By Matthew Dublin

More and more, the concept of exascale-class supercomputers (that's 1,000,000,000,000,000,000 calculations per second) is moving from merely being the subject of whimsical high-performance computing conference panels to serious thought about how and when these machines will become a reality. The Obama administration's Strategy for American Innovation listed exascale computing among the Grand Challenges of the 21st century, explicitly calling for a "an exascale supercomputer capable of a million trillion calculations per second – dramatically increasing our ability to understand the world around us through simulation and slashing the time needed to design complex products such as therapeutics, advanced materials, and highly-efficient autos and aircraft.”

Last week, J. Craig Venter's op-ed piece in Nature made mention of the need for exascale class computers in the not-so-distant future when DNA sequencing becomes a commodity and researchers turn their attention to phenotypes. Experiments that look at the connections between human genetic variation and biological outcomes will require the complete genomes of tens of thousands of humans combined with a range of data that incorporates family history and clinical manifestations associated with disease.

Venter writes that "even if we had all this information today, we wouldn't be able to make use of it because we don't have the computational infrastructure to compare even thousands of genotypes and phenotypes with each other The need for such an analysis could be the best justification for building a proposed 'exascale' supercomputer, which would run 1,000 times faster than today's fastest computers."

And in another opinion piece over at International Science Grid This Week, Irving Wladawsky-Berger, a 37-year IBM veteran, lays out some of the big challenges on the way to achieving exascale computing. Whereas the evolution from terascale to petascale went smoothly using tens to hundreds of thousands of processors from the PC and Unix markets, they will not get us to exascale, writes Wladawsky-Berger. Exascale will require some other kind of major transition in chip architecture, not to mention a completely new programming paradigm.