Sign up for our newsletter and get the latest HPC news and analysis.
Send me information from insideHPC:

Video: Argonne’s Pete Beckman Describes the Challenges of Exascale

In this video, Pete Beckman from Argonne National Lab describes the challenges of Exascale computing.

Exascale computing is the next step in how big and how fast we can do science and so it’s ten to the eighteenth operations a second. That’s about a hundred or a thousand times faster than the computers we have now. Right now, we solve a lot of fantastic science problems by using supercomputers. We model the climate. We can design better jet engines. We can look at the human genome. And these problems are fantastic in order to advance mankind, advance our understanding of science. However, at some point, we run up against the limits of our supercomputer. We need it to be faster. And so as we look at the next level of performance, at exascale computing, that’s our challenge. Can we build and design machines that can run a thousand times faster? Now, one of the challenges is software. The software that we have today is designed in a certain way to run on a certain number of processors and, as we look to future machines, the number of processors, the number of cores is dramatically increasing. The future computers that we’re looking at at exascale have millions and millions of cores, which means, as a scientist, programming that is an enormous challenge. Argonne National Laboratory is one of the laboratories helping to lead the exascale push for the nation with the DOE. We lead in a numbers of areas with software and storage systems and applied math. And we’re really focusing, our expertise is focusing on those new ideas, those novel new things that will allow us to sort of leapfrog the standard slow evolution of technology and get something further out ahead, three years, five years out ahead. And that’s where our research is focused.

In related news, Argonne recently held their ATPESC training program on extreme-scale computing. At ATPESC, a group of 65 students and early career researchers spent two weeks in an arduous training program designed to teach them the key skills and tools needed to efficiently use leading-edge supercomputers.

Sign up for our insideHPC Newsletter.

Resource Links: