aiidalab / aiidalab/aiidalab-qe
HPC resource settings must be expanded for PBS schedulers
- Dominant language
- Python
- Stars
- 33
- Forks
- 25
- Avg merge
- 6h 7m
- Merged PRs (30d)
- 8
Description
**Is your feature request related to a problem? Please describe.**
Currently, the App allows users to set only the number of nodes and CPUs for a particular workflow. For the SLURM scheduler (and maybe others, though I have not tested those), one can specify additional options *via* the `prepend_text` section of the Computer node itself instead of setting them in the `builder.metadata.options` namespace. For PBS schedulers (torque, PBSpro, etc) this cannot be done as the plugin for PBS in `aiida-core` appends `cd $PBS_O_WORKDIR` to the end of the scheduler options like so:
```
#!/bin/bash
#PBS -r n
#PBS -m n
#PBS -N aiida-696
#PBS -V
#PBS -o _scheduler-stdout.txt
#PBS -e _scheduler-stderr.txt
#PBS -l walltime=12:00:00
#PBS -l select=1:mpiprocs=16
cd "$PBS_O_WORKDIR"
#PBS -q s3par
module purge
module load IMPI/qe-7.1
```
The scheduler will stop parsing options in this case because it detects a shell command, so the queue name is not set for the job.
**Describe the solution you'd like**
A simple solution, which I have already written and tested locally, is to add a simple text box to Step 3 of the setup where the user can specify the name of the queue to use. This also obviously requires a slight refactoring of the backend code in order to set these changes correctly in the builder.
**Additional context**
Screenshot from my local version:

Contributor guide
No contributing guide indexed for this repository
Assessment
This issue has not been assessed yet.