This essay by Charles Day first appeared on page 88 of the January/February 2012 issue of Computing in Science & Engineering, a bimonthly magazine published jointly by the American Institute of Physics and IEEE Computer Society:
My title comes from a comment made on Physics Today‘s Facebook page by Fernanda Foertter, a physicist who programs high-performance computers for a biotechnology company.
Although Foertter’s computational science background lies mostly in molecular dynamics simulations of polymers, her comment was about this post I wrote on colliding galaxies:
Here’s a great example of using computer simulation to help interpret observations. Jennifer Lotz of Space Telescope Science Institute and her colleagues modeled pairs of galaxies merging into each other. Stills from her movies were then compared with Hubble images of galaxies that looked as though they had just merged or were about to merge. The comparison yielded a new, more accurate estimate of the galaxy merger rate.
Until I encountered Foertter’s enthusiastic outburst, I hadn’t thought of supercomputers as being inspirational. As a science writer, I’ve seen plenty of stunning simulations of exploding supernovae, wiggling proteins, and other phenomena. I’ve written about climate calculations that gobbled up weeks of supercomputer time. Several Nobel Prizes, I know, have been awarded for work that required the services of high-performance computers.
But now I’ve come to realize that supercomputers are not just useful, they’re glamorous, too. What’s more, their awesome power could be used to encourage schoolchildren to think about careers in computational science.
To see what I mean, consider what is perhaps the most ambitious, most glamorous field of physics: particle physics. When I was in high school, I read Nigel Calder’s The Key to the Universe: A Report on the New Physics (Viking Press, 1977), which I found in my local library. There within its pages, in accessible prose accompanied by photos and diagrams, was the quest to discover the ultimate constituents of matter and the laws that govern their behavior.
Back in 1977, the world’s most powerful particle accelerator was Fermilab’s Main Ring, whose circumference and maximum collision energy were 6.4 km and 400 gigaelectronvolts. The current record holder, CERN’s Large Hadron Collider, is 27 km in circumference and is designed to reach 7 teraelectronvolts. When the LHC ended its latest science run in October, it had smashed together 7 × 1014 protons and antiprotons.
To me, supercomputing—or high-performance computing, if you prefer—is the particle physics of computational science. The world’s fastest computer, K, consumes 10 megawatts of electricity to carry out 8 × 1015 floating-point operations per second. The problems that K and other supercomputers are programmed to tackle are among the toughest and most important in all of science, such as understanding Earth’s changing climate and figuring out how 1011 interconnected neurons form a thinking human brain.
As I write this column, Supercomputing 2011 is being held at the Washington State Convention Center in Seattle. I was glad to see that the meeting’s education track has 19 talks altogether, including one entitled “Parallel: HPC Overview” by Charlie Peck and his colleagues.
Attending a lecture or class is still work to a student, no matter how interesting the topic. But reading a captivating book is play, and therefore more likely to fire a student’s imagination. I’ve just looked on Amazon for an inspiring book on supercomputing. I couldn’t find one.