University of Virginia Computer Science Professors Receive National Science Foundation CAREER Awards

April 1, 2008 — Kim Hazelwood and Jason Lawrence, both assistant professors in the University of Virginia School of Engineering and Applied Science's Department of Computer Science, have recently won National Science Foundation CAREER Awards, recognizing their promising research and teaching.

The National Science Foundation's Faculty Early Career Development (CAREER) Program makes awards to faculty who, early in their careers, effectively integrate research and teaching and are deemed to have promising futures of continuing that work. The awards, which range from $200,000 to $500,000, are distributed over four or five years. 
Hazelwood and Lawrence's pioneering work in computer science could lead to significant changes in the field. Hazelwood's research aims to comprehensively improve the function of personal computers with an eye toward green solutions for power usage. Lawrence is focused on improving the tools available to designers for digitizing the complex visual appearance of three-dimensional objects. The $400,000 award that each received will help move their groundbreaking research forward.

Hazelwood's CAREER Award was presented in January for research that looks at cross-layer solutions to performance, power, reliability and temperature problems in computers. This research could ensure that the level of speed and reliability improvements the public has become accustomed to continues into the foreseeable future.

Traditionally, researchers concentrated on layers of a computer system stack in isolation. For example, they may focus on the operating system, or the microarchitecture of the circuits themselves. Hazelwood's team is looking at the entire design stack. This approach aims to create solutions where the various layers work together to solve computing challenges, Hazelwood said.

"A nice side-effect of this choice is that it allows us to design 'reactive' solutions," Hazelwood said. "For instance, the underlying hardware is best at detecting temperature problems, but higher layers, such as the operating system or virtual machine, are a much better place to solve those problems. The catch is that those higher layers are not good at predicting when temperature problems may arise. By working together, these various layers can effectively detect and react to problems as they arise."

While consumer demand for increased computer speed will likely remain, Hazelwood expects a trend toward performance that is more environmentally conscious.

"We anticipate users moving away from focusing on raw performance in isolation," Hazelwood said. "While users will always want faster computers, other metrics, such as whether their laptop can last on a coast-to-coast flight on battery power, are moving to the forefront. We imagine a new line of 'green' computers that sacrifice some level of performance for environmental factors such as power usage. Our work explores various ways to make that performance/power trade-off on a case-by-case basis."

In addition to significantly impacting the way computers operate, Hazelwood's work integrates education. She teaches a graduate seminar on virtual-execution environments where students in the classes are given access to all of the research team's simulators and tools and can explore their own cross-layer solutions.

Recently, Hazelwood also received a Career Enhancement Fellowship for Junior Faculty from the Woodrow Wilson National Fellowship Foundation. The goal of this fellowship program is to help talented junior faculty pursue scholarly research and writing for a one-year period; nationwide, only 20 such fellows were selected from all disciplines. Hazelwood plans to spend this time writing a book on the Pin Dynamic Instrumentation System, a project she collaborates on with Intel that enables programmers to easily add new functionality into any arbitrary software even if they do not have the source code for that software.

Computer science colleague Jason Lawrence received a CAREER Award for his research in computer graphics, work that could make it easier for graphic designers to author digital models of the rich visual appearance of real-world materials such as polished wood, brushed metal and human hair. Presently, creating realistic computer models of natural materials is burdened by tedious measurement techniques and a scarcity of structured representations and editing tools. Lawrence's research could be the basis for technology that saves designers from having to manually hand-tune numerous parameters with complex three-dimensional modeling software.

"My vision is to replace the current workflow with a much easier-to-use physical acquisition process," Lawrence said. "Imagine a device physically very similar to a digital camera that allows you to simply snap a few 'pictures' of an object you'd like to import into your virtual world. A piece of computer software built on the principles this research will develop automatically recovers a meaningful and accurate model from that type of unstructured input."

This technology promises to ease the work of graphic designers creating interactive media, historians cataloging 3-D images of statues or investigators recreating a murder scene. Until then, students will be the main beneficiaries.

"Both the creators and consumers of this research will be students," Lawrence said. "I'll be doing this in collaboration with undergraduate and graduate students and working with people in other parts of the University where this research would be relevant — medicine, architecture, design. Besides the fascinating theoretical problems we will study, the unique reflectometry equipment this award will help purchase will give students crucial hands-on experience."

— By Zak Richards

About the University of Virginia School of Engineering and Applied Science
Founded in 1836, the University of Virginia School of Engineering and Applied Science combines research and educational opportunities at the undergraduate and graduate levels. Within the undergraduate programs, courses in engineering, ethics, mathematics, the sciences and the humanities are available to build a strong foundation for careers in engineering and other professions. Its abundant research opportunities complement the curriculum and educate young men and women to become thoughtful leaders in technology and society. At the graduate level, the Engineering School collaborates with the University's highly ranked medical and business schools on interdisciplinary research projects and entrepreneurial initiatives. With a distinguished faculty and a student body of 2,000 undergraduates and 650 graduate students, the Engineering School offers an array of engineering disciplines, including cutting-edge research programs in computer and information science and engineering, bioengineering and nanotechnology. For information, visit