GPU Experts Panel: Ampere Explained

9:00 AM - 10:00 AM PDT

Speakers:
Olivier Giroux, Distinguished Engineer, NVIDIA
Carter Edwards, Principal Systems Software Engineer, NVIDIA
Timothy Costa, Product Manager, HPC Software, NVIDIA
CJ Newburn, Principal Architect, HPC, NVIDIA
Christopher Lamb, VP of Compute Software, NVIDIA
Michael Houston, Chief Architect - AI Systems, NVIDIA

An interactive panel of NVIDIA GPU experts to ask questions, give feedback, and get advice on the latest GPU developments, including libraries, hardware, platform support, developer tools, and more.

VIEW SESSION

HPC Summit Developer Forum

9:00 AM - 11:00 AM PDT

Speakers:
CJ Newburn, Principal Architect, HPC, NVIDIA
Ian Buck, VP of Accelerated Computing, NVIDIA
Jeff Larkin, Senior DevTech Software Engineer, NVIDIA
Timothy Costa, Product Manager, HPC Software, NVIDIA
Julie Bernauer, Director, Deep Learning Systems Engineering, NVIDIA
Stephen Jones, CUDA Architect, NVIDIA
Joshua Patterson, Director, Engineering, NVIDIA
Geetika Gupta, Product Lead AI+HPC, NVIDIA

A forum to share your GPU experiences, learn from fellow HPC developers and GPU experts, and help shape future directions in the NVIDIA ecosystem. The forum will kick off with a description of some past success stories where developer feedback helped drive NVIDIA’s direction, continue with short presentations from developers who'll share their top requests, and then proceed to a set of Q&A discussions with our panel of experts. Topics for the Q&A will include programming models and languages, math libraries, HPC + AI, data science, mixed precision, Arm, multi-GPU programming and message passing, and more.

VIEW SESSION

HPC Summit Developer Roundtable Discussions

11:30 AM - 12:30 PM PDT

Attendance is Limited to 150 People for This Session. This is to Ensure the Discussions Can Remain Personal. An Additional Short Registration Step is Therefore Required When Joining the Session.

Speakers:
Max Katz, Senior Solutions Architect, NVIDIA
Harun Bayraktar, Manager, CUDA Math Libraries, NVIDIA
Jim Dinan, Principal Software Engineer, NVIDIA
Jaydeep Marathe, Sr. Compiler Software Engineer, NVIDIA

A set of virtual roundtable discussions between developers and NVIDIA experts occurring in breakout rooms where you’ll be able to provide feedback and pose requests from NVIDIA engineers and product managers. We’ll provide our thoughts on the feedback during the session and afterward. Topics include programming models and languages, math libraries, HPC + AI, data science, mixed precision, Arm, multi-GPU programming and message passing, and more.

This session has concluded.

HPC + AI Networking and Storage

9:00 AM - 10:00 AM PDT

Speakers:
Craig Tierney, Solutions Architect, NVIDIA
Julie Bernauer, Director, Deep Learning Systems Engineering, NVIDIA
Gilad Shainer, Mellanox SVP Marketing, NVIDIA

A presentation followed by a discussion, starting with a closer look at the network data processing for HPC workflows: The ever-increasing demands for higher computation performance to handle the large data volumes and scientific problems complexity drove the creation of In-Network Computing or network based data processing engines. These engines can participate in the application runtime and analyze application data as it's being transferred within the data center, or at the edge. This technology is already in use in leading HPC and deep-learning data centers, have been integrated into multiple MPI and deep learning frameworks, and have demonstrated accelerated performance with a variety of compute- and data-intensive applications. We'll review the InfiniBand state of the art, and the roadmap for future in-network computing technologies. Following this presentation will be a community- driven Q&A session on networking and storage topics.

VIEW SESSION

Integrating Cloud Tools to HPC Workflows

10:00 AM - 11:00 AM PDT

Speakers:
Adam DeConinck, Solutions Architect, NVIDIA
Alan Chalker, Director of Strategic Programs, Ohio Supercomputer Center

A presentation followed by a discussion on traditional command-line based methods for accessing computational tools on HPC systems provides a barrier to many potential researchers that could benefit greatly from those resources. Just as many consumer and enterprise related computational providers discovered over the past decade, by providing a modern, web-based gateway to supercomputer resources, many HPC providers are lowering that barrier to entry and improving productivity of their clients across a broad range of science and engineering disciplines. This presentation will provide some details and lessons learned from the Open OnDemand project, one of the most popular web platforms currently deployed at numerous HPC centers. Following this presentation will be a community driven Q&A session.

VIEW SESSION

What the Arm + HPC Ecosystem Can Enable You to Do

11:00 AM - 12:00 PM PDT

Speakers:
Scott McMillan, Solutions Architect, NVIDIA
Brent Gorda, Senior HPC Executive, Arm

A presentation followed by a discussion, sharing the rapid progress made by the Arm and HPC ecosystem. This rapidly changing market opens new avenues for innovation and innovative uses of HPC. We'll start with some examples of Arm HPC projects, and encourage you to tell us what's most important and what you see as challenges. Following this presentation will be a community driven Q&A session.

VIEW SESSION