Publication: Hierarchical Automatic Differentiation by Vertex Elimination and Source Transformation
Introduction
Applications
Tools
Research Groups
Workshops
Publications
   List Publications
   Advanced Search
   Info
   Add Publications
My Account
About
Impress

Hierarchical Automatic Differentiation by Vertex Elimination and Source Transformation

- Part of a collection -
 

Area
Computational Fluid Dynamics

Author(s)
Mohamed Tadjouddine , Shaun A. Forth , John D. Pryce

Published in
Computational Science and Its Applications -- ICCSA 2003, Proceedings of the International Conference on Computational Science and its Applications, Montreal, Canada, May 18--21, 2003. Part II

Editor(s)
V. Kumar, M. L. Gavrilova, C. J. K. Tan, P. L'Ecuyer

Year
2003

Publisher
Springer

Abstract
We present a hierarchical scheme to extend the applicability of automatic differentiation (ad) by vertex elimination from the basic block level to code with branches and subroutine calls. We introduce the ELIAD tool that implements our scheme. Results from computational fluid dynamics (CFD) flux linearisations show runtime speedup by a typical factor of two over both finite-differencing and traditional forward and reverse modes of ad.

Cross-References
Kumar2003CSa

AD Tools
EliAD

AD Theory and Techniques
Hierarchical Approach

BibTeX
@INPROCEEDINGS{
         Tadjouddine2003HAD,
       author = "Mohamed Tadjouddine and Shaun A. Forth and John D. Pryce",
       title = "Hierarchical Automatic Differentiation by Vertex Elimination and Source
         Transformation",
       booktitle = "Computational Science and Its Applications -- ICCSA~2003, Proceedings of the
         International Conference on Computational Science and its Applications, Montreal, Canada,
         May~18--21, 2003. Part~II",
       editor = "V. Kumar and M. L. Gavrilova and C. J. K. Tan and P. {L'Ecuyer}",
       abstract = "We present a hierarchical scheme to extend the applicability of automatic
         differentiation (AD) by vertex elimination from the basic block level to code with branches and
         subroutine calls. We introduce the ELIAD tool that implements our scheme. Results from computational
         fluid dynamics (CFD) flux linearisations show runtime speedup by a typical factor of two over both
         finite-differencing and traditional forward and reverse modes of AD.",
       volume = "2668",
       series = "Lecture Notes in Computer Science",
       publisher = "Springer",
       pages = "115--124",
       address = "Berlin",
       reference = "http://www.autodiff.org/?module=Workshops&submenu=iccsa03",
       ad_area = "Computational Fluid Dynamics",
       ad_tools = "EliAD",
       ad_theotech = "Hierarchical Approach",
       year = "2003",
       crossref = "Kumar2003CSa"
}


back
  

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