HPC | ECBE | SIU

High Performance Computing Facility

High-performance computing (HPC) refers to the use of supercomputers and/or computer clusters to accelerate the solution of fundamental problems in science, engineering and business that have broad scientific and economic impact. With a generous support from NSF (Award No. 0855221), SIUC has developed an HPC center (SIHPSI: Southern Illinois HPC Infrastructure), a facility first-of-its-kind not only within the campus but in the greater Southern Illinois region also. SIHPCI initially consists of 106 dual quad-core Dell PowerEdge R410 compute nodes running Red Hat Enterprise Linux and has been fully operational since Spring 2010. Each node has 8 (eight) 64-bit Intel Xeon 2.23 GHz E5520 CPUs and 8 GB of RAM and is connected with Gigabit Ethernet and supported by appropriate hardware. 90 TB of storage is installed behind the Master Node. The machine offers access to scratch space.

SIHPSI is expected to expand the scope and quality of research at SIUC in two broad areas: (1)Computational nanoscience and engineering (CNE): Projects include quantum simulations of nanoscale devices with tens of millions of complex degrees of freedom; computational design of catalysts at the molecular level; large scale computations of multiphase flows such as solidification of binary alloys, dynamics of red and white blood cells in arteries, boiling phenomena in energy generation and electronic cooling, and enhancement of heat and mass transfer by bubble columns in bioprocesses; molecular dynamics studies of polymer morphology at interfaces; fundamental studies of the non-equilibrium states of matter; and quantum information processing exploring new properties of atomic nuclei. (2) Geographic Information Science (GIS): Involves strategic research to investigate new algorithms for representation and transformation of massive dynamic data allowing a cognitive and visual interpretation for analysts by exploiting invariant geometric properties. New representations of the large-scale datasets have the potential to change the way people utilize the data for knowledge discovery.

Computational research in the nanoscience area will support experimental sciences in the pursuit of scientific discovery and technical innovation and is expected to have significant impact in a wide range of technological applications including low-power and fast transistors, coatings, lithography, adhesives to light emitting diodes and sensors, various smart and functionalized materials, and quantum computation. While the GIS research will facilitate efficient data streaming, crime and health studies, medical imaging, and genome mapping. SIHPCI, as the first community HPC infrastructure at SIUC, will serve as a nucleation center for further purchase/addition of HPC resources. The facility would offer new faculty members a much quicker time frame (1-2 weeks) to be up and start computing as compared to a custom cluster configuration which usually takes about 6 months from start to production.

SIHPCI will be closely tied to major educational activities within the campus, and will have a significant impact on SIUC's curriculum development at both undergraduate and graduate levels. SIHPCI is expected to address the needs of more than 20 faculty members (including the SIHPCI investigators Shaikh Ahmed, Mesfin Tsige, Mark Byrd, Tonny Oyana, and Qiang Cheng) spanning over 7 departments. Also, in accordance with SIUC's long tradition of service to its community and region, SIHPCI will play an outstanding role in the greater Southern Illinois area through training a diverse community of college teachers/instructors as well as K-12 students in the field of parallel computing and data analysis and thus will support the overall economic vitality and sustainability of the region.

Earlier in 2009, SIHPSI principal investigator Shaikh Ahmed, an Associate Professor in the Department of Electrical and Computer Engineering in the College of Engineering, was one of four researchers nationwide included in Oak Ridge National Laboratory's first High-Performance Computing Grants Competition that allowed his research group access to ORNL's Jaguar supercomputer and other top-end computing platforms and staff housed at the site in Tennessee. Ahmed's group uses the massive computing power to conduct research in petascale modeling and designing of nanoscale devices for use in harsh-environments.

Visit the High-Performance Computing website for more information.