes | No |
| Debug queue availability | 24 hrs., every day | 8am - 8pm, M-F |
| System drains | Lemieux: 8 am Wed. Ben: As needed |
8am & 8pm, Monday-Friday |
| Scheduling | FIFO | Largest jobs first |
| Max. walltime | Lemieux: 48 hrs. Ben: 36 hrs. |
8 hrs. |
| PBS options | -l rmsnodes=N:n | -l size=N:M where N=nodes, M=processes can also use -l size=M |
| prun | pbsyod | |
| $RMS_NODES | $PBS_NNODES | |
| $RMS_PROCS | $PBS_NPROCS | |
| -m cyclic | default behavior | |
| Lemieux: -l rmsrails Ben: N/A |
N/A | |
| -l rmsproject | -w group_list | |
| mprun | N/A | |
| Chaining jobs | Yes | No |
| Copy files from $SCRATCH to golem | tcscp | far |
| IO to/from $LOCAL | TCSIO | N/A |
| AGW nodes | Lemieux: Yes Ben: No |
No |
| Vis nodes | Lemieux: Yes Ben: No |
No |
| Node and queue information | rinfo | shownids |
| nodestat | N/A | |
| backlog | N/A | |
| qest | N/A | |
| Checkpointing for system failures | Lemieux: Yes Ben: No | No |
| Messages from running jobs | tcs-postmessage | N/A |