Trace-based simulation

Summary

In computer science, trace-based simulation refers to system simulation performed by looking at traces of program execution or system component access with the purpose of performance prediction.[1]

Trace-based simulation may be used in a variety of applications, from the analysis of solid state disks to the message passing performance on very large computer clusters.[1][2]

Traced-based simulators usually have two components: one that executes actions and stores the results (i.e. traces) and another which reads the log files of traces and interpolates them to new (and often more complex) scenarios.[2]

For instance, in the case of large computer cluster design, the execution takes place on a small number of nodes, and traces are left in log files. The simulator reads those log files and simulates performance on a much larger number of nodes, thus providing a view of the performance of very large applications, based on the execution traces on a much smaller number of nodes.[2][3]

See also edit

References edit

  1. ^ a b Software Technologies for Embedded and Ubiquitous Systems edited by Sunggu Lee and Priya Narasimhan 2009 ISBN 3642102646 page 28
  2. ^ a b c Languages and Compilers for Parallel Computing edited by Keith Cooper, John Mellor-Crummey and Vivek Sarkar 2011 ISBN 3642195946 pages 202-203
  3. ^ Petascale Computing: Algorithms and Applications by David A. Bader 2007 ISBN 1584889098 pages 435-435