Thanks to a $27.5 million grant from the Semiconductor Research Corp., the University of Virginia’s Department of Computer Science is tackling one of the most pressing problems in computer science and engineering — the so-called “memory wall,” reports the Daily Progress.
As health care, science and technology systems grow more and more data-intensive and analytics become more sophisticated, current computer systems are unable to feed data to the processor fast enough, wasting time and energy. That gap between needed power and ability, first articulated by UVa professor emeritus William Wulf and graduate student Sally McKee, is often called the “memory wall.”
The 20 faculty members on Skadron’s team will investigate how to rebuild the entire computer processing system, from better memory chips and wiring to new software that can process complex and fragmented problems. They are focusing on several specific areas of application, including advanced genomics, new cancer biomarkers and predictive home health care.
“The volume of data here, as well as the ways that we have to mine it, are basically impossible to work with in today’s systems,” Skadron said. “But if we can make that tractable, we can help do things like diagnose cancer patients.”
From an economic development perspective, the grant sounds like unalloyed good news. Kevin Skadron, chairman of the computer science department, says that he expects the university will garner some intellectual property rights from the research, providing funding for more research and innovation at the university. Who knows, perhaps UVa’s computer science school will form the nucleus of a Charlottesville-based innovation ecosystem that spins off new enterprises and highly paid private-sector jobs.
The program also could contribute to Virginia’s workforce development. According to a university press release, the center will create opportunities for undergraduate students to get involved in research as well as internships with companies that are program sponsors. Tens of thousands of IT jobs in Virginia are going unfilled, and this program conceivably could increase the supply of qualified workers (although it’s not clear if Ph.D.s with expertise in computer design match up with the skills demanded by Virginia’s IT companies).
Before we get too excited, it’s also worthwhile asking, “Who’s paying for all this?” As the Daily Progress notes, “Over the past few years, UVa’s School of Engineering & Applied Sciences has been pouring investment into cyber infrastructure.”
In other words, UVa had to make a big investment to get the computer-science program to the point where it could attract the “memory wall” research grant. How big of an investment? The Daily Progress did not think to ask, and UVa apparently did not volunteer a number. Moreover, there is the matter of overhead. Research contracts typically allocate a negotiated percentage of the grant to cover overhead such as lab space, computers, the labor-intensive grant-application apparatus, and general administration. It would be useful to know how much this research grant pays for, and whether it fully covers all costs.
In recent posts, Bacon’s Rebellion has asked to what extend university R&D programs are subsidized by undergraduate tuition. Interestingly, the new Center for Research in Intelligent Storage and Processing in Memory, or CRISP, will fund positions for about 100 Ph.D. students. One might infer from the wording of the press release that the graduate-student positions will be fully funded by outside money. If that inference is correct, perhaps we can likewise assume that graduate student stipends are not being subsidized by undergraduate tuition. But we don’t know for sure until someone asks.
UVa and Virginia’s other research universities are not accustomed to much oversight of their research programs. Research institutions have been free to pursue their dreams of R&D glory by running monies through an accounting black box. Meanwhile, state policy is to foster university research, even to support it with General Fund dollars, in the name of economic development, so no one in state government is asking questions.
The Daily Progress reporter didn’t ask where the money is coming from, what investments UVa made to reach the point where it qualified for the grant, and what ongoing obligations it might incur that aren’t covered by the grant. That’s no slight on the reporter — until a few months ago, I would have assumed that the grant meant, “Whoopee! Free money!” Indeed, I still don’t pretend to know how the system works. But I am going up the learning curve, and I’ll keep asking questions.