Coeus Cluster

COEUS CLUSTER

Coeus cluster

Specs:

The Coeus cluster consists of 128 compute nodes, 12 KNL Xeon Phi processor nodes (768 cores), 2 large-memory nodes (768GB each), 2 login nodes, 2 management nodes, and 2 storage servers, all connected using Intel'snew Omni-Path interconnect fabric. In total, the compute nodes provide with 2560 cores and 16,384GB RAM.

Getting Started on Coeus:

Please read this document for basics on accessing Coeus:

https://sites.google.com/pdx.edu/research-computing/getting-started/coeus-hpc-cluster

Slurm Scheduler:

SLURM (Simple Linux Utility for Resource Management) is a job scheduler and workload manager. You will need it to manage your work on Coeus. Please read this if you are unfamiliar with SLURM:

https://sites.google.com/pdx.edu/research-computing/getting-started/slurm-scheduler

Current system status:

http://coeus.rc.pdx.edu/ganglia/
(Ganglia is only available within PSU campus network or through the VPN.)

BECOME A MEMBER!

PICS membership is free and is required for access to the Coeus cluster

Coeus cluster installation

All basic equipment parts were acquired by January 2016. The equipment was released in a soft opening for friendly user mode during February 2017.