Subutai: Speeding Up Legacy Parallel Applications Through Data Synchronization

Rodrigo Cataldo (1, 2) , Ramon Fernandes (2) , Kevin Martin (1) , Jarbas Silveira (3) , Gustavo Sanchez (2) , Johanna Sepulveda (4) , Cesar Marcon (2) , Jean-Phillipe Diguet (1)

1 Lab-STICC_UBS_CACS_MOCS
2 PUC-RS - Pontifícia Universidade Católica do Rio Grande do Sul [Brasil] = Pontifical Catholic University of Rio Grande do Sul [Brazil] = Université catholique pontificale de Rio Grande do Sul [Brésil]
3 UFC - Universidade Federal do Ceará = Federal University of Ceará
4 TUM - Technische Universität Munchen - Technical University Munich - Université Technique de Munich

The decrease of the performance gain dictated by Moore's Law boosted the development of manycore architectures to replace single-core architectures. These new architectures must employ parallel applications and distribute its workload over a multitude of cores to reach the desired performance. Parallel applications are harder to develop than sequential ones since the developer must guarantee data integrity using synchronization primitives. While multiple novel solutions have been proposed to speed up parallel applications through handling one type of data synchronization primitive, exceptionally few works support multiple types of synchronization primitives and legacy code. This article proposes Subutai, a hardware/software co-design solution for accelerating multiple synchronization primitives without modifying the application source code. By providing a new user library, while retaining an existing synchronization API, legacy and novel applications can benefit from our solution. Our experimental evaluation, which provides a POSIX Threads implementation, demonstrates Subutai speeds up to 2.71× and 4.61× the execution of single- and multiple-application executions, respectively.

Mots clés: Legacy Parallel Applications; PThreads; Network-on-Chip; Distributed Scheduler

Domaines: Architectures Matérielles [cs.AR]

Rodrigo Cataldo, Ramon Fernandes, Kevin Martin, Jarbas Silveira, Gustavo Sanchez, et al.. Subutai: Speeding Up Legacy Parallel Applications Through Data Synchronization. IEEE Transactions on Parallel and Distributed Systems, 2021, 32 (5), pp.1102-1116. ⟨10.1109/TPDS.2020.3040066⟩. ⟨hal-03082831⟩