Abstract
The reverse or adjoint mode of automatic differentiation is a software engineering technique that permits efficient computation of gradients. However, this technique requires a lot of temporary memory. In this chapter, we present a refinement that reduces memory consumption in the case of parallel loops, and we give a proof of its correctness based on properties of the data-dependence graph of adjoint programs and parallel loops. This technique is particularly suitable for assembly loops that dominate in mesh-based computations. Application is done on the kernel of a realistic Navier-Stokes solver.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2002 Springer Science+Business Media New York
About this chapter
Cite this chapter
Hascoët, L., Fidanova, S., Held, C. (2002). Adjoining Independent Computations. In: Corliss, G., Faure, C., Griewank, A., Hascoët, L., Naumann, U. (eds) Automatic Differentiation of Algorithms. Springer, New York, NY. https://doi.org/10.1007/978-1-4613-0075-5_35
Download citation
DOI: https://doi.org/10.1007/978-1-4613-0075-5_35
Published:
Publisher Name: Springer, New York, NY
Print ISBN: 978-1-4612-6543-6
Online ISBN: 978-1-4613-0075-5
eBook Packages: Springer Book Archive