Increasing the degree of parallelism using speculative execution in task-based runtime systems
Task-based programming models have demonstrated their efficiency in the development of scientific applications on modern high-performance platforms. They allow delegation of the management of parallelization to the runtime system (RS), which is in charge of the data coherency, the scheduling, and th...
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
PeerJ Inc.
2019-03-01
|
| Series: | PeerJ Computer Science |
| Subjects: | |
| Online Access: | https://peerj.com/articles/cs-183.pdf |
| _version_ | 1818893674370564096 |
|---|---|
| author | Bérenger Bramas |
| author_facet | Bérenger Bramas |
| author_sort | Bérenger Bramas |
| collection | DOAJ |
| description | Task-based programming models have demonstrated their efficiency in the development of scientific applications on modern high-performance platforms. They allow delegation of the management of parallelization to the runtime system (RS), which is in charge of the data coherency, the scheduling, and the assignment of the work to the computational units. However, some applications have a limited degree of parallelism such that no matter how efficient the RS implementation, they may not scale on modern multicore CPUs. In this paper, we propose using speculation to unleash the parallelism when it is uncertain if some tasks will modify data, and we formalize a new methodology to enable speculative execution in a graph of tasks. This description is partially implemented in our new C++ RS called SPETABARU, which is capable of executing tasks in advance if some others are not certain to modify the data. We study the behavior of our approach to compute Monte Carlo and replica exchange Monte Carlo simulations. |
| first_indexed | 2024-12-19T18:16:21Z |
| format | Article |
| id | doaj.art-3b9ce9d2e64d416983bbe2a846d35ff7 |
| institution | Directory Open Access Journal |
| issn | 2376-5992 |
| language | English |
| last_indexed | 2024-12-19T18:16:21Z |
| publishDate | 2019-03-01 |
| publisher | PeerJ Inc. |
| record_format | Article |
| series | PeerJ Computer Science |
| spelling | doaj.art-3b9ce9d2e64d416983bbe2a846d35ff72022-12-21T20:11:05ZengPeerJ Inc.PeerJ Computer Science2376-59922019-03-015e18310.7717/peerj-cs.183Increasing the degree of parallelism using speculative execution in task-based runtime systemsBérenger Bramas0CAMUS Team, Inria Nancy—Grand Est, Illkirch-Graffenstaden, FranceTask-based programming models have demonstrated their efficiency in the development of scientific applications on modern high-performance platforms. They allow delegation of the management of parallelization to the runtime system (RS), which is in charge of the data coherency, the scheduling, and the assignment of the work to the computational units. However, some applications have a limited degree of parallelism such that no matter how efficient the RS implementation, they may not scale on modern multicore CPUs. In this paper, we propose using speculation to unleash the parallelism when it is uncertain if some tasks will modify data, and we formalize a new methodology to enable speculative execution in a graph of tasks. This description is partially implemented in our new C++ RS called SPETABARU, which is capable of executing tasks in advance if some others are not certain to modify the data. We study the behavior of our approach to compute Monte Carlo and replica exchange Monte Carlo simulations.https://peerj.com/articles/cs-183.pdfSTFMonte-CarloSpeculationTask-based |
| spellingShingle | Bérenger Bramas Increasing the degree of parallelism using speculative execution in task-based runtime systems PeerJ Computer Science STF Monte-Carlo Speculation Task-based |
| title | Increasing the degree of parallelism using speculative execution in task-based runtime systems |
| title_full | Increasing the degree of parallelism using speculative execution in task-based runtime systems |
| title_fullStr | Increasing the degree of parallelism using speculative execution in task-based runtime systems |
| title_full_unstemmed | Increasing the degree of parallelism using speculative execution in task-based runtime systems |
| title_short | Increasing the degree of parallelism using speculative execution in task-based runtime systems |
| title_sort | increasing the degree of parallelism using speculative execution in task based runtime systems |
| topic | STF Monte-Carlo Speculation Task-based |
| url | https://peerj.com/articles/cs-183.pdf |
| work_keys_str_mv | AT berengerbramas increasingthedegreeofparallelismusingspeculativeexecutionintaskbasedruntimesystems |