Publication: On Combining Computational Differentiation and Toolkits for Parallel Scientific Computing
Introduction
Applications
Tools
Research Groups
Workshops
Publications
   List Publications
   Advanced Search
   Info
   Add Publications
My Account
About

On Combining Computational Differentiation and Toolkits for Parallel Scientific Computing

- Part of a collection -
 

Author(s)
C. H. Bischof , H. M. Bücker , P. D. Hovland

Published in
Euro-Par 2000 -- Parallel Processing, Proceedings of the 6th International Euro-Par Conference, Munich, Germany, August/September 2000

Editor(s)
A. Bode, T. Ludwig, W. Karl, R. Wismüller

Year
2000

Publisher
Springer

Abstract
Automatic differentiation is a powerful technique for evaluating derivatives of functions given in the form of a high-level programming language such as Fortran, C, or C++. The program is treated as a potentially very long sequence of elementary statements to which the chain rule of differential calculus is applied over and over again. Combining automatic differentiation and the organizational structure of toolkits for parallel scientific computing provides a mechanism for evaluating derivatives by exploiting mathematical insight on a higher level. In these toolkits, algorithmic structures such as BLAS-like operations, linear and nonlinear solvers, or integrators for ordinary differential equations can be identified by their standardized interfaces and recognized as high-level mathematical objects rather than as a sequence of elementary statements. In this note, the differentiation of a linear solver with respect to some parameter vector is taken as an example. Mathematical insight is used to reformulate this problem into the solution of multiple linear systems that share the same coefficient matrix but differ in their right-hand sides. The experiments reported here use ADIC, a tool for the automatic differentiation of C programs, and PETSc, an object-oriented toolkit for the parallel solution of scientific problems modeled by partial differential equations.

AD Theory and Techniques
Toolkits

BibTeX
@INPROCEEDINGS{
         Bischof2000OCC,
       author = "C.~H.~Bischof and H.~M.~B{\"u}cker and P.~D.~Hovland",
       title = "On Combining Computational Differentiation and Toolkits for Parallel Scientific
         Computing",
       booktitle = "Euro-Par~2000 -- Parallel Processing, Proceedings of the 6th International
         Euro-Par Conference, Munich, Germany, August/September~2000",
       editor = "A.~Bode and T.~Ludwig and W.~Karl and R.~Wism{\"u}ller",
       volume = "1900",
       series = "Lecture Notes in Computer Science",
       pages = "86--94",
       address = "Berlin",
       publisher = "Springer",
       abstract = "Automatic differentiation is a powerful technique for evaluating derivatives of
         functions given in the form of a high-level programming language such as Fortran, C, or C++. The
         program is treated as a potentially very long sequence of elementary statements to which the chain
         rule of differential calculus is applied over and over again. Combining automatic differentiation
         and the organizational structure of toolkits for parallel scientific computing provides a mechanism
         for evaluating derivatives by exploiting mathematical insight on a higher level. In these toolkits,
         algorithmic structures such as BLAS-like operations, linear and nonlinear solvers, or integrators
         for ordinary differential equations can be identified by their standardized interfaces and
         recognized as high-level mathematical objects rather than as a sequence of elementary statements. In
         this note, the differentiation of a linear solver with respect to some parameter vector is taken as
         an example. Mathematical insight is used to reformulate this problem into the solution of multiple
         linear systems that share the same coefficient matrix but differ in their right-hand sides. The
         experiments reported here use ADIC, a tool for the automatic differentiation of C~programs, and
         PETSc, an object-oriented toolkit for the parallel solution of scientific problems modeled by
         partial differential equations.",
       ad_theotech = "Toolkits",
       year = "2000",
       doi = "10.1007/3-540-44520-X_12"
}


back
  

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