Management host selection
To achieve the highest degree of performance and scalability, use a powerful management host.
LSF has no minimum CPU requirement. For the platforms on which LSF is supported, any host with sufficient physical memory can run LSF as management host. Swap space is normally configured as twice the physical memory. LSF daemons in a cluster on Linux x86-64 use about 488 MB of memory when no jobs are running. Active jobs use most of the memory that LSF requires.
Note: If a Windows host must be installed as the management host, only supported
Windows Server operating systems are recommended as LSF
management
hosts.
The following table outlines recommendations for host selection:
| Cluster size | Active jobs | Recommended system memory | Recommended server CPU (Intel, AMD, Power, ARM or equivalent) |
|---|---|---|---|
| Less than 100 hosts | 100 | 32 GB | 4 cores |
| 10,000 | 64 GB | 8 cores | |
| 100 to 1,000 hosts | 10,000 | 128 GB | 8 to 16 cores |
| 50,000 | 128 GB | 8 to 16 cores | |
| 1,000 to 5,000 hosts | 50,000 | 256 GB | 16 to 24 cores |
| 500,000 | 256 GB to 512 GB | 24 to 48 cores | |
| 5,000 to 10,000 hosts | 500,000 | 512 GB to 1 TB | 48 to 96 cores |
Additionally, note for your host selection and the table recommendations:
- Larger clusters benefit from higher clock speed, faster memory, and more cores.
- Memory requirements vary depending on the number of hosts and the number of jobs retained in memory (pending, running and completed).
- The numbers in the table are for guidance only. Actual performance can vary based on site-specific variables, such as job submission rates, query rates, network performance, total number of jobs in the system, and the performance of external services.
- For very large deployments, discuss these with your account technical lead or technical account manager.