Migration from Marconi A1: general informations and update about RCM

Dear users,

in order to support the migration of your projects from Marconi A1 to Galileo/Marconi A2/Marconi A3 we summarize the main differences between Broadwell, Knights Landing and Skylake nodes:

BDW (Galileo):
- 36 cores per node and no hyper threading
- 118000 MB of memory per node

KNL (Marconi A2):
- 68 cores per node and hyperthreading enabled: 4 threads per core, for a total of 272 virtual cpus
- 86000 MB of memory per node
- exclusive use of the nodes
- you may need to specify the binding of MPI tasks and OpenMP threads to cpus

SKL (Marconi A3)
- 48 cores per node and no hyperthreading
- 182000 MB of memory per node
- exclusive use of the nodes
- you may need to specify the binding of MPI tasks and OpenMP threads to cpus

You can find the documentation concerning these arguments at the following web pages:
https://wiki.u-gov.it/confluence/display/SCAIUS/UG3.1%3A+MARCONI+UserGuide
https://wiki.u-gov.it/confluence/display/SCAIUS/UG3.3%3A+GALILEO+UserGuide
https://wiki.u-gov.it/confluence/display/SCAIUS/UG2.5.1%3A+Batch+Scheduler+SLURM
https://wiki.u-gov.it/confluence/pages/viewpage.action?pageId=182255681

Finally, an update about RCM and visualization on MARCONI: as some of you has already noticed, the service is currently unavailable. This is also because of the shutdown of A1; we are working towards a migration of the service on SkyLake nodes, and we will keep you updated as soon as it will become available again.

Regards,
HPC User Support - CINECA