Monitoring MPI running nodes status for load balance

Authors:
Qianni Deng;Xugang Wang;Dehua Zang
Affiliations:
Department of Computer Science, Grid Computing Center, Shanghai Jiao Tong University, China;Department of Computer Science, Grid Computing Center, Shanghai Jiao Tong University, China;Department of Computer Science, Grid Computing Center, Shanghai Jiao Tong University, China
Venue:
GCC'05 Proceedings of the 4th international conference on Grid and Cooperative Computing
Year:
2005

Citing 1
Cited 1

The Paradyn Parallel Performance Measurement Tool

Computer

Dynamic load balancing for I/O-intensive applications on clusters

ACM Transactions on Storage (TOS)

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper presents an improvement on XMPI which is a MPI debugging tool based on LAM/MPI platform. The improvement is to monitor the involving physical nodes timely, gather the status information and display the status snapshots of the physical nodes. This function is an effective debugging method, which makes developers find their program bugs and load balance problems more effectively and efficiently. Especially after releasing of MPI-2 spec which allows developers to add and delete nodes dynamically during the MPI program running, it makes more significant in monitoring the physical nodes status.