Virginia Tech to build new supercomputer

Aug 05, 2010
Pictured is a small cluster testbed that houses the graphical processing units (GPUs) accelerator cards. HokieSpeed is being designed at Virginia Tech as a new heterogeneous supercomputing instrument based on a combination of central processing units (CPUs) and GPUs. Traditionally, only the CPU has served as the brains in computing instruments, and this new combination is expected to greatly improve a supercomputer’s problem solving capabilities. Credit: Virginia Tech Photo

Two million dollars is coming to Virginia Tech to create HokieSpeed, a versatile new supercomputing instrument for accelerating and transforming discovery and innovation across a myriad of disciplines.

Wu Feng, associate professor of computer science as well as electrical and computer engineering and principal investigator on the grant, along with Khidir Hilu, professor of biological sciences, and Scott King, professor of geosciences, all at Virginia Tech, successfully applied for the $2 million grant through the National Science Foundation's Major Instrumentation Program to invent and build "HokieSpeed."

Feng, who joined Virginia Tech in 2006, is an expert in efficient . He is the founder of the Green500 List, a ranking of the most energy-efficient supercomputers in the world, and considered to be a complementary listing to the Top500 List of supercomputers (http://www.cs.vt.edu/node/5219?nid=5219). He also serves as the faculty co-director for the NSF Center on High Performance Reconfigurable Computing (CHREC), an NSF Industry/University Cooperative Research Center (I/UCRC).

Virginia Tech became a key player in the field of high-end computing in 2003, when it designed and built System X, when the machine was ranked as the fastest academic in the world (November, 2003 TOP500 List). Srindhi Varadarajan, associate professor of computer science, was the chief architect of System X (http://www.cs.vt.edu/facilities/systemx).

Feng described HokieSpeed as a new heterogeneous supercomputing instrument based on a combination of central processing units (CPUs) and graphical processing units (GPUs). In terms of raw performance, he noted that HokieSpeed is expected to deliver 35 times better peak performance, 70 times better peak , and 14 times better peak space efficiency than System X.

"HokieSpeed is expected to catalyze new approaches for conducting research via the synergistic amalgamation of heterogeneous supercomputing and cyber-enabled tools to enhance ease of use. In particular, it will give end users the ability to perform in-situ visualization for rapid visual information synthesis and analysis. It will also control their level of immersion in the discovery process - from being completely immersed, making real-time intuitive decisions via a large-scale gigapixel display, to observing the instrument automatically collect, organize, and analyze data in support of visual analytics. A large set of Virginia Tech researchers from across the university will be actively involved in research using HokieSpeed," Feng said.

The expectations for HokieSpeed come from the researchers' plan to have each compute node of the supercomputer consist of both CPUs and GPUs. Since the 1960s, the CPU has served as the brains in computing instruments. Recent trends, however, have exposed the CPU as a "jack of all (computing) trades, master of none," Feng said, "thus giving rise to heterogeneous computing instruments with multiple types of brains that include both CPUs and GPUs.

"This transformative instrument will empower faculty, students, and staff across disciplines to tackle problems previously viewed as intractable or that required heroic efforts and significant domain-specific expertise to solve," Feng said.

As an example, Feng cited his leadership of a team of more than 20 interdisciplinary researchers from eight institutions around the world in 2007. They developed software cybertool instruments to integrate a set of distributed supercomputers, totaling 12,000 plus CPUs, to enable the discovery of missing genes in 699 microbial genomes. Originally thought to be computationally infeasible, his team completed the task in 10 weeks.

Virginia Tech's Office of the Vice President for Information Technology and the Office of the Vice President for Research provided key enabling cost sharing support for this project.

The three-year NSF grant is from American Recovery and Reinvestment Act funds.

Explore further: Researcher develops method for monitoring whether private information is sufficiently protected

add to favorites email to friend print save as pdf

Related Stories

Georgia Tech wins NSF award for next-gen supercomputing

Oct 21, 2009

The Georgia Institute of Technology today announced its receipt of a five-year, $12 million Track 2 award from the National Science Foundation's Office of Cyberinfrastructure to lead a partnership of academic, industry and ...

New computer cluster gets its grunt from games

Nov 25, 2009

Technology designed to blast aliens in computer games is part of a new GPU (Graphics Processing Units) computer cluster that will process CSIRO research data thousands of times faster and more efficiently ...

New Supercomputing Center in Pittsburgh

Sep 29, 2004

The U.S. academic community will soon have access to a new supercomputer modeled on the highest-performance systems currently being built in the United States, through a $9.7 million award to the Pittsburgh Supercomputin ...

Recommended for you

Tackling urban problems with Big Data

16 hours ago

Paul Waddell, a city planning professor at the University of California, Berkeley, with a penchant for conducting research with what he calls his "big urban data," is putting his work to a real-world test ...

Computer-assisted accelerator design

Apr 22, 2014

Stephen Brooks uses his own custom software tool to fire electron beams into a virtual model of proposed accelerator designs for eRHIC. The goal: Keep the cost down and be sure the beams will circulate in ...

User comments : 1

Adjust slider to filter visible comments by rank

Display comments: newest first

dirk_bruere
not rated yet Aug 05, 2010
Some numbers eg TFLOPS, would have been nice

More news stories

Genetic code of the deadly tsetse fly unraveled

Mining the genome of the disease-transmitting tsetse fly, researchers have revealed the genetic adaptions that allow it to have such unique biology and transmit disease to both humans and animals.