By tapping into and consolidating unused computing power, the World Community Grid is helping fast-track research—including finding new PV materials.
Essentially a virtual supercomputer, the World Community Grid (WCG) pools the spare time of volunteered personal computers across more than 80 countries to provide nonprofit research projects with high-speed computing that their researchers otherwise could not afford. Its massive power can accelerate research time at a fraction of the cost.
“It’s safe to say that World Community Grid would consistently place in the top 15 supercomputers if it were a traditional, physical supercomputer,” says IBM spokesperson Ari Fishkind. “When a university or other institution purchases a supercomputer, it is usually shared by hundreds, if not thousands, of researchers. Rarely does one single research project have access to the equivalent of an entire supercomputer to itself, 24/7. However, this is what the researchers effectively get with World Community Grid.”
WCG uses the idle time of Internet-connected computers to perform research calculations. Once downloaded by the volunteer user, WCG software works in the background, using spare system resources to process small computing assignments. When computations are completed, the software sends the results back to the server and requests another assignment. To ensure accuracy, the WCG servers send out multiple copies of each work assignment, and researchers validate the results.
“While the central processing unit [of the user’s personal computer] does consume additional power when it is processing an item on the to-do list, the rest of the machine is the major consumer of power—just by being plugged in and turned on. World Community Grid puts that otherwise wasted power to good use by computing some research that will ultimately help humanity and the world,” says Viktors Berstis, the WCG’s lead architect. “And, we don’t ask people to keep their computers turned on more than they normally would. The system is designed to handle the pauses and resumption in calculations without any problems.”
The WCG can be completely customized by users so that their PCs are always running quickly and efficiently. For instance, WCG can be the lowest-priority task for the PC so that it instantaneously relinquishes control to the user’s “normal” work. The default is for WCG to use a computer’s central processing unit (CPU) at levels of 60%—meaning that if the computer hits 60% for all tasks it is doing, then WCG stops using it. Users may lower the utilization even further if desired.
“Users can instruct WCG to start crunching numbers when it’s clear that the machine isn’t in constant use, such as if their PCs have been idle for a certain number of minutes. On the flip side, if a user has a modern, fast PC and wants to make greater contributions to WCG, then the computer can crunch numbers in the background while a user does lightweight tasks such as checking email,” Berstis says.
Since WCG’s launch in 2004, nearly 600,000 individuals and organizations have volunteered spare time on 2 million computers and completed 1 billion computation assignments to advance a total of 21 research initiatives. At any given time, the WCG supports a variety of projects—12 currently. Such projects include efforts to develop cancer, malaria, and AIDS drugs; identify healthier and hardier strains of rice for developing countries; and engineer better means for filtering water. Users are included in all projects by default but may opt out of projects as they choose.
The research supported by WCG has yielded some 33 peer-reviewed papers, including one from Harvard University that discusses a new organic compound that may have the potential to form a new generation of flexible and lightweight solar cells. Because physically creating material with those molecules on a trial-and-error basis is too slow, the Harvard Clean Energy Project, led by theoretical chemist Alán Aspuru-Guzik, is relying on the WCG to complete computational models that will identify molecules that may make good semiconductors. All told, the team plans to screen as many as 10 million molecules using their automated quantum chemical calculations. According to Aspuru-Guzik, the WCG allows them to characterize about 25,000 candidates every day, compared to the few tens or hundreds that a conventional study could investigate in a month. So far, the initiative has inspected more than 3 million molecules. The team identified one new compound and shared the findings with researchers at Stanford University, who synthesized the molecule and confirmed its properties as a potential semiconductor.
“Nobody has ever computed so much quantum chemistry calculations together. This is actually terabytes and terabytes of data. The question is how to mine it, how to understand it,” says Aspuru-Guzik. “This is actually what we are concentrating on right now, and the WCG allows us to do that.”
Learn more about the World Community Grid (WCG) at worldcommunitygrid.org.