Wildebeest Cluster

Wildebeest is a 132-processor Beowulf cluster, located at City University of New York, and administered by the Center for Algorithms and Interactive Scientific Software (CAISS) at City College. It is a joint project between Lehman College, City College, and CAISS, supporting primarily the research activities of Professors Katherine St. John (Lehman College), Sean Cleary (CCNY), Damian Rouson (CCNY), and Gilbert Baumslag (CCNY). The research projects are evolutionary tree reconstruction in computational biology, genetic algorithms in combinatorial group theory, computer experiments on parafree groups, modelling turbulent fluid flows arising in the early stages of combustion, and questions and simulations in computer networking.

The cluster is currently configured with two head nodes (one for queuing and servicing jobs on the slave nodes, and one for developement and testing) and 64 dual-processor slave nodes .

Why the Name?

Two reasons, really. First, Prof. Baumslag is from South Africa and all of the machines he's involved with are either named for South African animals or past pets. The second is wildebeest is a synonym for gnu:

From the American Heritage Dictionary: WILDEBEEST: See GNU. ETYMOLOGY: Obsolete Afrikaans: Middle Dutch wild, wild + Middle Dutch beeste, beast (from Old French beste; see beast).

Information for Research Students

To use wildebeest, you need to:

write up a short description with a few sentences about your project and what programs you expect to use. (Also, including estimates of how much space, time, and processors you need would be very helpful.)
have Prof. St. John send a note to the system adminstrator to have your account set up,
read Moret's article on designing good computer experiments:
Moret, B.M.E., "Towards a discipline of experimental algorithmics," Proc. 5th and 6th DIMACS Implementation Challenges, M. Goldwasser, D.S. Johnson, and C.C. McGeoch, eds, DIMACS Monographs 59, American Mathematical Society, Providence, 2002. PS
(TexShop can translate postscript to pdf, or use the command ps2pdf at a terminal window.)
read about the queuing system: see http://research.gc.cuny.edu for an overview and documentation.

After debugging and testing your programs, you need to contact Prof. St. John with a time and resource estimates. Since the machine is a shared resource, the estimates are needed to plan usage and system downtime.

We gratefully acknowledge the NSF for their generous support of this project via
NSF-MRI 02-15942: "MRI: Parallel Computing Environment for Computational Mathematics"
(PI: St. John, co-PIs: Baumslag, Cleary, and Rouson).