Efficient message passing interface (MPI) for parallel computing on clusters of workstations

Jehoshua Bruck; Danny Dolev; Ching Tien Ho; Marcel Catalin Rosu; Ray Strong

doi:10.1145/215399.215421

Efficient message passing interface (MPI) for parallel computing on clusters of workstations

Jehoshua Bruck^*
, Danny Dolev
, Ching Tien Ho
, Marcel Catalin Rosu
, Ray Strong

^*Corresponding author for this work

Research output: Contribution to conference › Paper › peer-review

19 Scopus citations

Abstract

An efficient design and implementation of the collective communication part in a Message Passing Interface (MPI) that is optimized for clusters of workstations is described. The system which consist of two main components, the MPI-CCL layer and a User-level Reliable Transport Protocol (URTP), is integrated with the operating system via an efficient kernel extension mechanism. The system is then implemented on a collection of IBM RS/6000 workstations connected via a 10Mbit Ethernet LAN. Results indicate that the performance of the MPI Broadcast (on top of Ethernet) is about twice as fast as a recently published software implementation of broadcast on top of ATM.

Original language	English
Pages	64-73
Number of pages	10
DOIs	https://doi.org/10.1145/215399.215421
State	Published - 1995
Externally published	Yes
Event	Proceedings of the 7th Annual ACM Symposium on Parallel Algorithms and Architectures, SPAA'95 - Santa Barbara, CA, USA Duration: 17 Jul 1995 → 19 Jul 1995

Conference

Conference	Proceedings of the 7th Annual ACM Symposium on Parallel Algorithms and Architectures, SPAA'95
City	Santa Barbara, CA, USA
Period	17/07/95 → 19/07/95

Access to Document

10.1145/215399.215421

Cite this

@conference{992350971c0648798d44f1165f52d6cf,

title = "Efficient message passing interface (MPI) for parallel computing on clusters of workstations",

abstract = "An efficient design and implementation of the collective communication part in a Message Passing Interface (MPI) that is optimized for clusters of workstations is described. The system which consist of two main components, the MPI-CCL layer and a User-level Reliable Transport Protocol (URTP), is integrated with the operating system via an efficient kernel extension mechanism. The system is then implemented on a collection of IBM RS/6000 workstations connected via a 10Mbit Ethernet LAN. Results indicate that the performance of the MPI Broadcast (on top of Ethernet) is about twice as fast as a recently published software implementation of broadcast on top of ATM.",

author = "Jehoshua Bruck and Danny Dolev and Ho, \{Ching Tien\} and Rosu, \{Marcel Catalin\} and Ray Strong",

year = "1995",

doi = "10.1145/215399.215421",

language = "אנגלית",

pages = "64--73",

note = "Proceedings of the 7th Annual ACM Symposium on Parallel Algorithms and Architectures, SPAA'95 ; Conference date: 17-07-1995 Through 19-07-1995",

}

TY - CONF

T1 - Efficient message passing interface (MPI) for parallel computing on clusters of workstations

AU - Bruck, Jehoshua

AU - Dolev, Danny

AU - Ho, Ching Tien

AU - Rosu, Marcel Catalin

AU - Strong, Ray

PY - 1995

Y1 - 1995

N2 - An efficient design and implementation of the collective communication part in a Message Passing Interface (MPI) that is optimized for clusters of workstations is described. The system which consist of two main components, the MPI-CCL layer and a User-level Reliable Transport Protocol (URTP), is integrated with the operating system via an efficient kernel extension mechanism. The system is then implemented on a collection of IBM RS/6000 workstations connected via a 10Mbit Ethernet LAN. Results indicate that the performance of the MPI Broadcast (on top of Ethernet) is about twice as fast as a recently published software implementation of broadcast on top of ATM.

AB - An efficient design and implementation of the collective communication part in a Message Passing Interface (MPI) that is optimized for clusters of workstations is described. The system which consist of two main components, the MPI-CCL layer and a User-level Reliable Transport Protocol (URTP), is integrated with the operating system via an efficient kernel extension mechanism. The system is then implemented on a collection of IBM RS/6000 workstations connected via a 10Mbit Ethernet LAN. Results indicate that the performance of the MPI Broadcast (on top of Ethernet) is about twice as fast as a recently published software implementation of broadcast on top of ATM.

UR - https://www.scopus.com/pages/publications/0029195611

U2 - 10.1145/215399.215421

DO - 10.1145/215399.215421

M3 - ???researchoutput.researchoutputtypes.contributiontoconference.paper???

AN - SCOPUS:0029195611

SP - 64

EP - 73

T2 - Proceedings of the 7th Annual ACM Symposium on Parallel Algorithms and Architectures, SPAA'95

Y2 - 17 July 1995 through 19 July 1995

ER -

Efficient message passing interface (MPI) for parallel computing on clusters of workstations

Abstract

Conference

Access to Document

Other files and links

Fingerprint

Cite this