Profiling checkpointing schedules in adjoint ST-AD - Inria - Institut national de recherche en sciences et technologies du numérique
Communication Dans Un Congrès Année : 2024

Profiling checkpointing schedules in adjoint ST-AD

Résumé

Checkpointing is a cornerstone of data-flow reversal in adjoint algorithmic differentiation. Checkpointing is a storage/recomputation trade-off that can be applied at different levels, one of which being the call tree. We are looking for good placements of checkpoints onto the call tree of a given application, to reduce run time and memory footprint of its adjoint. There is no known optimal solution to this problem other than a combinatorial search on all placements. We propose a heuristics based on run-time profiling of the adjoint code. We describe implementation of this profiling tool in an existing source-transformation AD tool. We demonstrate the interest of this approach on test cases taken from the MITgcm ocean and atmospheric global circulation model. We discuss the limitations of our approach and propose directions to lift them.

Fichier principal
Vignette du fichier
halSource03Sep24.pdf (326.35 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04693450 , version 1 (10-09-2024)

Licence

Identifiants

  • HAL Id : hal-04693450 , version 1

Citer

Laurent Hascoët, Jean-Luc Bouchot, Shreyas Sunil Gaikwad, Sri Hari Krishna Narayanan, Jan Hückelheim. Profiling checkpointing schedules in adjoint ST-AD. AD 2024 - 8th International Conference on Algorithmic Differentiation, Sep 2024, CHICAGO, United States. ⟨hal-04693450⟩
0 Consultations
0 Téléchargements

Partager

More