Speaker
Dr
Michal Svatoš
(FZU, AV CR)
Description
The distributed computing system of the ATLAS experiment at the Large Hadron Collider (LHC) uses computing resources of IT4Innovations for several years. So far, only selected workloads were used, i.e. those most suitable for many-core environment. The computing needs of the experiment do change and with that there is a need to be able to run more workloads. The HyperQueue, a tool designed to simplify execution of large workflows at HPCs, allows us to do that. Its schedulers were updated to allow full utilization of available resources by ATLAS jobs which request different number of CPU cores or memory. This work started in 2025 and it is now concluded.
Authors
Jiří Chudoba
(FZU AV ČR)
Dr
Michal Svatoš
(FZU, AV CR)
Mr
Petr Vokáč
(FNSPE, CTU)