×

zbMATH — the first resource for mathematics

The data-flow equations of checkpointing in reverse automatic differentiation. (English) Zbl 1157.65334
Alexandrov, Vassil N. (ed.) et al., Computational science – ICCS 2006. 6th international conference, Reading, UK, May 28–31, 2006. Proceedings, Part IV. Berlin: Springer (ISBN 3-540-34385-7/pbk). Lecture Notes in Computer Science 3994, 566-573 (2006).
Summary: Checkpointing is a technique to reduce the memory consumption of adjoint programs produced by reverse Automatic Differentiation. However, checkpointing also uses a non-negligible memory space for the so-called “snapshots”. We analyze the data-flow of checkpointing, yielding a precise characterization of all possible memory-optimal options for snapshots. This characterization is formally derived from the structure of checkpoints and from classical data-flow equations. In particular, we select two very different options and study their behavior on a number of real codes. Although no option is uniformly better, the so-called “lazy-snapshot” option appears preferable in general.
For the entire collection see [Zbl 1107.68001].

MSC:
65D25 Numerical differentiation
Software:
Maple; TAPENADE
PDF BibTeX XML Cite
Full Text: DOI