Early experience with message-passing on the SHRIMP multicomputer

Edward W. Felten, Richard D. Alpert, Angelos Bilas, Matthias A. Blumrich, Douglas W. Clark, Stefanos N. Damianakis, Cezary Dubnicki, Liviu Iftode, Kai Li

Research output: Contribution to journalConference article

21 Citations (Scopus)

Abstract

The SHRIMP multicomputer provides virtual memory-mapped communication (VMMC), which supports protected, user-level message passing, allows user programs to perform their own buffer management, and separates data transfers from control transfers so that a data transfer can be done without the intervention of the receiving node CPU. An important question is whether such a mechanism can indeed deliver all of the available hardware performance to applications which use conventional message-passing libraries. This paper reports our early experience with message-passing on a small, working SHRIMP multicomputer. We have implemented several user-level communication libraries on top of the VMMC mechanism, including the NX message-passing interface, Sun RPC, stream sockets, and specialized RPC. The first three are fully compatible with existing systems. Our experience shows that the VMMC mechanism supports these message-passing interfaces well. When zero-copy protocols are allowed by the semantics of the interface, VMMC can effectively deliver to applications almost all of the raw hardware's communication performance.

Original languageEnglish (US)
Pages (from-to)296-307
Number of pages12
JournalConference Proceedings - Annual International Symposium on Computer Architecture, ISCA
StatePublished - Jan 1 1996
EventProceedings of the 1996 23rd Annual International Symposium on Computer Architecture - Philadelphia, PA, USA
Duration: May 22 1996May 24 1996

Fingerprint

Message passing
Communication
Interfaces (computer)
Data storage equipment
Data transfer
Hardware
Sun
Program processors
Semantics
Network protocols

All Science Journal Classification (ASJC) codes

  • Hardware and Architecture

Cite this

Felten, Edward W. ; Alpert, Richard D. ; Bilas, Angelos ; Blumrich, Matthias A. ; Clark, Douglas W. ; Damianakis, Stefanos N. ; Dubnicki, Cezary ; Iftode, Liviu ; Li, Kai. / Early experience with message-passing on the SHRIMP multicomputer. In: Conference Proceedings - Annual International Symposium on Computer Architecture, ISCA. 1996 ; pp. 296-307.
@article{4a207d4ff07b412a81ac1d477490f4be,
title = "Early experience with message-passing on the SHRIMP multicomputer",
abstract = "The SHRIMP multicomputer provides virtual memory-mapped communication (VMMC), which supports protected, user-level message passing, allows user programs to perform their own buffer management, and separates data transfers from control transfers so that a data transfer can be done without the intervention of the receiving node CPU. An important question is whether such a mechanism can indeed deliver all of the available hardware performance to applications which use conventional message-passing libraries. This paper reports our early experience with message-passing on a small, working SHRIMP multicomputer. We have implemented several user-level communication libraries on top of the VMMC mechanism, including the NX message-passing interface, Sun RPC, stream sockets, and specialized RPC. The first three are fully compatible with existing systems. Our experience shows that the VMMC mechanism supports these message-passing interfaces well. When zero-copy protocols are allowed by the semantics of the interface, VMMC can effectively deliver to applications almost all of the raw hardware's communication performance.",
author = "Felten, {Edward W.} and Alpert, {Richard D.} and Angelos Bilas and Blumrich, {Matthias A.} and Clark, {Douglas W.} and Damianakis, {Stefanos N.} and Cezary Dubnicki and Liviu Iftode and Kai Li",
year = "1996",
month = "1",
day = "1",
language = "English (US)",
pages = "296--307",
journal = "Proceedings - International Symposium on Computer Architecture",
issn = "1063-6897",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

Early experience with message-passing on the SHRIMP multicomputer. / Felten, Edward W.; Alpert, Richard D.; Bilas, Angelos; Blumrich, Matthias A.; Clark, Douglas W.; Damianakis, Stefanos N.; Dubnicki, Cezary; Iftode, Liviu; Li, Kai.

In: Conference Proceedings - Annual International Symposium on Computer Architecture, ISCA, 01.01.1996, p. 296-307.

Research output: Contribution to journalConference article

TY - JOUR

T1 - Early experience with message-passing on the SHRIMP multicomputer

AU - Felten, Edward W.

AU - Alpert, Richard D.

AU - Bilas, Angelos

AU - Blumrich, Matthias A.

AU - Clark, Douglas W.

AU - Damianakis, Stefanos N.

AU - Dubnicki, Cezary

AU - Iftode, Liviu

AU - Li, Kai

PY - 1996/1/1

Y1 - 1996/1/1

N2 - The SHRIMP multicomputer provides virtual memory-mapped communication (VMMC), which supports protected, user-level message passing, allows user programs to perform their own buffer management, and separates data transfers from control transfers so that a data transfer can be done without the intervention of the receiving node CPU. An important question is whether such a mechanism can indeed deliver all of the available hardware performance to applications which use conventional message-passing libraries. This paper reports our early experience with message-passing on a small, working SHRIMP multicomputer. We have implemented several user-level communication libraries on top of the VMMC mechanism, including the NX message-passing interface, Sun RPC, stream sockets, and specialized RPC. The first three are fully compatible with existing systems. Our experience shows that the VMMC mechanism supports these message-passing interfaces well. When zero-copy protocols are allowed by the semantics of the interface, VMMC can effectively deliver to applications almost all of the raw hardware's communication performance.

AB - The SHRIMP multicomputer provides virtual memory-mapped communication (VMMC), which supports protected, user-level message passing, allows user programs to perform their own buffer management, and separates data transfers from control transfers so that a data transfer can be done without the intervention of the receiving node CPU. An important question is whether such a mechanism can indeed deliver all of the available hardware performance to applications which use conventional message-passing libraries. This paper reports our early experience with message-passing on a small, working SHRIMP multicomputer. We have implemented several user-level communication libraries on top of the VMMC mechanism, including the NX message-passing interface, Sun RPC, stream sockets, and specialized RPC. The first three are fully compatible with existing systems. Our experience shows that the VMMC mechanism supports these message-passing interfaces well. When zero-copy protocols are allowed by the semantics of the interface, VMMC can effectively deliver to applications almost all of the raw hardware's communication performance.

UR - http://www.scopus.com/inward/record.url?scp=0029666631&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0029666631&partnerID=8YFLogxK

M3 - Conference article

SP - 296

EP - 307

JO - Proceedings - International Symposium on Computer Architecture

JF - Proceedings - International Symposium on Computer Architecture

SN - 1063-6897

ER -