Exascale Intelligence: Leveraging GB300 and A4X Max for the Next Frontier of Scientific HPC and Generative AI
March 3, 2026
9:00 AM - 10:00 AM PST
Online
Embedded Files
About the Session
The arrival of the NVIDIA GB300 Grace Blackwell Ultra Superchip marks a paradigm shift in how we approach the most demanding computational challenges. In this session, Dan Ernst will explore the architecture and real-world impact of the GB300.
The presentation will provide a deep dive into the "Ultra" enhancements of the Blackwell architecture, including the second-generation Transformer Engine with native FP4 support and the 1.8 TB/s NVLink-C2C interconnect that fuses the Grace CPU and Blackwell GPUs into a single, high-bandwidth logical unit.
Dan will also outline the software ecosystem essential for maximizing this hardware, highlighting CUDA-X libraries, NVIDIA Magnum IO, and the NeMo Framework. Attendees will see performance benchmarks demonstrating how the GB300 delivers up to a 1.27x training speedup over the GB200 and a staggering 50x increase in AI factory output compared to previous generations. Finally, the session will showcase notable results in scientific simulations—from climate modeling to drug discovery—and the latest in agentic and multimodal AI reasoning.
Speaker
Dan Ernst
Senior Director of Supercomputing System Technology, NVIDIA
Dan Ernst is the Senior Director of Supercomputing System Technology at NVIDIA, where he leads pathfinding and architectural efforts for the next generation of NVIDIA’s supercomputing products. With a career spanning over 20 years in high-performance computing, Dan previously served as a Principal Architect at Microsoft and held senior technical leadership roles at Hewlett Packard Enterprise (HPE) and Cray Inc. Dan holds a Ph.D. in Computer Science and Engineering from the University of Michigan, specializing in high-performance, low-power, and fault-tolerant microarchitectures. A recognized leader in the HPC community, he frequently contributes to the TOP500 and ISC High Performance forums, focusing on the intersection of advanced system architecture and scientific discovery.
Google Sites
Report abuse