Berkeley Lab Researchers Lead Parallel Computing Research Center at UC Berkeley
March 18, 2008
Computing sciences researchers at Lawrence Berkeley National Laboratory are part of a team leading a new research center in a partnership with Intel and Microsoft to accelerate developments in parallel computing and advance the powerful benefits of multicore processing to mainstream consumer and business computers.
Microsoft Corp. and Intel Corp. announced today the creation of two Universal Parallel Computing Research Centers (UPCRC), the first at UC Berkeley and another at the University of Illinois at Urbana-Champaign. This is considered the first joint industry and university research alliance of this magnitude in the United States focused on mainstream parallel computing.
"This is a once-in-a-career opportunity to recast the foundations of information technology and influence the entire IT industry for decades to come," said David Patterson, UC Berkeley professor of computer sciences. A pioneering expert in computer architecture, Patterson also is a scientist in Berkeley Lab’s Computational Research Division. "We are excited and proud to be a part of this ambitious effort."
The funding for UC Berkeley's UPCRC, which Patterson directs, forms the foundation for the campus's Parallel Computing Laboratory, or Par Lab, a multidisciplinary research project exploring the future of parallel processing.
Over the next five years, Intel and Microsoft expect to invest a combined $20 million in the two university centers, including $10 million at UC Berkeley. Researchers at the UC Berkeley center have also applied for matching funds through UC Discovery Grants, a program supported by the state of California to foster economic development through partnerships with industry and academia.
Research from this effort is expected to lead to powerful, new mainstream applications, such as cell phones that can recognize the face of an approaching acquaintance and, more importantly, whisper that person's name into the user's ear, or speech recognition software that can act as a court reporter by providing an accurate, written record of what was said by numerous people in a court or conference room.
"It is important for industry to work in tandem with academia to unleash the immense power of parallel computing," said Tony Hey, corporate vice president of External Research at Microsoft Research. "We are privileged to have a dedicated research partner like UC Berkeley and look forward to partnering with them to transform the way multi-core technology is developed and used in the future."
In addition to Patterson, the center includes seven other UC Berkeley faculty members from the Department of Electrical Engineering and Computer Sciences as principal investigators: Krste Asanovic, Ras Bodik, James Demmel, John Kubiatowicz, Kurt Keutzer, Koushik Sen and Kathy Yelick.
Yelick also is the new director of the National Energy Research Scientific Computing Center (NERSC) at Berkeley Lab. Demmel is a researcher in Berkeley Lab’s Computational Research Division. John Shalf, a researcher in computer architectures at NERSC, also participated in meetings that led to the creation of Par Lab.
The Par Lab not only bring together researchers focused on software and hardware, but also experts in disciplines ranging from music to medical technology to help push state-of-the-art computing applications in a wide range of fields.
"The researchers at UC Berkeley have a long track record of making important contributions in computing software and hardware," said Andrew Chien, vice president of Intel's Corporate Technology Group and director of Intel Research. "As importantly, UC Berkeley has a long tradition of leadership in the academic research community in catalyzing infrastructure and change as exemplified by the Unix Berkeley Software Distribution, which was the de facto system software infrastructure for the academic community for over a decade."
The concept of parallel computing, in which a task is divided up into smaller bits to be completed simultaneously to achieve greater speed and performance, has been around for more than 40 years, but it has been the domain of highly specialized programmers and scientists.
New "duo-core" and "quad-core" products that squeeze multiple processors onto a single, integrated circuit have successfully met the increased demand for performance in desktop and mobile computers, but researchers and industry experts are looking to a day when a computer chip will hold hundreds of processors.
"These so-called many-core processors present some significant challenges to computer programming that we will address at this center," said Patterson. "As an analogy, eight reporters writing the same story could potentially write a story eight times faster. But to achieve this increased speed, one would need to break up the task so that each reporter had the same amount of work to do. If anything went wrong and just one reporter took longer than the seven others, then the benefits of having eight writers would be diminished. You would also fall short if one part of the story, such as the conclusion, couldn't be written until all of the other parts were completed, creating a bottleneck."
The challenge ahead for the technology industry lies in bringing to mainstream developers and, ultimately, consumers, the benefits of multi-core processing based on tens or even hundreds of cores.
"The challenges of scheduling, load balancing and limits to parallelism apply to parallel programming as well as to the analogy of parallel story writing," said Yelick. "And as you might expect, the more people or processors involved, the stiffer the challenges."
UC Berkeley's Department of Electrical Engineering and Computer Sciences was selected to host one of the two UPCRC sites from among 25 of the country's leading computer science departments by a joint Intel-Microsoft technical evaluation committee.
"This selection reflects UC Berkeley's outstanding reputation in computing," said Patterson, who is also part of the Center for Information Technology Research in the Interest of Society (CITRIS), an innovative public-private partnership between four UC campuses, the state of California and industry. "Our computer science department is consistently ranked among the top three programs in the country, and it is second-to-none in the specialty of computer systems. Moreover, a recent National Academies report noted that UC Berkeley researchers have invented technologies that led to seven multi-billion dollar IT industries, more than any other university in the country."
Research funds for the new center will be used to advance parallel programming applications, architecture and operating systems software. In addition to the director and seven principal investigators, the center at UC Berkeley will host six members of the UC Berkeley faculty and 50 doctoral students and postdoctoral researchers. Software developed by the center will be provided to the technology community for additional development.
More information about the Par Lab is online at http://parlab.eecs.berkeley.edu.
About NERSC and Berkeley Lab
The National Energy Research Scientific Computing Center (NERSC) is a U.S. Department of Energy Office of Science User Facility that serves as the primary high-performance computing center for scientific research sponsored by the Office of Science. Located at Lawrence Berkeley National Laboratory, the NERSC Center serves more than 7,000 scientists at national laboratories and universities researching a wide range of problems in combustion, climate modeling, fusion energy, materials science, physics, chemistry, computational biology, and other disciplines. Berkeley Lab is a DOE national laboratory located in Berkeley, California. It conducts unclassified scientific research and is managed by the University of California for the U.S. Department of Energy. »Learn more about computing sciences at Berkeley Lab.