Zobrazit v EDS

A Data Forwarding Scheme for Task-based Programming Models

Uloženo v:

Podrobná bibliografie
Název:	A Data Forwarding Scheme for Task-based Programming Models
Autoři:	Manivannan, Madhavan, Negi, Anurag, 1980, Stenström, Per, 1957
Zdroj:	Proceedings of the Fifth Swedish Workshop on Multicore Computing.
Popis:	Task-based programming models are increasingly being adopted due to their ability to express parallelism. Communication is an inherent aspect of this model and is expected toplay an important part in application scalability on multi-core architectures. In this paper we focus specifically oncommunication arising due to producer-consumer sharing intask-based applications. Existing approaches that optimizefor producer-consumer sharing, by predicting the identity ofthe consumers and forwarding data in advance, rely on producers and consumers to exhibit stable communication to be effective. We show that task-based parallel applications do not exhibit stable communication as the mapping of tasksto cores changes based on runtime conditions thereby lim-iting applicability of such optimizations. As an alternativesolution, we propose a forwarding scheme in the context ofchip multiprocessors that helps in reducing communicationoverheads by enabling cache-to-cache transfers of updatedlines to the consumer after the consumer identifies itself.Our evaluation indicates that the proposed scheme is ableto reduce communication overheads due to such producer-consumer sharing.
Přístupová URL adresa:	https://research.chalmers.se/publication/172050
Databáze:	SwePub

Popis
Abstrakt:	Task-based programming models are increasingly being adopted due to their ability to express parallelism. Communication is an inherent aspect of this model and is expected toplay an important part in application scalability on multi-core architectures. In this paper we focus specifically oncommunication arising due to producer-consumer sharing intask-based applications. Existing approaches that optimizefor producer-consumer sharing, by predicting the identity ofthe consumers and forwarding data in advance, rely on producers and consumers to exhibit stable communication to be effective. We show that task-based parallel applications do not exhibit stable communication as the mapping of tasksto cores changes based on runtime conditions thereby lim-iting applicability of such optimizations. As an alternativesolution, we propose a forwarding scheme in the context ofchip multiprocessors that helps in reducing communicationoverheads by enabling cache-to-cache transfers of updatedlines to the consumer after the consumer identifies itself.Our evaluation indicates that the proposed scheme is ableto reduce communication overheads due to such producer-consumer sharing.