Close

Presentation

HiPar23: 4th Workshop on Hierarchical Parallelism for Exascale Computing
DescriptionTwo conflicting time scales are emerging in High Performance Computing (HPC): the fastest dictated by the progress of "hardware'', and the slowest by the "software''. Hardware is rapidly changing: compute nodes feature increasingly more physical cores over multiple sockets and accelerators, and complex memory hierarchies. This is driven by several factors, including intense industrial competition, the demands of AI and machine learning, and the always present leading-edge scientific research. Software is progressing more slowly, thus increasing the urgency of finding solutions to narrow this gap in a sustainable way. Hierarchical parallelism is one approach that shows great promise. Its main strength is to embrace the hardware complexity by exploiting parallelism at all levels: compute, memory and network. This workshop aims at bringing together hardware and software practitioners proposing new strategies to fully exploit computational hierarchies, and examples to illustrate their benefits for extreme scale computing.
Event Type
Workshop
TimeMonday, 13 November 20239am - 12:30pm MST
Location704-706