High-performance Computing

Science and engineering increasingly rely on computational simulation to better understand the cosmos, genomics, the design of space and unmanned vehicles, the atmosphere and oceans, biochemistry, biodiversity and forest fires, to list but a sample of ongoing large-scale scientific projects. These computing requirements have grown by order of magnitude. In order to investigate "big science," research teams have turned to supercomputers - parallel and distributed computing platforms. The field of high-performance computing (HPC) has developed to further the design of the architecture, algorithms, visualization and numerics associated with large-scale scientific computing.

At WPI, active HPC research exists in the areas of genomics, two-phase fluid flow, turbulence, combustion, quantum devices and numerical algorithm development.

The following resources are available to support the efforts of computational researchers at WPI:

Display Wall (Power Wall)

The all is a large-scale data visualization system producing a 2x2 tiled stereo or a 2x4 tiled mono display. The images are generated by an eight-node cluster of commodity PCs that operate eight high-resolution projectors. The use of the wall allows researchers to investigate large data sets impossible to view on standard workstations. The wall provides a development environment for computer science visualization research.

Linux Cluster

As part of the global shift toward cluster computing, WPI's HPC group provides a distributed computing platform consisting of 10 compute nodes (20 Intel P IV Xeon processors). Available to the whole community for large-scale computing, this platform provides supercomputing capability at a fraction of the traditional cost. The initial user community includes members of Biology, Bioinformatics, Bioengineering, Mechanical Engineering and Chemical Engineering.

For additional information, contact Sia Najafi. (snajafi@wpi.edu).

Maintained by itweb
Last modified: Jun 17, 2009, 11:35 EDT
[WPI] [IT] [Back] [Top]