Ken Kennedy, director of Rice's Center for High Performance Software Research (HiPerSoft), has been tapped to lead a high-profile, five-year $8.25 million project to develop Virtual Grid Application Development Software. VGrADS are software tools that will simplify and accelerate the development of applications for grid computing -- massive systems of linked supercomputers that form a single, virtual supercomputer. This project will be a partnership with six other schools: University of California, Santa Barbara; University of California, San Diego; University of Houston; University of Illinois, Urbana-Champaign; University of Southern California Information Sciences Institute; University of Tennessee, Knoxville.
"It takes a degree in computer science to understand today's parallel computers and several more years of training to write code to use them efficiently," said Kennedy, University Professor, the Ann and John Doerr Professor in Computational Engineering and Computer Science and professor of computer and electrical engineering. "VGrADS hopes to change that with tools that bring the power of grid computing within reach of the average scientist."
The VGrADS grant is one of 183 ITR grants announced today by the NSF. The awards, which total $169 million, support projects chosen from a field of 2,200 proposals. Rice won nine awards and also garnered a lead role in the $7.5 million 100 X 100 Project, which, along with VGrADS, is one of the eight large multi-institutional ITR projects awarded this year.
The 100 X 100 Project aims to develop the fundamental technologies needed to provide 100-megabit-per-second Internet access to 100 million households nationwide. That's about 100 times faster than DSL or cable modems, and equipping most of the households and businesses in the country with connections that fast will require revolutionary retooling of the Internet, said Rice's Ed Knightly, one of five principal investigators on the project.
"Comparing the 100 X 100 Internet to today's Internet is like comparing today's Internet to the global telephone networks of the early 1970s," said Knightly, associate professor of electrical and computer engineering. "In everything from the backbone to the last-mile connections, a revolutionary approach -- as opposed to an evolutionary one -- will be required to realized the 100 X 100 vision."
Other Rice ITR grants awarded today include:
* Optical Control in Semiconductors for Spintronics and Quantum Information Processing
Principal investigator: Junichiro Kono, assistant professor of electrical and computer engineering.
This $2.7 million program will develop optical methods for controlling electronic, magnetic, vibrational and excitonic properties of semiconductors for ultrafast information processing. Optical control methods are much faster than the electrical switching methods used for today's semiconductors. This project will look at four promising methods of optical control: (1) Optical control of ferromagnetism in magnetic III-V semiconductors, (2) optical control of band structure via the dynamic Franz-Keldysh effect, (3) optical control of electric fields in GaN/InGaN strain superlattices and (4) optical control of excitons in coupled quantum wells. The collaborative research program involves theorists, experimentalists and materials scientists from Rice, the University of Florida, the University of California at San Diego and the Tokyo Institute of Technology.
* Wireless Transit Access Points: Enabling a Scalable, Deployable, High-Performance Wireless Internet
Principal investigator: Ed Knightly, associate professor of electrical and computer engineering
The goal of this $2.4 million project is to develop the hardware and protocols needed for a high-performance, scalable, wireless Internet that supports true wireless "broadband" to residences and public spaces at rates of several megabits per second. Today's cellular and WiFi wireless networks cannot achieve this vision, due to excessive costs of the wired backhaul network, poor performance scaling and excessive costs of spectral license fees. This project will design an architecture based on Transit Access Points (TAPs), devices that use high-performance directional-antenna wireless links, operating in the unlicensed band, to form a wireless backbone mesh.
* Model Reduction of Dynamical Systems for Real-time Control
Principal investigator: Dan Sorensen, Noah Harding Professor of Computational and Applied Mathematics
This $2.2 million project will provide efficient and robust methods for producing reduced order models of large state-space systems. This activity is expected to have an impact on system theory of complex systems, parallel numerical linear algebra for large-scale problems and the efficient implementation of these schemes on parallel and distributed computing platforms. Once the theory and computational methods are developed, the investigators expect that high-quality software will result and have applications in many areas of engineering. This will enable the design of real-time controllers for complex systems. The project is a collaboration between Rice, Purdue University and Florida State University.
* Single Spin Measurement for Quantum Information Processing
Principal investigator: Alex Rimberg, assistant professor of physics and astronomy
Quantum information processing and computing offer the prospect of new technologies that rely fundamentally on quantum phenomena for their operation. For such technologies to be successful, means of storing and reading out quantum information must be developed. This $680,000 project will focus on using the electronic spin to store quantum information. Spin, unlike charge, interacts weakly with its surroundings and better maintains quantum coherence. The same weakness of interaction, however, makes readout of individual electronic spins difficult. This project, which is in collaboration with the University of Wisconsin, Madison, will focus on developing two different schemes for detecting the individual spins in semiconductor quantum dots.
* Novel Finite Element Techniques for Modeling of Flows of Microstructured Liquids in Complex Geometries with Free Surfaces
Principal investigator: Matteo Pasquali, assistant professor of chemical engineering
This collaborative project between Rice and researchers at the Technical University of Munich, Germany, involves the development of new computational methods and software for modeling complex fluid flows in processes where free surfaces, elastic boundaries and moving boundaries are present. The software will be used for modeling important engineering applications, such as blood flow and blood damage in blood pumps, coating and polymer processing flows, ink-jet printing, flow in microfluidic devices and studies of the effects of stress on white blood cells and on the operation of the immune system. In particular, this $360,000 project will directly benefit the design of centrifugal left-ventricular assist devices, which offer hope for thousands of heart-disease patients waiting for donor hearts.
* Prototype High-Performance, Threat-Adaptive, Space Storm Simulation and Forecast Model Supported by a Data Assimilative, Grid Computing Infrastructure
Principal investigator: Richard Wolf, professor emeritus of physics and astronomy
This collaborative project, which involves the University of Michigan, Rice, and several other institutions, centers on research on how to improve space weather modeling by incorporating data assimilation and on-demand grid computing capability. The aim of the project, for which Rice will receive $260,000, is to develop a first-principles, physics-based model of the Sun-Earth system that has the capability of making complete space plasma simulations, from the Sun's surface to the upper atmosphere of Earth. One of the goals is to develop an adaptive system that can respond to significant events in the magnetosphere by incorporating additional computational grid resources as required. If successful, this research could lead to better tools for forecasting space weather.
* Grid-Embedded Interactive Analysis Framework Based on ROOT
Principal investigator: Pablo Yepes, senior faculty fellow in physics and astronomy
This collaborative project between MIT and Rice aims at implementing and testing computer-grid technologies in ROOT, a broadly used analysis framework that supports the analysis of large datasets. Testing is planned in a fully operational prototype environment. The main research topics for this $200,000 project are the optimization of resource discovery and allocation algorithms, as well as real-time monitoring, and fault tolerance in an interactive environment. Prototypes of the system will be used in studies for the Compact Muon Solenoid (CMS), one of the experiments under construction at the Large Hadron Collider (LHC) at the CERN Laboratory in Geneva, Switzerland. The data from this experiment is on the scale of several petabytes per year.
Rice University is consistently ranked one of America's best teaching and research universities. It is distinguished by its: size--2,700 undergraduates and 1,850 graduate students; selectivity--10 applicants for each place in the freshman class; resources--an undergraduate student-to-faculty ratio of 5-to-1, and the fifth largest endowment per student among American universities; residential college system, which builds communities that are both close-knit and diverse; and collaborative culture, which crosses disciplines, integrates teaching and research, and intermingles undergraduate and graduate work. Rice's wooded campus is located in the nation's fourth largest city and on America's South Coast.