The Itasca compute cluster provides resources for running production batch and interactive jobs

What is Itasca?

Itasca is an HP Linux cluster with a total of 9,712 compute cores and 25 TiB of main memory:

  • 1,091, 8 core Nehalem nodes, where each node is an HP ProLiant BL280c G6 blade server, each with two quad-core 2.8 GHz Intel Xeon X5560 "Nehalem EP" processors sharing 24 GiB of system memory.

  • 51, 16 core Sandy Bridge nodes, where each node is an HP Proliant BL460c G8 blade servers, each with two 8-core 2.6 GHz E5-2670 Xeon "Sandy Bridge EP" processors with 64, 128 or 256 GiB of memory.

These nodes are connected with a 40-gigabit QDR InfiniBand (IB) interconnect.

What can I do with Itasca?

Itasca batch jobs are scheduled on a per node basis, so that each job has the requested nodes dedicated to that user (cores are not shared with other users). So users can submit single node jobs, or large multi-node MPI jobs to the Itasca batch queues, and have those nodes dedicated to their job. Users can also schedule an interactive sessions and similarly reserve compute nodes for use in an interactive environment.

See also the documentation for job queues.

How do I access Itasca?

Itasca is most efficently accessed though a terminal enviroment, MSI users can follow directions in the Connecting to HPC Resources quick start guide to access Itasca and all other MSI HPC systems. Once connected to Itasca you will be able to submit jobs to Itasca's queues using scripts with specific PBS commands also known as PBS scripts.