A Checkpoint and Restart Mechanism for Parallel Programming Systems

Sameer Paranjpye

Thesis 2000

Pulication Type: MS Thesis

Download: pdf ps

Abstract

This MS Thesis describes an application-independent checkpointing system for Charm++. It uses the pup framework to save in-flight messages, all running objects, and all Charm++/Converse state. The restart can occur on a different number of processors, or a different machine architecture.

Research Areas

Text Ref

Sameer Paranjpye "A Checkpoint and Restart Mechanism for Parallel Programming 
Systems", University of Illinois at Urbana-Champaign, 2000.

BibTex

@mastersthesis{SameerThesis,
	author = "Sameer Paranjpye",
	title = "A Checkpoint and Restart Mechanism for Parallel Programming Systems",
	year = "2000",
	school = {University of Illinois at Urbana-Champaign},
}