Curbing Supercomputers' Growing Drain On Energy
RENEE MONTAGNE, host:
We turn now to the growing power of supercomputers.
Supercomputers are crucial for analyzing complex problems like climate change and particle physics. But as they grow, so does their appetite for electricity.
Reporter Lauren Sommer of member station KQED looks at efforts to make supercomputing more energy efficient.
(Soundbite of background noise)
LAUREN SOMMER: Meet the Cray One.
Mr. DAG SPICER (Senior Curator, Computer History Museum): If you needed an icon for a supercomputer, you would use the Cray One.
SOMMER: This six-foot-tall tower of wires was the most powerful supercomputer the world had ever seen in 1976. Today, it's part of the Computer History Museum in Silicon Valley, where Dag Spicer is a senior curator.
Mr. SPICER: It blew people's minds. It was so powerful, so fast.
SOMMER: Of course, in today's terms...
Mr. SPCIER: It's roughly equivalent to a first-generation iPhone from Apple.
SOMMER: The reason we don't play "Angry Birds" on a supercomputer today is thanks to something called Moore's Law.
Mr. SPICER: Moore's Law is a predication made by Intel co-founder Gordon Moore in 1965 that the number of transistors incorporated in a chip will double approximately every 12 months.
SOMMER: What that means is computer chips have gotten smaller and faster at an incredible rate over the last five decades. Which leads us to...
(Soundbite of door opening)
Mr. JOHN SHALF (Computer Scientist, Lawrence Berkeley National Lab): This is our new Cray XE6 supercomputing system.
SOMMER: Also known as Hopper, the nickname that's used by computer scientist John Shalf and his colleagues at Lawrence Berkeley National Lab. And right now, it's chewing on some complicated problems.
Mr. SHALF: Particle accelerator design. We have fusion energy, and then we also have laser plasma inertial fusion simulation.
SOMMER: Okay, some very complicated problems. It's as powerful as 10,000 desktop computers. But in just a few years, Berkeley Lab will get a new model that's as powerful as 100,000 desktops. According to Moore's Law, that next-generation supercomputer should be smaller. But...
Mr. SHALF: The problem is that now we can't make them go any faster.
SOMMER: Shalf says computer chips have hit a wall.
Mr. SHALF: So we can cram more things on the chip, but if you make them go fast, it's so hot they'll melt.
SOMMER: If chips themselves aren't faster, supercomputers will simply have to add more and more of them to increase computing power, which means much more energy use. Hopper uses around three megawatts of electricity, about as much as 2,000 homes.
Mr. SHALF: Projections say that at the end of the decade, we'd be at 100 megawatts if we continue.
SOMMER: That's enough power for a small city. The electricity bill alone would be roughly $100 million a year.
Mr. SHALF: What that says is that our current approach to doing supercomputing is a dead end, and that we need to think of dramatically new ways to improve the efficiency of computing.
SOMMER: Shalf says that could be done with some very familiar technology.
(Soundbite of duck quacking cell phone ring)
Mr. SHALF: Hello?
SOMMER: Cell phones have computer chips that are designed to be energy efficient, so your phone lasts longer on a charge. Shalf says why not build a supercomputer with chips that combine millions of these small cell phone processors?
Mr. SHALF: In using that approach, we're able to demonstrate an additional 80 times more energy efficiency than business as usual.
SOMMER: Which would mean a much more practical supercomputer. Whether or not it gets built depends on companies like Intel, who would design the chips.
Shalf says Berkeley Lab hopes to use this cell phone-inspired computer to model some of the trickier effects of climate change like rainfall patterns, ice sheet melt and the effects of clouds.
Mr. SHALF: It enables policymakers to have the tools they need to make important decisions that have trillion-dollar consequences.
SOMMER: That way, Shalf says, supercomputers could work on tomorrow's climate and energy challenges without contributing to the very problem they're trying to solve.
For NPR News, I'm Lauren Sommer, in San Francisco.
NPR transcripts are created on a rush deadline by a contractor for NPR, and accuracy and availability may vary. This text may not be in its final form and may be updated or revised in the future. Please be aware that the authoritative record of NPR's programming is the audio.