Booth or Dare: Our Show Within a Show at SC16

by Brian Caulfield

We put together a supercomputing conference inside this week’s SC16 supercomputing conference, hosting a series of talks at our booth that had the event’s record 14,000 attendees crowding the aisles and craning their necks for a chance to learn more.

More than three dozen sessions are unspooling virtually every half hour this week in our gleaming green-and-black trimmed booth, smack in the center of the Salt Lake City convention center’s exhibition hall.

Held at our Technology Theater, the series of talks and panel discussions are drawing crowds ranging from dozens to hundreds. Among the luminaries featured are top figures in the supercomputing field from University of Michigan, the Barcelona Supercomputing Center, the Southern California Earthquake Center (SCEC), Baidu, the U.S. National Oceanic and Atmospheric Administration, Sandia National and Microsoft. And that was just Tuesday.

The series kicked off Monday night, in the opening hours of SC16, when NVIDIA CEO Jen-Hsun Huang explained how the AI boom will create a path to exascale computing, one of the supercomputing world’s loftiest goals.

“Several years ago deep learning came along, like Thor’s hammer falling from the sky, and gave us an incredibly powerful tool to solve some of the most difficult problems in the world,” Jen-Hsun said. “Every industry has awoken to AI.”

Other highlights include:

  • Deep learning researchers like Baidu’s Greg Diamos called on supercomputer scientists to build them bigger, better deep-learning systems. “If you don’t have a large enough computer, you can be stuck waiting years or decades for a result,” he said.
  • Phil Maechling, from the SCEC, explained how GPUs power a new generation of earthquake models that give scientists the ability to build sophisticated earthquake propagation models.
  • NVIDIA Chief Scientist Bill Dally spoke about the role GPUs played in sparking the development of deep learning, and how GPUs can accelerate both deep learning and high performance computing. Deep learning is even being used, he said, to solve complex scheduling problems on supercomputers — using AI computing to help HPC run more efficiently.
  • A panel including Jack Wells, of Oak Ridge National Laboratory; Jackie Chen, from Sandia National Lab; SCEC’s Philip Maechling; Galen Shipman, from Los Alamos National Laboratory; and Bill Kramer, from the NCSA at the University of Illinois at Urbana Champaign, spoke about the prospects for a new generation of exascale computers to drive scientific innovation.

In our Developer Zone, two dozen people at a time, crowd around tables equipped with headphones and laptops for hands-on training in deep learning and GPU programming.

Nearby, dozens of scientists are filing into our booth to talk, on camera, about how they were using GPUs as part of our “Share Your Science” series — giving those training up on GPU programming techniques something to aspire to.

Just a few steps away, visitors are donning VR headsets to explore a 100 million atom simulation of how bacteria turns sunlight into energy.

We’re also showing the hundreds of people crowding into our booth how you can tap into GPUs hosted on Amazon Web Services, IBM Cloud and Microsoft Azure.

If you find yourself waking up in Salt Lake City, stop by our booth at SC16, we’re here through Thursday.