Title
A lightweight task graph scheduler for distributed high-performance scientific computing
Abstract
The continually growing demand for increased simulation complexity introduces the need for scientific software frameworks to parallelize simulation tasks. We present our approach for a task graph scheduler based on modern programming techniques. The scheduler utilizes the Message Passing Interface to distribute the tasks among distributed computing nodes. We show that our approach does not only offer a concise user-level code but also provides a high degree of scalability.
Year
DOI
Venue
2012
10.1007/978-3-642-36803-5_47
PARA
Keywords
Field
DocType
high degree,increased simulation complexity,task graph scheduler,simulation task,message passing interface,modern programming technique,lightweight task graph scheduler,high-performance scientific computing,scientific software framework,concise user-level code
Graph,Scientific software,Computer science,Parallel computing,Theoretical computer science,Message Passing Interface,Distributed computing,Scalability
Conference
Volume
ISSN
Citations 
7782
0302-9743
0
PageRank 
References 
Authors
0.34
5
3
Name
Order
Citations
PageRank
Josef Weinbub1179.55
Karl Rupp25412.32
Siegfried Selberherr310539.95