24-29 March 2024
BHSS, Academia Sinica
Asia/Taipei timezone

A workunit scheduling system of the HEP@home project for reliable volunteer computers

Not scheduled
20m
BHSS, Academia Sinica

BHSS, Academia Sinica

Poster Presentation Track 9: Converging High Performance Computing Infrastructures: Supercomputers, clouds, accelerators

Speaker

嘉龙 程 (高能物理研究所)

Description

To run workunits effectively, the BOINC server will stop sending workunits to those volunteer computers that make too many errors for some time. However, during BOINC apps development, workunit errors may arise from time to time. Therefore, to make development and debug not affected by errors, a workunit scheduling system is designed for the HEP@home project. With this scheduling system, reliable volunteer computers, especially those only for BOINC app development will get new workunits as soon errors get corrected but not waiting for quite a long time. This scheduling system consists with a new scheduling algorithm for reliable computers, an interface for volunteers to report errors are corrected, and a pilot App to verify the reports from volunteers. This paper will present the design and development details of the scheduling system.

Primary author

嘉龙 程 (高能物理研究所)

Co-authors

Prof. 京燕 石 (中国科学院高能物理研究所) Prof. 然 杜 (中国科学院高能物理研究所)

Presentation materials

There are no materials yet.