Both sides previous revision Previous revision Next revision | Previous revisionLast revisionBoth sides next revision |
technical:slurm:scheduler-params [2020-02-20 15:03] – [Implementation] frey | technical:slurm:scheduler-params [2020-03-09 11:21] – frey |
---|
|fair-share| 4000|see ''sshare''| | |fair-share| 4000|see ''sshare''| |
|partition id| 2000|1.0 for all partitions| | |partition id| 2000|1.0 for all partitions| |
|job size| 1|all cores in cluster=1.0| | |job size| 1|all resources in partition=1.0| |
| |
Next to priority access, wait time is the largest factor: the longer a job waits to execute, the higher its priority to be scheduled. This seems appropriate, but it competes against the fair-share factor, which prioritizes jobs for underserved users. | Next to priority access, wait time is the largest factor: the longer a job waits to execute, the higher its priority to be scheduled. This seems appropriate, but it competes against the fair-share factor, which prioritizes jobs for underserved users. |
|''PriorityWeightPartition''| 2000| 0| | |''PriorityWeightPartition''| 2000| 0| |
|''PriorityWeightQOS''| 20000| 3221225472| | |''PriorityWeightQOS''| 20000| 3221225472| |
|''PriorityWeightTRES''| unset| ''cpu=819,mem=245,gpu=245,node=327''| | |''PriorityWeightTRES''| unset| ''cpu=819,mem=245,GRES/gpu=245,node=327''| |
|''PriorityFlags''| ''FAIR_TREE,SMALL_RELATIVE_TO_TIME''| ''FAIR_TREE''| | |''PriorityFlags''| ''FAIR_TREE,SMALL_RELATIVE_TO_TIME''| ''FAIR_TREE''| |
| |
| |
Since the ''PriorityWeightJobSize'' will not be used, the more complex "small-relative-to-time" algorithm will be disabled. | Since the ''PriorityWeightJobSize'' will not be used, the more complex "small-relative-to-time" algorithm will be disabled. |
| |
| The modifications to ''slurm.conf'' must be pushed to all systems. The ''scontrol reconfigure'' command should be all that is required to activate the altered priority caclulation scheme. |
| |
| ==== Addendum: lg-swap partition ==== |
| |
| One of the test nodes containing 6 TiB of NVMe storage has been reconstructed with the NMVe as swap devices rather than as file storage devices. This configuration is currently being tested for viability as a solution for jobs requiring extremely large amounts of allocatable memory: the node has 256 GiB of physical RAM and 6 TiB of NVMe swap. A job that allocates more than 256 GiB of RAM will force the OS to move 4 KiB memory pages between the NVMe storage and the 256 GiB of physical RAM (swapping); the idea is that the NVMe efficiency is more nearly a match to physical RAM (versus a hard disk) that the performance penalty may be lower, making this design an attractive option to some workgroups. |
| |
| A special-access partition has been added to feed jobs to this node. Access is by request only. |
| |
===== Impact ===== | ===== Impact ===== |
| |
^Date ^Time ^Goal/Description ^ | ^Date ^Time ^Goal/Description ^ |
|2019-10-24| |Authoring of this document| | |2019-10-24| |Authoring of this document| |
| |2020-03-09|10:45|Implementation| |