Performance analysis of a parallel molecular dynamics program

Amitabh B. Sinha; Klaus Schulten; Helmut Heller

doi:10.1016/0010-4655(94)90005-1

Performance analysis of a parallel molecular dynamics program

Amitabh B. Sinha, Klaus Schulten, Helmut Heller

Research output: Contribution to journal › Article › peer-review

Abstract

EGO is a parallel molecular dynamics program running on Transputers. We conducted a performance analysis of the EGO program in order to determine whether it was effectively using the computational resources of Transputers. Our first concern was whether communication was overlapped with computation, so that the overheads due to communication not overlapped with computation were less. With the assistance of performance tools such as UPSHOT, and with instrumentation of the EGO program itself, we were able to determine that only 8% of the execution time of the EGO program was spent in non-overlapping communication. Our next concern was that the MFLOPS rating of the EGO program was 0.25 MFLOPS, while the Transputers have a sustained rating of 1.5 MFLOPS. We measured MFLOPS ratings of small blocks of OCCAM code and determined that they matched the performance of the EGO code.

Original language	English (US)
Pages (from-to)	265-278
Number of pages	14
Journal	Computer Physics Communications
Volume	78
Issue number	3
DOIs	https://doi.org/10.1016/0010-4655(94)90005-1
State	Published - Jan 1994
Externally published	Yes

ASJC Scopus subject areas

Hardware and Architecture
General Physics and Astronomy

Online availability

10.1016/0010-4655(94)90005-1

Library availability

Discover UIUC Full Text

Cite this

@article{1cadb08b38bc4dc5bce6d30d2a030871,

title = "Performance analysis of a parallel molecular dynamics program",

abstract = "EGO is a parallel molecular dynamics program running on Transputers. We conducted a performance analysis of the EGO program in order to determine whether it was effectively using the computational resources of Transputers. Our first concern was whether communication was overlapped with computation, so that the overheads due to communication not overlapped with computation were less. With the assistance of performance tools such as UPSHOT, and with instrumentation of the EGO program itself, we were able to determine that only 8% of the execution time of the EGO program was spent in non-overlapping communication. Our next concern was that the MFLOPS rating of the EGO program was 0.25 MFLOPS, while the Transputers have a sustained rating of 1.5 MFLOPS. We measured MFLOPS ratings of small blocks of OCCAM code and determined that they matched the performance of the EGO code.",

author = "Sinha, {Amitabh B.} and Klaus Schulten and Helmut Heller",

note = "Funding Information: This work was supported by the National Institutes of Health Resource Concurrent Biological Computing under grant number 1 P41 RR-05969*01. We thank Rusty Lusk at Argonne National Laboratory for assistance with the Upshot",

year = "1994",

month = jan,

doi = "10.1016/0010-4655(94)90005-1",

language = "English (US)",

volume = "78",

pages = "265--278",

journal = "Computer Physics Communications",

issn = "0010-4655",

publisher = "Elsevier B.V.",

number = "3",

}

TY - JOUR

T1 - Performance analysis of a parallel molecular dynamics program

AU - Sinha, Amitabh B.

AU - Schulten, Klaus

AU - Heller, Helmut

N1 - Funding Information: This work was supported by the National Institutes of Health Resource Concurrent Biological Computing under grant number 1 P41 RR-05969*01. We thank Rusty Lusk at Argonne National Laboratory for assistance with the Upshot

PY - 1994/1

Y1 - 1994/1

N2 - EGO is a parallel molecular dynamics program running on Transputers. We conducted a performance analysis of the EGO program in order to determine whether it was effectively using the computational resources of Transputers. Our first concern was whether communication was overlapped with computation, so that the overheads due to communication not overlapped with computation were less. With the assistance of performance tools such as UPSHOT, and with instrumentation of the EGO program itself, we were able to determine that only 8% of the execution time of the EGO program was spent in non-overlapping communication. Our next concern was that the MFLOPS rating of the EGO program was 0.25 MFLOPS, while the Transputers have a sustained rating of 1.5 MFLOPS. We measured MFLOPS ratings of small blocks of OCCAM code and determined that they matched the performance of the EGO code.

AB - EGO is a parallel molecular dynamics program running on Transputers. We conducted a performance analysis of the EGO program in order to determine whether it was effectively using the computational resources of Transputers. Our first concern was whether communication was overlapped with computation, so that the overheads due to communication not overlapped with computation were less. With the assistance of performance tools such as UPSHOT, and with instrumentation of the EGO program itself, we were able to determine that only 8% of the execution time of the EGO program was spent in non-overlapping communication. Our next concern was that the MFLOPS rating of the EGO program was 0.25 MFLOPS, while the Transputers have a sustained rating of 1.5 MFLOPS. We measured MFLOPS ratings of small blocks of OCCAM code and determined that they matched the performance of the EGO code.

UR - http://www.scopus.com/inward/record.url?scp=0027989161&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0027989161&partnerID=8YFLogxK

U2 - 10.1016/0010-4655(94)90005-1

DO - 10.1016/0010-4655(94)90005-1

M3 - Article

AN - SCOPUS:0027989161

SN - 0010-4655

VL - 78

SP - 265

EP - 278

JO - Computer Physics Communications

JF - Computer Physics Communications

IS - 3

ER -

Performance analysis of a parallel molecular dynamics program

Abstract

ASJC Scopus subject areas

Online availability

Library availability

Related links

Fingerprint

Cite this