Difference between revisions of "About"
| Line 15: | Line 15: | ||
This cluster uses the SLURM [https://slurm.schedmd.com/documentation.html] workload manager, and some familiarity with | This cluster uses the SLURM [https://slurm.schedmd.com/documentation.html] workload manager, and some familiarity with | ||
it will be beneficial. | it will be beneficial. | ||
| + | |||
| + | Up: | ||
| + | [[ HPC_Usage| HPC_Usage]] | ||
Latest revision as of 15:25, 1 April 2025
Overview
This cluster is made up of a set of GPU servers, with upto 96 cores and x2 A30 NVidia compute cards. The interconnect is Gigabit ethernet, and storage is offered over NFS. Neither Lustre nor infiniband nor NVlink is available. We strongly encourage the use of this cluster with codes that are GPU enabled, and can take full advantage of the hardware. If as a user you are unaware of the capabilities of your code, please request for assistance through the KENET support email, we are happy to answer your questions.
Storage is available on the login node (limited), and in a $SCRATCH directory that is linked to through a soft link in your home directory. Please retain only necessary data, any generated files that have served their purpose should be deleted immediately to preserve space.
This cluster uses the SLURM [1] workload manager, and some familiarity with it will be beneficial.
Up: HPC_Usage
