Scheduling Methods in HPC System: Review

Abstract

Parallel and distributed computing is a research area that is complex and rapidly evolving. Researchers are trying to get more and more methods and technologies in parallel and distributed computing. Most users who use Pay as you do go design, elasticity, and virtualization mobility and automation make cloud computing an attractive option to meet the development needs of some HPC users. Over the years, the research algorithm has also changed accordingly. The basic principles of parallel computing, however, remain the same, such as inter - process and inter - processor communication schemes, parallelism methods and performance model. In this paper, HPC system, scheduling methods and challenges are discussed and given some potential solutions.