DEMO server: Add support for light local hyperqueue scheduler

aiidalab / aiidalab-docker-stack

Docker images with the basic software stack for AiiDAlab

https://aiidalab.net

Other

9 stars 14 forks source link

DEMO server: Add support for light local hyperqueue scheduler #463

Closed unkcpz closed 3 months ago

unkcpz commented 5 months ago

[ ] Use runOnce to manage the restart of hq.
[ ] For hq on demo-server image, test 1. time scaling of pw.x with mpiprocs number. 2. time of loading notebook when different number of pw.x is running.
[ ] Understanding the behavior when number of workers decreases.

danielhollas commented 4 months ago

Just a note: At the meeting it was decided to move the hyperqueue schedule to the QE image first to get some operational experience with it. Then we can consider putting it in the full-stack image.

@unkcpz thank you for working on this! I am more than happy to ditch SLURM in my image. I will test this once it is integrated to the QeApp image.

danielhollas commented 3 months ago

@unkcpz I think I this can be closed for now?

unkcpz commented 3 months ago

@danielhollas Sure, let's close this one and I'll continue on QeApp.