NVIDIA to Describe Path Forward for Accelerated Computing at SC15 Show

by Greg Estes

As the world of supercomputing gathers for its biggest event of the year this week in Austin, Texas, NVIDIA will be at the center of it – with a talk by our CEO, presentations by international researchers and demonstrations of the latest technology.

Starting tomorrow, we’ll describe our vision for just what’s possible today in the world of computation, machine learning and visualization to 10,000 of the world’s leading researchers and technologists attending SC15.

Jen-Hsun Huang headshot
NVIDIA CEO Jen-Hsun Huang will discuss the path forward for accelerated computing at SC15.

Kicking off the show will be a presentation on the path forward for accelerated computing by NVIDIA CEO Jen-Hsun Huang and Ian Buck, who runs our Accelerated Computing business. They’ll also describe our latest work that makes it possible for web-services companies to accelerate their huge machine learning workloads, using the new hyperscale datacenter line for our Tesla Accelerated Computing Platform.

At our show floor booth, we’ll feature a schedule of some 40 talks from the industry’s leading figures on parallel computing, machine learning and interactively visualizing data using supercomputers. For example:

  • Diego Rossinelli, of ETH Zurich – The ACM Gordon Bell finalist will discuss the uDeviceX, a petascale in-silico lab-on-a-chip. Nearly 19,000 Tesla K20X GPU accelerators are providing massive  muscle to help this microfluidic device track down tumor cells that are markers of metastatic cancer.
  • Mathieu Luisier, of ETH Zurich – Also an ACM Gordon Bell finalist, he will discuss GPU-enabled simulation and visualization of nanoelectronic devices, like transistors, memristors and solar cells. The work better predicts performance before fabrication, and sheds light on their working principles.
  • Bryan Catanzaro, of Baidu Research — Will detail how his team scaled the training of deep neural networks for end-to-end speech recognition to multiple GPUs on multiple nodes, as well as their use of GPUs to deploy DNNs to end users.
  • Ruchir Pori and Rajesh Bordawekar, of IBM Thomas J Watson Research Center – Will discuss how GPUs are increasingly the compute engines for achieving dramatic improvements in the performance of cognitive and machine learning workloads.

We’ll also feature in our booth demos of the latest advances in simulation, big data analytics and deep learning, including:

  • In-situ weather simulation
  • An interactive tornado visualization
  • Ray-traced quantum-level visualization
  • Drug reaction prediction
  • Image content classification

Developers can get hands-on training and discussion time with experts on subjects including parallel programming and deep learning. There will also be a special area for the community to tell their stories of scientific discovery, which we’ll be sharing via social media.

Those attending SC15 should swing by the NVIDIA Community Café, located adjacent to the Austin Convention Center. Lunch is on us, with a variety of food and a great chance to connect with the larger community to talk about advancements that are rocking the industry. Not able to make it? Catch the replays of the talks above at the SC15 GPU Technology Theater page starting Nov. 17.