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

Parallelism in the Reverse Mode

- incollection -
 

Author(s)
Jochen Benary

Published in
Computational Differentiation: Techniques, Applications, and Tools

Editor(s)
Martin Berz, Christian Bischof, George Corliss, Andreas Griewank

Year
1996

Publisher
SIAM

Abstract
In the basic form of the reverse mode for calculating derivatives, the amount of memory needed to record the intermediate values can become excessively large for problems of practical interest. If sequential checkpointing schemes are used, the memory requirement can be dramatically reduced, but the run time may be significantly increased. Implementing suitable checkpointing schemes on multiprocessor systems can decrease the run time to its theoretical minimum. Among the many possible scheduling strategies, we develop one that minimizes resource requirements. We present different communication structures that depend on the memory architecture of the multiprocessor system and the available resources. We also estimate the limits of the complexity and the memory requirements of the problem function.

Cross-References
Berz1996CDT

AD Theory and Techniques
Parallelism

BibTeX
@INCOLLECTION{
         Benary1996Pit,
       author = "Jochen Benary",
       editor = "Martin Berz and Christian Bischof and George Corliss and Andreas Griewank",
       title = "Parallelism in the Reverse Mode",
       booktitle = "Computational Differentiation: Techniques, Applications, and Tools",
       pages = "137--147",
       publisher = "SIAM",
       address = "Philadelphia, PA",
       key = "Benary1996Pit",
       crossref = "Berz1996CDT",
       abstract = "In the basic form of the reverse mode for calculating derivatives, the amount of
         memory needed to record the intermediate values can become excessively large for problems of
         practical interest. If sequential checkpointing schemes are used, the memory requirement can be
         dramatically reduced, but the run time may be significantly increased. Implementing suitable
         checkpointing schemes on multiprocessor systems can decrease the run time to its theoretical
         minimum. Among the many possible scheduling strategies, we develop one that minimizes resource
         requirements. We present different communication structures that depend on the memory architecture
         of the multiprocessor system and the available resources. We also estimate the limits of the
         complexity and the memory requirements of the problem function.",
       keywords = "Parallelism, reverse mode, checkpointing, snapshots.",
       referred = "[Mancini2002APH].",
       ad_theotech = "Parallelism",
       year = "1996"
}


back
  

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