Publication: Recomputations in Reverse Mode AD
Introduction
Applications
Tools
Research Groups
Workshops
Publications
   List Publications
   Advanced Search
   Info
   Add Publications
My Account
About

Recomputations in Reverse Mode AD

- incollection -
 

Area
General

Author(s)
Ralf Giering , Thomas Kaminski

Published in
Automatic Differentiation: From Simulation to Optimization

Editor(s)
George Corliss, Christèle Faure, Andreas Griewank, Laurent Hascoët, Uwe Naumann

Year
2002

Publisher
Springer

Abstract
The main challenge of the reverse (or adjoint) mode of automatic differentiation (ad) is providing the accurate values of required variables to the derivative code. We discuss different strategies to tackle this challenge. The ability to generate efficient adjoint code is crucial for handling large scale applications. For challenging applications, efficient adjoint code must provide at least a fraction of the values of required variables through recomputations, but it is essential to avoid unnecessary recomputations. This is achieved by the Efficient Recomputation Algorithm implemented in the Tangent linear and Adjoint Model Compiler and in Transformation of Algorithms in Fortran, which are source-to-source translation ad tools for Fortran programs. We describe the algorithm and discuss possible improvements.

Cross-References
Corliss2002ADo

AD Tools
TAF, TAMC

AD Theory and Techniques
Recomputation

BibTeX
@INCOLLECTION{
         Giering2002RiR,
       author = "Ralf Giering and Thomas Kaminski",
       editor = "George Corliss and Christ{\`e}le Faure and Andreas Griewank and Laurent
         Hasco{\"e}t and Uwe Naumann",
       title = "Recomputations in Reverse Mode {AD}",
       booktitle = "Automatic Differentiation: From Simulation to Optimization",
       series = "Computer and Information Science",
       pages = "283--291",
       publisher = "Springer",
       address = "New York",
       key = "Giering2002RiR",
       abstract = "The main challenge of the reverse (or adjoint) mode of automatic differentiation
         (AD) is providing the accurate values of required variables to the derivative code. We discuss
         different strategies to tackle this challenge. The ability to generate efficient adjoint code is
         crucial for handling large scale applications. For challenging applications, efficient adjoint code
         must provide at least a fraction of the values of required variables through recomputations, but it
         is essential to avoid unnecessary recomputations. This is achieved by the Efficient Recomputation
         Algorithm implemented in the Tangent linear and Adjoint Model Compiler and in Transformation of
         Algorithms in Fortran, which are source-to-source translation AD tools for Fortran programs. We
         describe the algorithm and discuss possible improvements.",
       referred = "[Faure2002ASf], [Griewank2002VJS], [Klein2002DMf].",
       year = "2002",
       ad_tools = "TAF, TAMC",
       ad_area = "General",
       ad_theotech = "Recomputation",
       chapter = "33",
       pdf = "http://www.FastOpt.com/papers/ad2000.pdf",
       url = "http://www.springer.de/cgi-bin/search_book.pl?isbn=0-387-95305-1",
       crossref = "Corliss2002ADo"
}


back
  

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