RCS High Performance Computing (HPC) Documentation
Contents
Chinook
In 2015, the Geophysical Institute launched Chinook as an energy efficient linux cluster purchased from Penguin Computing, Inc. Chinook is named in honor of long time GI colleague Kevin Engle's unique, strong, collaborative nature and passion for salmon and Alaska. Chinook is made possible by the Geophysical Institute, IARC, the Institute of Arctic Biology, the Vice Chancellor of Research, and your fellow colleagues who contribute shares. In 2016 and 2017, Chinook expansions were also supported with funding from the M. J. Murdock Charitable Trust.
Chinook is the foundation for an energy-efficient, condo-style HPC cluster for UA researchers. The computing environment hosted on Chinook includes:
105x Relion 1900 Compute Nodes each with dual Intel Xeon 12- or 14-core processors (24 or 28 cores per node) and 128 GB memory
26x Relion XE1112 Compute Nodes each with dual Intel Xeon Scalable 20-core processor (40 cores per node) and 192 GBs memory
3x Relion 1900 Compute Nodes each with dual Intel Xeon 14-core processors (28 cores per node) and 1.5 TB memory
2x Relion 1903GT Compute Nodes each with dual Intel Xeon 14-core processors (28 cores per node), 128 GB memory, and 3 dual GPU Tesla K80m accelerators
Multiple login nodes with dual Intel Xeon processors and 48 or more GBs memory
CentOS operating system, Slurm open-source workload management software, and Scyld ClusterWare HPC management software
Compute and login node access to the 307 TB Lustre scratch file system
Are you interested in using the Chinook HPC cluster in your computational work? Please read our directions on how to obtain RCS project and user accounts.
XSEDE/Campus Champions
Members of RCS staff are part of XSEDE's Campus Champions program, providing support and contacts for XSEDE systems and XSEDE startup allocations. Please contact RCS if you are interested in pursuing XSEDE resources.
Last updated