New GCS Large-Scale Projects Started in May 2021
Twice a year, the Gauss Centre for Supercomputing (GCS) issues a call for large-scale projects on its petascale supercomputers – Hawk (HLRS), JUWELS (JSC), and SuperMUC-NG (LRZ). Projects are classified as large-scale if they require at least 2 % of the systems’ annual production in terms of estimated availability.
After computing time quantities were previously specified in core hours, the modularity of JUWELS requires the introduction of a new computing time unit. JSC is currently working with the peak floating point operations per year (FLOP/a) of the computing devices (CPU or GPU) available to approved projects. The computing time on Hawk and SuperMUC-NG continues to be given in core hours. Projects in this case fall into the category of large-scale only if they require at least 100 Mcore-h on Hawk, or 45 x 1021 FLOP/a on JUWELS, or 45 Mcore-h on SuperMUC-NG.
The GCS Peer Review Board decided to award the status of a large-scale project to 14 projects from various scientific fields. Two of these projects requested computing time on two of the three available systems. In total, three projects were granted 950 Mcore-h on Hawk, another four projects were granted 227 x 1021 FLOP/a on JUWELS, and nine projects were granted 654 Mcore-h on SuperMUC-NG.
For more details on these projects, visit https://www.gauss-centre.eu/results/large-scale-projects/.
Contact: Dr. Alexander Trautmann, email@example.com
from JSC News No. 281, 16 June 2021