Running HPC and AI Workloads on the NVIDIA Grace Hopper Superchip

Explore the architecture of the Nvidia Grace Hopper superchip in this 26-minute talk that demonstrates how it enables more efficient scaling for both traditional HPC codes and AI workloads. Learn about key HPC sites that have adopted this superchip and how they're effectively utilizing the hardware. The presentation also covers Linaro Forge and its application in debugging and profiling applications running on heterogeneous architectures.