Architectural mechanisms for explicit communication in shared memory multiprocessors

Umakishore Ramachandran, Gautam Shah, Anand Sivasubramaniam, Aman Singla, Ivan Yanasak

Research output: Contribution to journalConference article

8 Citations (Scopus)

Abstract

The goal of this work is to explore architectural mechanisms for supporting explicit communication in cache-coherent shared memory multiprocessors. The motivation stems from the observation that applications display wide diversity in terms of sharing characteristics and hence impose different communication requirements on the system. Explicit communication mechanisms would allow tailoring the coherence management under software control to match these differing needs and strive to provide a close approximation to a zero overhead machine from the application perspective. Toward achieving these goals, we first analyze the characteristics of sharing observed in certain specific applications. We then use these characteristics to synthesize explicit communication primitives. The proposed primitives allow selectively updating a set of processors, or requesting a stream of data ahead of its intended use. These primitives are essentially generalizations of prefetch and poststore, with the ability to specify the sharer set for poststore either statically or dynamically. The proposed primitives are to be used in conjunction with an underlying invalidation based protocol. Used in this manner, the resulting memory system can dynamically adapt itself to performing either invalidations or updates to match the communication needs. Through application driven performance study we show the utility of these mechanisms in being able to reduce and tolerate communication latencies.

Original languageEnglish (US)
Pages (from-to)1737-1744
Number of pages8
JournalProceedings of the ACM/IEEE Supercomputing Conference
Volume2
StatePublished - Dec 1 1995
EventProceedings of the 1995 ACM/IEEE Supercomputing Conference. Part 2 (of 2) - San Diego, CA, USA
Duration: Dec 3 1995Dec 8 1995

Fingerprint

Data storage equipment
Communication

All Science Journal Classification (ASJC) codes

  • Electrical and Electronic Engineering

Cite this

Ramachandran, Umakishore ; Shah, Gautam ; Sivasubramaniam, Anand ; Singla, Aman ; Yanasak, Ivan. / Architectural mechanisms for explicit communication in shared memory multiprocessors. In: Proceedings of the ACM/IEEE Supercomputing Conference. 1995 ; Vol. 2. pp. 1737-1744.
@article{642d319374b34d3fb5bd25c01f94257e,
title = "Architectural mechanisms for explicit communication in shared memory multiprocessors",
abstract = "The goal of this work is to explore architectural mechanisms for supporting explicit communication in cache-coherent shared memory multiprocessors. The motivation stems from the observation that applications display wide diversity in terms of sharing characteristics and hence impose different communication requirements on the system. Explicit communication mechanisms would allow tailoring the coherence management under software control to match these differing needs and strive to provide a close approximation to a zero overhead machine from the application perspective. Toward achieving these goals, we first analyze the characteristics of sharing observed in certain specific applications. We then use these characteristics to synthesize explicit communication primitives. The proposed primitives allow selectively updating a set of processors, or requesting a stream of data ahead of its intended use. These primitives are essentially generalizations of prefetch and poststore, with the ability to specify the sharer set for poststore either statically or dynamically. The proposed primitives are to be used in conjunction with an underlying invalidation based protocol. Used in this manner, the resulting memory system can dynamically adapt itself to performing either invalidations or updates to match the communication needs. Through application driven performance study we show the utility of these mechanisms in being able to reduce and tolerate communication latencies.",
author = "Umakishore Ramachandran and Gautam Shah and Anand Sivasubramaniam and Aman Singla and Ivan Yanasak",
year = "1995",
month = "12",
day = "1",
language = "English (US)",
volume = "2",
pages = "1737--1744",
journal = "Proceedings of the ACM/IEEE Supercomputing Conference",
issn = "1063-9535",

}

Architectural mechanisms for explicit communication in shared memory multiprocessors. / Ramachandran, Umakishore; Shah, Gautam; Sivasubramaniam, Anand; Singla, Aman; Yanasak, Ivan.

In: Proceedings of the ACM/IEEE Supercomputing Conference, Vol. 2, 01.12.1995, p. 1737-1744.

Research output: Contribution to journalConference article

TY - JOUR

T1 - Architectural mechanisms for explicit communication in shared memory multiprocessors

AU - Ramachandran, Umakishore

AU - Shah, Gautam

AU - Sivasubramaniam, Anand

AU - Singla, Aman

AU - Yanasak, Ivan

PY - 1995/12/1

Y1 - 1995/12/1

N2 - The goal of this work is to explore architectural mechanisms for supporting explicit communication in cache-coherent shared memory multiprocessors. The motivation stems from the observation that applications display wide diversity in terms of sharing characteristics and hence impose different communication requirements on the system. Explicit communication mechanisms would allow tailoring the coherence management under software control to match these differing needs and strive to provide a close approximation to a zero overhead machine from the application perspective. Toward achieving these goals, we first analyze the characteristics of sharing observed in certain specific applications. We then use these characteristics to synthesize explicit communication primitives. The proposed primitives allow selectively updating a set of processors, or requesting a stream of data ahead of its intended use. These primitives are essentially generalizations of prefetch and poststore, with the ability to specify the sharer set for poststore either statically or dynamically. The proposed primitives are to be used in conjunction with an underlying invalidation based protocol. Used in this manner, the resulting memory system can dynamically adapt itself to performing either invalidations or updates to match the communication needs. Through application driven performance study we show the utility of these mechanisms in being able to reduce and tolerate communication latencies.

AB - The goal of this work is to explore architectural mechanisms for supporting explicit communication in cache-coherent shared memory multiprocessors. The motivation stems from the observation that applications display wide diversity in terms of sharing characteristics and hence impose different communication requirements on the system. Explicit communication mechanisms would allow tailoring the coherence management under software control to match these differing needs and strive to provide a close approximation to a zero overhead machine from the application perspective. Toward achieving these goals, we first analyze the characteristics of sharing observed in certain specific applications. We then use these characteristics to synthesize explicit communication primitives. The proposed primitives allow selectively updating a set of processors, or requesting a stream of data ahead of its intended use. These primitives are essentially generalizations of prefetch and poststore, with the ability to specify the sharer set for poststore either statically or dynamically. The proposed primitives are to be used in conjunction with an underlying invalidation based protocol. Used in this manner, the resulting memory system can dynamically adapt itself to performing either invalidations or updates to match the communication needs. Through application driven performance study we show the utility of these mechanisms in being able to reduce and tolerate communication latencies.

UR - http://www.scopus.com/inward/record.url?scp=0029428663&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0029428663&partnerID=8YFLogxK

M3 - Conference article

AN - SCOPUS:0029428663

VL - 2

SP - 1737

EP - 1744

JO - Proceedings of the ACM/IEEE Supercomputing Conference

JF - Proceedings of the ACM/IEEE Supercomputing Conference

SN - 1063-9535

ER -