Diagnosing and reporting operational issues to optimise system usage and performance is challenging on large-scale HPC systems due to their sheer complexity. At the Jülich Supercomputing Centre (JSC), we address this challenge with LLview, an open-source system and job reporting framework. LLview provides near real-time metrics for analysis through a web portal with role-based access for users, administrators, and support staff. In this paper, we present a series of use cases demonstrating how LLview enables efficient diagnosis and resolution of system and application issues, enhancing both reactive and proactive support for HPC users.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Supporting HPC Users with LLview

  • Filipe Souza Mendes Guimarães,
  • Aravind Sankaran,
  • Wolfgang Frings

摘要

Diagnosing and reporting operational issues to optimise system usage and performance is challenging on large-scale HPC systems due to their sheer complexity. At the Jülich Supercomputing Centre (JSC), we address this challenge with LLview, an open-source system and job reporting framework. LLview provides near real-time metrics for analysis through a web portal with role-based access for users, administrators, and support staff. In this paper, we present a series of use cases demonstrating how LLview enables efficient diagnosis and resolution of system and application issues, enhancing both reactive and proactive support for HPC users.