Publication: Bringing Together Automatic Differentiation and OpenMP
Introduction
Applications
Tools
Research Groups
Workshops
Publications
   List Publications
   Advanced Search
   Info
   Add Publications
My Account
About

Bringing Together Automatic Differentiation and OpenMP

- Part of a collection -
 

Author(s)
H. Martin Bücker , Bruno Lang , an Mey , Dieter , Christian H. Bischof

Published in
Proceedings of the 15th ACM International Conference on Supercomputing, Sorrento, Italy, June 17--21, 2001

Year
2001

Publisher
ACM Press

Abstract
Derivatives of almost arbitrary functions can be evaluated efficiently by automatic differentiation whenever the functions are given in the form of computer programs in a high-level programming language such as Fortran, C, or C++. Furthermore, in contrast to numerical differentiation where derivatives are approximated, automatic differentiation generates derivatives that are accurate up to machine precision. The so-called forward mode of automatic differentiation computes derivatives by carrying forward a gradient associated with each intermediate variable simultaneously with the evaluation of the function itself. It is shown how software tools implementing the technology of automatic differentiation can benefit from simple concepts of shared memory programming to parallelize the gradient operations. The feasibility of our approach is demonstrated by numerical experiments. They were performed with a code that was generated automatically by the ADIFOR system and augmented with OpenMP directives.

AD Tools
ADIFOR

AD Theory and Techniques
Parallelism

BibTeX
@INPROCEEDINGS{
         Bucker2001BTA,
       title = "Bringing Together Automatic Differentiation and {OpenMP}",
       booktitle = "Proceedings of the 15th ACM International Conference on Supercomputing, Sorrento,
         Italy, June~17--21, 2001",
       publisher = "ACM Press",
       pages = "246--251",
       address = "New York",
       mynote = "Also available as SC preprint RWTH--CS--SC--01--04",
       mywwwtype = "proceedings",
       abstract = "Derivatives of almost arbitrary functions can be evaluated efficiently by automatic
         differentiation whenever the functions are given in the form of computer programs in a high-level
         programming language such as Fortran, C, or C++. Furthermore, in contrast to numerical
         differentiation where derivatives are approximated, automatic differentiation generates derivatives
         that are accurate up to machine precision. The so-called forward mode of automatic differentiation
         computes derivatives by carrying forward a gradient associated with each intermediate variable
         simultaneously with the evaluation of the function itself. It is shown how software tools
         implementing the technology of automatic differentiation can benefit from simple concepts of shared
         memory programming to parallelize the gradient operations. The feasibility of our approach is
         demonstrated by numerical experiments. They were performed with a code that was generated
         automatically by the Adifor system and augmented with OpenMP directives.",
       ad_tools = "ADIFOR",
       ad_theotech = "Parallelism",
       year = "2001",
       url = "http://doi.acm.org/10.1145/377792.377842",
       doi = "10.1145/377792.377842",
       author = "H. Martin B{\"u}cker and Bruno Lang and an Mey, Dieter and Christian
         H.~Bischof"
}


back
  

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