Efficient Doacross execution on distributed shared-memory multiprocessors

Hong Men Su, Pen Chung Yew

Research output: Chapter in Book/Report/Conference proceedingConference contribution

7 Scopus citations

Abstract

The authors propose a Doacross execution scheme for the distributed shared memory (DSM) architecture. The scheme uses direct communication and static message passing. The communication is direct because a producer writes (or sends) data directly to the nearest shared memory (NeSM) of its consumer. The message passing is static in the sense that a message is bound to a buffer location determined at compile time. The authors present compiler algorithms that transform a Doacross loop into the proposed execution scheme and derive conditions that can reduce the extra storage needed for message buffers. The preliminary results show that the proposed scheme can outperform approaches that ignore the features of the DSM architecture, i.e., simply utilize the DSM architecture as a UMA (uniform memory access) architecture.

Original languageEnglish (US)
Title of host publicationProc Supercomput 91
PublisherPubl by IEEE
Pages842-853
Number of pages12
ISBN (Print)0818621583
StatePublished - Dec 1 1991
EventProceedings of Supercomputing '91 - Albuquerque, NM, USA
Duration: Nov 18 1991Nov 22 1991

Publication series

NameProc Supercomput 91

Other

OtherProceedings of Supercomputing '91
CityAlbuquerque, NM, USA
Period11/18/9111/22/91

Fingerprint Dive into the research topics of 'Efficient Doacross execution on distributed shared-memory multiprocessors'. Together they form a unique fingerprint.

Cite this