Technical presentation - 30 minutes (including q&a)
This talk looks at the architecture of the Grace Hopper super chip and explains how the architecture is enabling traditional HPC codes as well as AI workloads to scale more efficiently. The talk will touch on some key HPC sites who have adopted the Grace Hopper super chip and explain how those sites are making effective use of the hardware. Finally the talk will touch on Linaro Forge and how it is used to debug and profile applications running on heterogeneous architectures.
Rudy Graduated from the University of Reading in 2011, where he studied Cybernetics and Electronic Engineering. He held various Engineering positions in Arm for over 10 years, mainly focused on development and quality aspects of Arm based debuggers and profilers. Rudy has been involved in HPC tools since 2017 and now works as a Principal Field Application Engineer at Linaro, spreading his technical expertise and knowledge to help users optimize, debug and port their HPC codes.