You do not have permission to delete messages in this group
Copy link
Report message
Show original message
Either email addresses are anonymous for this group or you need the view member email addresses permission to view the original message
to cp2k
Dear All,
The computer cluster of my university uses SLURM as workload manager.
For parallel runs, it imposes a memory usage limit per task (4GB by
default). I am trying to run a simulation using the attached input files
but the workload manager kills the simulation before it can even write
an output file, I suppose because one of the mpi tasks (possibly the
master) is using much more memory than the others (and that is allowed)
at the beginning. SLURM is giving me the following error:
The task which had the largest memory consumption differs by 260.42%
from the average task max memory consumption
I have tried adding TRACE and TRACE_MASTER but as I said, cp2k.psmp is
killed before it writes an output file.
This simulation runs fine using CP2K 2024.3 on the same cluster.
You do not have permission to delete messages in this group
Copy link
Report message
Show original message
Either email addresses are anonymous for this group or you need the view member email addresses permission to view the original message
to cp2k
Dear Edwin,
You may
- switch from RUN_TYPE MD to RUN_TYPE ENERGY_FORCE (or even ENERGY) to ensure that the base calculation can be run at all and all relevant parameters are converged (EPS_SCF)
- omit the `-o <OUTPUT_FILE>` part in the command to run CP2k to redirect the output to standard output
Best,
Frederick
Edwin Helbert Aponte Angarita
unread,
Aug 6, 2026, 4:20:54 PMAug 6
Reply to author
Sign in to reply to author
Forward
Sign in to forward
Delete
You do not have permission to delete messages in this group
Copy link
Report message
Show original message
Either email addresses are anonymous for this group or you need the view member email addresses permission to view the original message
to cp...@googlegroups.com
Dear All,
Thank your Frederick for your reply. Sorry for my late reply.
Forcing SLURM to allocate a number of CPUs per node solved the issue.
The minimum number of required CPUs per node seems to depends on the
molecule being simulated.