Reducing Global Schedulers' Complexity Through Runtime System Decoupling

Alexandre Santana 1 Vinicius Freitas 1 Marcio Castro 1 Laércio Lima Pilla 2 Jean-François Méhaut 2
1 UFSC - Universidade Federal de Santa Catarina [Florianópolis]
2 CORSE - Compiler Optimization and Run-time Systems
Inria Grenoble - Rhône-Alpes, LIG - Laboratoire d'Informatique de Grenoble
Abstract : Global schedulers are components used in parallel solutions, specially in dynamic applications, to optimize resource usage. Nonetheless, their development is a cumbersome process due to necessary adaptations to cope with the programming interfaces and abstractions of runtime systems. This paper proposes a model to dissociate sched-ulers from runtime systems to lower software complexity. Our model is based on the scheduler breakdown into modular and reusable concepts that better express the scheduler requirements. Through the use of meta-programming and design patterns, we were able to achieve fully reusable workload-aware scheduling strategies with up to 63% fewer lines of code with negligible run time overhead.
Document type :
Conference papers
Complete list of metadatas

Cited literature [17 references]  Display  Hide  Download
https://hal.inria.fr/hal-01873526
Contributor : Laércio Lima Pilla <>
Submitted on : Thursday, September 13, 2018 - 1:15:26 PM
Last modification on : Monday, July 8, 2019 - 3:11:31 PM
Long-term archiving on : Friday, December 14, 2018 - 2:34:58 PM

File

mogs.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01873526, version 1

Collections

LIG-SRCPR-CORSE | LIG_SRCPR | GRID5000 | INRIA | LIG | UGA | POLYTECH-GRENOBLE

Citation

Alexandre Santana, Vinicius Freitas, Marcio Castro, Laércio Lima Pilla, Jean-François Méhaut. Reducing Global Schedulers' Complexity Through Runtime System Decoupling. WSCAD 2018 - XIX Simpósio de Sistemas Computacionais de Alto Desempenho, Oct 2018, São Paulo, Brazil. pp.1-12. ⟨hal-01873526⟩

Export

BibTeX TEI DC DCterms EndNote

Share

Metrics

Record views

375

Files downloads

136

Contact


archive-ouverte@inria.fr