Great Lakes is the shared U-M campus Linux high-performance computing (HPC) cluster. It is available for use in research and teaching by all faculty, graduate students, undergrads, postdocs, research scientists, etc. on campus, including in Michigan Engineering. CAEN and U-M Advanced Research Computing (ARC), are jointly responsible for supporting members of the Michigan Engineering community who use the Great Lakes.
The Great Lakes cluster is composed of 13,000 cores across 380 nodes, with at least 5 GB of RAM per core. Great Lakes is a batch-oriented system, so compute nodes are intended to allow long-running CPU, memory, or I/O-intensive jobs. Login nodes are used to submit jobs to the batch scheduling system or compile software but may not be used to run jobs. More information about Great Lakes is available on the ARC website.
The following information will help you get started using Great Lakes:
- In order to use Great Lakes, you will need a Great Lakes user login. To apply, fill out the login application.
- Great Lakes requires multi-factor authentication using Duo. If you haven’t already, you should Enroll in Duo.
- Review the quick overview of using Great Lakes, which can be found in the “Great Lakes User Guide.”
- Michigan Engineering provides a shared Great Lakes resource account, engin1, that can be used by faculty, students, and staff in the College. This account is meant for small-scale exploration and is not suitable for research projects or coursework. Courses that require the use of Great Lakes should have their own account, and researchers should purchase an account for their lab or project.
- Great Lakes uses Slurm to handle the scheduling of jobs. Jobs can be submitted to the job queue using a Slurm script, or can be run interactively.
- Software: A number of open-source and commercial software packages are available on Great Lakes. The LMOD environmental module system is used on Great Lakes to manage different versions of available software. LMOD can provide an up-to-date listing of available software by using the module spider command.
Help with Great Lakes
Contact us at [email protected] with any questions or concerns you have about using Great Lakes.