Publication: A Differentiation-Enabled Fortran 95 Compiler
Introduction
Applications
Tools
Research Groups
Workshops
Publications
   List Publications
   Advanced Search
   Info
   Add Publications
My Account
About
Impress

A Differentiation-Enabled Fortran 95 Compiler

- Article in a journal -
 

Author(s)
Uwe Naumann , Jan Riehme

Published in
ACM Transactions on Mathematical Software

Year
2005

Abstract
The availability of first derivatives of vector functions is crucial for the robustness and effciency of a large number of numerical algorithms. An upcoming new version of the differentiation-enabled NAGWare Fortran 95 compiler is described that uses programming language extensions and a semantic code transformation known as automatic differentiation to provide Jacobians of numerical programs with machine accuracy. We describe a new user interface as well as the relevant algorithmic details. In particular, we focus on the source transformation approach that generates locally optimal gradient code for single assignments by vertex elimination in the linearized computational graph. Extensive tests show the superiority of this method over the current overloading-based approach. The robustness and convenience of the new compiler-feature is illustrated by various case studies.

AD Tools
NAGWare Fortran 95

BibTeX
@ARTICLE{
         Naumann2005ADE,
       author = "Uwe Naumann and Jan Riehme",
       title = "A Differentiation-Enabled {Fortran} 95 Compiler",
       journal = "{ACM} Transactions on Mathematical Software",
       volume = "31",
       number = "4",
       year = "2005",
       pages = "458--474",
       ad_tools = "NAGWare Fortran 95",
       url = "http://doi.acm.org/10.1145/1114268.1114270",
       abstract = "The availability of first derivatives of vector functions is crucial for the
         robustness and effciency of a large number of numerical algorithms. An upcoming new version of the
         differentiation-enabled NAGWare Fortran 95 compiler is described that uses programming language
         extensions and a semantic code transformation known as automatic differentiation to provide
         Jacobians of numerical programs with machine accuracy. We describe a new user interface as well as
         the relevant algorithmic details. In particular, we focus on the source transformation approach that
         generates locally optimal gradient code for single assignments by vertex elimination in the
         linearized computational graph. Extensive tests show the superiority of this method over the current
         overloading-based approach. The robustness and convenience of the new compiler-feature is
         illustrated by various case studies."
}


back
  

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