Toward performance prediction for Multi-BSP programs in ML

Abstract : bsml and multi-ml are functional parallel programming languages "à la ml" based of the respectively the bsp and multi-bsp bridging models. multi-bsp extends bsp to take into account hierarchical archi-tectures. For both models, it is possible to predict the performances of algorithms thanks to embedded cost models. To do so, we propose formal operational semantics with cost annotations for the two aforementioned languages. This work has been done in a incremental manner. First we recall the cost semantics of core-ml language. Then, we adapt it to bsml and then to multi-ml. It is then possible to evaluate the cost of a program following the annotated semantics. Finally, we compare the theoretical approach with the current implementation on a code example.
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-01941250
Contributor : Victor Allombert <>
Submitted on : Monday, December 10, 2018 - 11:08:27 AM
Last modification on : Thursday, February 7, 2019 - 3:38:29 PM
Document(s) archivé(s) le : Monday, March 11, 2019 - 1:35:06 PM

File

main.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01941250, version 2

Citation

Victor Allombert, Frédéric Gava, Julien Tesson. Toward performance prediction for Multi-BSP programs in ML. 18th International Conference on Algorithms and Architectures for Parallel Processing (ICA3PP), Nov 2018, Guangzhou, China. ⟨hal-01941250v2⟩

Share

Metrics

Record views

15

Files downloads

13