Over the past decade, there has been a growing awareness of the multi-faceted benefits we can derive from data-driven strategies like Quantitative Codesign. This increasing awareness, along with improvements in Machine Learning (ML) technologies, has driven vendors, operations staff, and application developers to provide an ever-increasing level of instrumentation into their products. The time is right for turning this vast trove of available information and the incredible advances in analysis technologies it represents into appropriate knowledge and understanding. Multiple reports have proposed quantitative codesign strategies to create a feedback loop that could assist vendors and software developers in their design processes. [1,2,3,4,5]. Indeed, Quantitative Codesign is essential for addressing challenges brought about by the recent trend of increasing heterogeneity and varied accelerators in HPC architectures.
● Jim Ang- Pacific Northwest National Laboratory, USA
● Jim Brandt – Sandia National Laboratories, USA
● Ann Gentile – Sandia National Laboratories, USA
● Michael Jantz – the University of Tennessee, USA
● Terry Jones – Oak Ridge National Laboratory, USA (Chair)
● Estela Suarez- Jülich Supercomputing Centre, Germany