CAC Systems

The following systems provide the computing power and architectural diversity needed to develop, evaluate, and improve diverse approaches to parallel computing.  By exploiting the synergy between computer science, engineering research, and application development, we hope to bring efficient parallel computing into the mainstream of academic and industrial research and development.  The CAC has assembled a "computational testbed" of moderate sized parallel computing systems. 

The CAC plans to expand on this computational testbed in the future. Potential users are encouraged to contact us to learn of our current configurations, new developments, and further affiliations.

Nyx - AMD Opteron/Intel Nehalem 64-bit cluster running Red Hat Linux

Over 5,000 Opteron and Nehalem cores with an average of 2 GB RAM/each
Head node: nyx-login.engin.umich.edu

This system is available for use by everyone in our community after you request an account.

Flux - Intel Nehalem/I7 64-bit cluster running Red Hat Linux

Over 6,000 Nehalem cores with an average of 4 GB RAM/core
Head node shares: flux-login.engin.umich.edu

This system is available for use by everyone in our community after you request an account.

System Downtime

We'll announce our downtimes on each systems /etc/motd (which is shown each time you login to a particular system. We're also announcing downtimes on the main page of the CAC site. You can keep an eye on that page, or better yet, use your RSS reader such as Google Reader or Bloglines ) to let you know when anything is happening here at the CAC.