What is a compute cluster?
A compute cluster, also known as a “supercomputer,” is a set of computers that all work together to form one functional system. The cluster has many individual computers in it (called compute nodes) that are all set up to run computational jobs on. Each node can run independently or can pool resources with other nodes to increase computing capacity. These computers are all linked to a central controller computer with similar hardware to a compute node that runs software like the cluster’s job scheduler.
In these guides, the terms “compute cluster,” “cluster,” and “supercomputer” are all used interchangeably. “Campus compute cluster” or “campus cluster” refers to R2 and Borah, the two clusters Boise State has purchased.
Why use a compute cluster?
The advantage to utilizing a cluster comes in the form of parallel computing, or the ability to run computational jobs over many CPUs or GPUs at once. This can improve the performance of a job and decrease the amount of time it takes to normally run by orders of magnitude. Parallel computing is applicable to many different fields, including education, raptor biology, environmental science, and genetics – researchers all across campus have had success utilizing our resources to improve the efficiency of their research.