Personal tools
You are here: Home Publications Activity Analysis: Algorithms and Effectiveness
Document Actions

Alan Carle and Mike Fagan (2004)

Activity Analysis: Algorithms and Effectiveness

In: Proceedings of the International Conference on Automatic Differentiation, Gleatcher Center, Chicago, IL, USA, Argonne National Laboratory.

In order to improve the performance of derivative code, many modern automatic differentiation tools use a source code analysis technique called activity analysis. Activity analysis potentially reduces both the computation time and space needed for a sensitivity calculations. This paper describes the static activity analysis algorithm used by Adifor, the special 'run time' activity analysis method. This paper also compares the static activity analysis with the exact run time information for a few large codes that we have processed over the past few years.

by admin last modified 2007-12-10 21:05
« October 2009 »
Su Mo Tu We Th Fr Sa
123
45678910
11121314151617
18192021222324
25262728293031
 

Powered by Plone

LACSI Collaborators include:

Rice University LANL UH UNM UIUC UNC UTK