NSF's TeraGrid network is off and running
- By Patricia Daukantas
- Feb 06, 2004
The National Science Foundation's distributed TeraGrid has gone into production mode with about a dozen teams using it to study groundwater pollution, the dynamics of biological molecules and the universe's evolution.
Rob Pennington, interim director of the National Center for Supercomputing Applications in Illinois, said TeraGrid is planned to become the world's largest distributed computing infrastructure, open to researchers chosen by peer review.
TeraGrid supercomputers have a combined peak performance of 4.5 trillion floating-point operations per second. TeraGrid has IBM Corp. Linux clusters, with more than 800 Intel Itanium 2 processors, located at NCSA and the San Diego Supercomputing Center.
Officials hope to add another 11 TFLOPS when newly installed Linux clusters go into production this spring at the Illinois and San Diego centers as well as the Energy Department's Argonne National Laboratory in Illinois and the Center for Advanced Computing Research at California Institute of Technology.
TeraGrid's backbone is a 40-Gbps network between Chicago and Los Angeles, with 30-Gbps pipes connecting the grid sites to the backbone.