Publication: Parallel Derivative Computation Using ADOL-C
Introduction
Applications
Tools
Research Groups
Workshops
Publications
   List Publications
   Advanced Search
   Info
   Add Publications
My Account
About

Parallel Derivative Computation Using ADOL-C

- Part of a collection -
 

Area
Physics

Author(s)
A. Kowarz , A. Walther

Published in
Proceedings of PASA 2008

Editor(s)
W. Nagel, R. Hoffmann, A. Koch

Year
2008

Abstract
Derivative computation using Automatic Differentiation (ad) is often considered to operate purely serial. Performing the differentiation task in parallel may require the applied ad-tool to extract parallelization information from the user function, transform it, and apply this new strategy in the differentiation process. Furthermore, when using the reverse mode of ad, it must be ensured that no data races are introduced due to the reversed data access scheme. Considering an operator overloading based ad-tool, an additional challenge is to be met: Parallelization statements are typically not recognized. In this paper, we present and discuss the parallelization approach that we have integrated into ADOL-C, an operator overloading based ad-tool for the differentiation of C/C++ programs. The advantages of the approach are clarified by means of the parallel differentiation of a function that handles the time evolution of a 1D-quantum plasma.

AD Tools
ADOL-C

AD Theory and Techniques
Parallelism

BibTeX
@INPROCEEDINGS{
         Kowarz2008PDC,
       title = "Parallel Derivative Computation Using {ADOL-C}",
       author = "A. Kowarz and A. Walther",
       editor = "W. Nagel and R. Hoffmann and A. Koch",
       year = "2008",
       booktitle = "Proceedings of PASA 2008",
       series = "Lecture Notes in Informatics, Vol. 124",
       pages = "83--92",
       abstract = "Derivative computation using Automatic Differentiation (AD) is often considered to
         operate purely serial. Performing the differentiation task in parallel may require the applied
         AD-tool to extract parallelization information from the user function, transform it, and apply this
         new strategy in the differentiation process. Furthermore, when using the reverse mode of AD, it must
         be ensured that no data races are introduced due to the reversed data access scheme. Considering an
         operator overloading based AD-tool, an additional challenge is to be met: Parallelization statements
         are typically not recognized. In this paper, we present and discuss the parallelization approach
         that we have integrated into ADOL-C, an operator overloading based AD-tool for the differentiation
         of C/C++ programs. The advantages of the approach are clarified by means of the parallel
         differentiation of a function that handles the time evolution of a 1D-quantum plasma.",
       ad_area = "Physics",
       ad_tools = "ADOL-C",
       ad_theotech = "Parallelism",
       url = "http://subs.emis.de/LNI/Proceedings/Proceedings124/article2036.html"
}


back
  

Contact:
autodiff.org
Username:
Password:
(lost password)