Publication: Elimination Techniques for Cheap Jacobians
Introduction
Applications
Tools
Research Groups
Workshops
Publications
   List Publications
   Advanced Search
   Info
   Add Publications
My Account
About

Elimination Techniques for Cheap Jacobians

- incollection -
 

Author(s)
Uwe Naumann

Published in
Automatic Differentiation of Algorithms: From Simulation to Optimization

Editor(s)
G. Corliss, C. Faure, A. Griewank, L. Hascoët, U. Naumann

Year
2002

Publisher
Springer

Abstract
The generation of optimized code for evaluating the Jacobian matrix of a vector function is known to result in a remarkable speedup of three and more compared to standard methods of Automatic Differentiation in most cases. So far, this optimization has been built on the elimination of vertices in the computational graph. We show that vertex elimination in general does not lead to optimal Jacobian code. We introduce two new elimination methods and demonstrate their superiority over the vertex elimination approach.

Cross-References
Corliss2002ADo

AD Theory and Techniques
General

BibTeX
@INCOLLECTION{
         Naumann2002ETf,
       author = "Uwe Naumann",
       title = "Elimination Techniques for Cheap {J}acobians",
       pages = "247--253",
       chapter = "29",
       crossref = "Corliss2002ADo",
       booktitle = "Automatic Differentiation of Algorithms: From Simulation to Optimization",
       year = "2002",
       editor = "G. Corliss and C. Faure and A. Griewank and L. Hasco{\"e}t and U. Naumann",
       series = "Computer and Information Science",
       publisher = "Springer",
       address = "New York, NY",
       abstract = "The generation of optimized code for evaluating the Jacobian matrix of a vector
         function is known to result in a remarkable speedup of three and more compared to standard methods
         of Automatic Differentiation in most cases. So far, this optimization has been built on the
         elimination of vertices in the computational graph. We show that vertex elimination in general does
         not lead to optimal Jacobian code. We introduce two new elimination methods and demonstrate their
         superiority over the vertex elimination approach.",
       referred = "[Faure2002MtT], [Klein2002DMf].",
       ad_theotech = "General"
}


back
  

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