Early career stars
rise to data challenge

Posted February 21, 2012

Streaming computations partition a 3-D domain into smaller blocks. Click image to enlarge and for more information.

Fast in supercomputing used to be so simple. The fastest supercomputer was the one that could perform the most flops, perhaps landing the machine a coveted spot on the TOP500 listing of the world’s supercomputers.

Now scientists doing high-performance computing are slowing down and taking a new look at what fast really means with today’s supercomputers and the new problems they’re trying to solve.

Two ASCR-supported Department of Energy Early Career Research Award recipients (see sidebar, “Personal data movement”) are at the leading edge of this transformation to supercomputing that can do more with far less energy, and that can mix speed and scale to mine genetic data for meaningful links as no computer has done before.

Supercomputing’s data-intensive energy crisis

Early Career Research awardee Peter Lindstrom of Lawrence Livermore National Laboratory (LLNL) wants to head off “a huge energy problem when it comes to next-generation computers. Today when you get time on a supercomputer, you’re typically charged in terms of the amount of time spent, and that’s going to change because power’s really going to be the key factor.”

It’s estimated that a next-generation exascale supercomputer (capable of a million trillion flops) envisioned for LLNL will consume about 50 megawatts of electricity – the same amount as the entire city of Livermore, Calif., population 80,000. A power-guzzling supercomputer doesn’t just raise costs; it slows computing. There are concerns that during warmer months a next-generation supercomputer would run at only half speed, given the challenges of keeping it cool enough to function. Already this balance between energy use, computational speed and actual usability has resulted in the creation of the Green500, a ranking of the world’s most energy-efficient supercomputers.

“We’re going to have to rewrite our codes to be more power efficient,” Lindstrom says. “And one of the key aspects of this is the power used in data movement. On next-generation supercomputers the power cost of moving data is really the critical metric for software.”

1   |   2   |   3   |   4   |   Print       Next »