摘要
神威计算机系统提供了强大的并行计算和批处理能力,代表了高性能计算机发展的新方向。作为系统软件的重要组成部分,作业管理系统可以根据用户的需求,统一管理和调度系统的软硬件资源,保证用户作业合理地使用机器资源,提高了系统利用率和吞吐率。该文主要介绍了神威高性能计算机系统的作业管理系统及其批式作业调度模块的设计思路和实现。
Sunway serial supercomputing system provides parallel computing and a powerful batch capability and represents the trend of high performance computer development. As an important component of system software, job management system provides centralized management and scheduling of hardware and software resources according to the users requirements, ensuring that computer resources are available to all users jobs and increasing the utilization and throughput of the system. This paper mainly introduces the job management system and the design and implementation of its batch job management system scheduling module in sunway.
出处
《计算机工程》
CAS
CSCD
北大核心
2004年第13期47-49,186,共4页
Computer Engineering
关键词
作业管理
大规模并行机
资源管理
Job management
Massive parallel process
Resource management