Sub-blocking technique available on FERMI

Dear Users,

   we are pleased to announce you that the sub-blocking technique is now available on FERMI.

You are now able to lunch multiple runs in the minimum allocatable block of 64 compute nodes, by dividing it in smaller blocks of either 32, 16, 8, 4, 2, 1 compute nodes. This enables you to submit jobs in which 2, 4, 8, 16, 32, or 64 simulations are simultaneously running, allowing to exploit at best codes that do not scale up to 1024 cores.

You can find more info at the web page

http://www.hpc.cineca.it/content/batch-scheduler-loadleveler-0#sub-block

Best Regards,

HPC-CINECA user support