COMPUTING FACILITIES IN PEYTON HALL

We pride ourselves on our very strong computational resources. Historically, the computational resources available to researchers in the Department of Astrophysical Sciences have always been state-of-the-art, and our goal is to remain in the forefront of computational astrophysics. We began using Unix circa 1980 (bsd for cognoscenti); at the present time the majority of our systems are running Red Hat Linux.  We are major partners in Tigress, the University high performance facilities.  These faciltities are our big computer engines.

In addition, all faculty, postdocs, and graduate students have modern networked workstations in their offices, with user home directories stored on a large departmental file server. Many users also have laptops; graduate students may purchase these from their research funds.

In addition, the Department maintains a number of computers that are available for heavy computational tasks. These include

• Xeon-based systems with large amounts of RAM for modest computational tasks.
• Xeon-based systems with large amounts of RAM and large amounts of disk, for processing large data sets.
• A number of computers dedicated to processing data from the SDSS.
• A large Beowulf cluster, to which faculty, postdocs and graduate students have access. The cluster consists of a master node, a storage node, and 92 slave nodes. All nodes are dual-Xeon systems, connected by a dedicated switched gigabit ethernet network. The master node and storage node together provide about 2.5 TByte of RAID disk storage. The majority of the slave nodes have 3.06 GHz clock speeds. 42 of the slave nodes have 4 GB of RAM, 42 have 2 GB, and 8 have 1 GB, for a total of 260 GB of RAM

Various research groups also operate computers configured for special projects (including a dual AMD Opteron system with 10 GB of RAM), and graduate students are routinely given access to these systems.

The systems administration staff keeps this hardware working smoothly, along with appropriate software.

Other large Linux and shared-memory clusters are also available for use on campus, and Department researchers regularly use the largest supercomputers available at a variety of national facilities around the country.

On-line technical documentation of our system.