Physical chemist recounts how and why he fell in love with distributive computing
When I was an undergraduate student in Mexico in the late 1990s, I was fascinated by the power of distributed computing projects, which harness many individual computers to carry out complex calculations and analysis. These projects are, in a way, the greenest form of computing. They use otherwise unused CPU (central processing unit) cycles from volunteer donors around the world.
In particular, the SETI@home project at the University of California, Berkeley, launched in 1999, was an inspiration: Over the years, this search for extraterrestrial life in radio telescope signals has been powered by the idle CPU cycles of hundreds of thousands of volunteer machines around the world.
When I became an assistant professor at Harvard University back in 2006, I was excited about the possibility of using distributed computing to run the theoretical calculations needed to discover novel materials.
So in collaboration with the IBM World Community Grid, my group and I started the Harvard Clean Energy Project (CEP), an effort to find novel organic electronic materials capable of converting sunlight into energy. Having employed more than 35,000 CPU years of computer time, CEP is now the largest computational quantum chemistry project that’s been carried out to date.
CEP and subsequent projects in my group have taught us how to more efficiently design materials. With our experimental collaborators, we have discovered new types of organic molecules for flow batteries and organic light-emitting diodes.
One of the most satisfying aspects of the project over the years has been the interaction with the project participants in the online forums. Seeing the enthusiasm for scientific discovery among the citizens of the world makes me optimistic that the Internet will continue delivering revolutionary tools that can help us tackle the scientific challenges associated with the 21st century.
Alán Aspuru-Guzik is a professor of chemistry at Harvard University.