Changes after the update of August 2023

LUMI is still on maintenance break. This notification is sent to allow you to update your job scripts before LUMI becomes available again.

This ongoing maintenance break will introduce a change in the SLURM configuration that will make resource usage on LUMI-G more efficient (the upgrade will change the reserved core configuration on LUMI-G to be 8 cores instead of 1). Due to the reconfiguration of SLURM, many job scripts will fail. Please review the detailed description of this change with examples of affected job scripts, and adapt your job scripts if needed.

Jobs for LUMI-G that are currently in the queue that are incompatible with the new configuration will either get stuck indefinitely in the queue and will have to be removed with scancel (when the job script requested more than 56 cores per node explicitly) or will fail while running, likely with an error message about conflicting bindings or cores not being available for binding.

Besides, responsible use of LUMI can help reducing the waiting times for everybody and helps the priority system to function as designed. Please use small or small-g for small jobs! Please see the guidance here as well as the policy change regarding dev-g and eap.

Please don’t hesitate to contact the LUMI User Support Team at https://lumi-supercomputer.eu/user-support/need-help/ if you need any assistance.

We apologize for the inconvenience caused by the maintenance break.