allow to limit cpu usage to the number of slots requested

Hello,, currently, when a user submit a job, there's no way in LSF to limit the cpu consumption. recently, we discover that it's possible to limit cpu usage OS side using cgroup, by putting a value in the 'cpu.cfs_quota_us' file of the '/sys/fs/cgroup/cpu/lsf/gnb/job.' directory (multiple of 'cpu.cfs_period_us' file): cd /sys/fs/cgroup/cpu/lsf/gnb/job. #multiply cpu.cfs_period_us value by the number of cpus to limit #ex to limit to 20 cores: cat cpu.cfs_period_us 100000 echo `expr 20 \* \`cat cpu.cfs_period_us\` ` > cpu.cfs_quota_us cat cpu.cfs_quota_us 2000000 in this ex, the sum of the cpu consumed by the job won't be able to go over 2000% (viewable using top), corresponding to a -n20 bsub option. maybe it's possible to change the effect of the 'LSB_RESOURCE_ENFORCE' parameter ! smart additional feature, is possible: a factor 'tolerance' like the CPU_FACTOR parameter: a way to weight the user request: 'factor_value' x 'slot_request' or 'slot request' + 'factor_value' (both ways are interesting). for our 20 slots job ex: 20x1.05 or 20+1 to limit the cpu consumption to 21 cpus (2100% from top) with this new feature, we will reduce drastically cpu overload on servers and guaranty jobs run time for all end-users. Regards, Olivier PS: my last request, the 'cores limitations using affinity [SPCLSF-I-1439]' is less versatile: tasks jobs are bound on cpu and we cannot allocate more cpus than the number available on the server.