Engineering Compute Clusters
These compute clusters are part of the College of Engineering's high performance computing (HPC) resources for scientific research and teaching. They consist of state of the art computing nodes, high compute power and memory, scheduling software, and general purpose engineering applications.
Buying into the Cluster
Researchers may purchase compute power in terms of compute nodes. Costs are listed below:
1 node with dual 6-core Intel Sandy Bridge E5-2640 processors and 256GB RAM (plus usage of other compute nodes in the cluster as available*) is $1,900 per year, or $160 per month.
1 GPU node with dual 6-core Intel Sandy Bridge E5-2640 processors, 64GB RAM, 3 GTX780 GPU cards, and 8TB scratch storage space (plus usage of other GPU nodes in the cluster as available*) is $4,130 per year, or $350 per month.
You may also choose to purchase compute nodes to add to the cluster at the current rate of:
Purchasing a node to be placed into the cluster will guarantee 100% compute resources to the nodes purchased* plus all other compute nodes as they are available. If you choose this option, you will be responsible to refresh your compute nodes every four years. We will allow a compute node to remain in the cluster for a maximum of six years.
To purchase compute power, email Shaila Parashar with the following information:
When there are additional resources available at the time of job submission, they will be assigned to the job, thus increasing the compute power of your jobs at no additional cost. For example: A user group has purchased one node with 6 processors. At the time of job submission, the job will be assigned more than 6 processors if they are available. There is a high likelihood of getting more compute power than your group has purchased.
This document last modified Thursday May 24, 2018