Skip to Main content Skip to Navigation
Reports

Expressiveness and Data-Flow Compilation of OpenMP Streaming Programs

Antoniu Pop 1 Albert Cohen 1
1 Parkas - Parallélisme de Kahn Synchrone
DI-ENS - Département d'informatique de l'École normale supérieure, Inria Paris-Rocquencourt, CNRS - Centre National de la Recherche Scientifique : UMR 8548
Abstract : We present a data-flow extension of OpenMP to express highly dynamic control and data flow over nested, dependent tasks. The language supports dynamic creation, modular composition, variable and unbounded sets of producers/consumers, separate compilation, and first-class streams. These features, enabled by our original compilation flow, allow translating high-level parallel programming patterns, like dependences arising from StarSs' array regions, or universal low-level primitives like futures. In particular, these dynamic features can be embedded efficiently and naturally into an unmanaged imperative language, avoiding the complexity and overhead of a concurrent garbage collector. We demonstrate the performance advantages of a data-flow execution model compared to more restricted task and barrier models. We also demonstrate the efficiency of our compilation and runtime algorithms for the support of complex dependence patterns arising from StarSs benchmarks.
Complete list of metadata

https://hal.inria.fr/hal-00710409
Contributor : Antoniu Pop <>
Submitted on : Monday, July 2, 2012 - 3:23:55 PM
Last modification on : Thursday, July 1, 2021 - 5:58:07 PM
Long-term archiving on: : Wednesday, October 3, 2012 - 3:15:15 AM

File

RR-8001.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-00710409, version 2

Collections

Citation

Antoniu Pop, Albert Cohen. Expressiveness and Data-Flow Compilation of OpenMP Streaming Programs. [Research Report] RR-8001, INRIA. 2012, pp.28. ⟨hal-00710409v2⟩

Share

Metrics

Record views

540

Files downloads

399