Running jobs
ACTnowHPC scheduler
The ACTnowHPC system uses SLURM as the workload manager. To ensure fair access to all resources, you will not have direct access to the compute nodes, but must run all jobs through the scheduler.
Choosing your partition
The scheduler has multiple partitions (this is often referred to as a queue in other systems). These partitions have different hardware characteristics.
Each partition has a different billing rate, and you can find your cost for each partition when you login to the customer portal. GPU and Phi nodes will be coming online soon. Watch this space for more information.
| Partition | CPU | Freq | Cores per node | Memory per node | Accelerators | Notes |
|---|---|---|---|---|---|---|
| hsw-fdr | 2x Intel Xeon E5-2650v3 "haswell" | 2.3GHz | 20 | 64GB | none | Default time limit is 5 min, maximum is 24 hours. If you need longer running jobs contact us at support@actnowhpc.com |
| bw-fdr | 2x Intel Xeon E5-2550v4 | 2.2GHz | 24 | 64GB | none | Default time limit is 5 min, maximum is 24 hours. If you need longer running jobs contact us at support@actnowhpc.com |
More details