Skip to Content

Department of Cognitive Science

High Performance Computing Facility

The MACCS High Performance Computing facility (known as Fluffy) is a 28 node (56 CPU) beowulf-style cluster available to MACCS researchers to run computationally intensive models.

HPC Cluster

54 CPU Xserve HPC Cluster

Cluster technical detailsSample pic.

The management/login node is an Apple Xserve server running Mac OS X Server 10.4.11. It is powered by dual 2.0Ghz G5 processors with 3GB of ECC memory and 160GB of storage.

The compute nodes are Apple Xserve cluster nodes running Mac OS X 10.4.7. They are powered by dual 2.0Ghz G5 processors with 512MB of ECC memory and 80GB of storage. The cluster has a theoretical peak performance of 243 Gigaflops (Flops = floating point operations per second) for double precision floating point calculations.

A Cisco Catalyst 2948G-GE-TX gigabit Ethernet switch is used for the interconnect as well as the management network.  Job submission is managed with Apple's Xgrid software.

More Information

Xgrid monitor

Screenshot from "XGrid Admin" application showing cluster running 6 jobs