Cell processor implementation of a MILC lattice QCD application

Guochun Shi, Volodymyr Kindratenko, Steven Gottlieb

Research output: Contribution to journalConference article

Abstract

We present results of the implementation of one MILC lattice QCD application-simulation with dynamical clover fermions using the hybrid-molecular dynamics R algorithm-on the Cell Broadband Engine processor. Fifty-four individual computational kernels responsible for 98.8% of the overall execution time were ported to the Cell's Synergistic Processing Elements (SPEs). The remaining application framework, including MPI-based distributed code execution, was left to the Cell's PowerPC processor. We observe that we only infrequently achieve more than 10 GFLOPS with any of the kernels, which is just over 4% of the Cell's peak performance. At the same time, many of the kernels are sustaining a bandwidth close to 20 GB/s, which is 78% of the Cell's peak. This indicates that the application performance is limited by the bandwidth between the main memory and the SPEs. In spite of this limitation, speedups of 8.7× (for 8×8×16×16 lattice) and 9.6× (for 16×16×16×16 lattice) were achieved when comparing a 3.2 GHz Cell processor to a single core of a 2.33 GHz Intel Xeon processor. When comparing the code scaled up to execute on a dual-Cell blade and a quad-core dual-chip Intel Xeon blade, the speedups are 1.5× (8×8×16×16 lattice) and 4.1× (16×16×16×16 lattice).

Original languageEnglish (US)
JournalProceedings of Science
Volume66
StatePublished - Jan 1 2008
Event26th International Symposium on Lattice Field Theory, LATTICE 2008 - Williamsburg, United States
Duration: Jul 14 2008Jul 19 2008

Fingerprint

Bandwidth
Fermions
Processing
Molecular dynamics
Engines
Data storage equipment

ASJC Scopus subject areas

  • General

Cite this

Cell processor implementation of a MILC lattice QCD application. / Shi, Guochun; Kindratenko, Volodymyr; Gottlieb, Steven.

In: Proceedings of Science, Vol. 66, 01.01.2008.

Research output: Contribution to journalConference article

@article{fbb85a9eb0554e82890844ea7e1d2288,
title = "Cell processor implementation of a MILC lattice QCD application",
abstract = "We present results of the implementation of one MILC lattice QCD application-simulation with dynamical clover fermions using the hybrid-molecular dynamics R algorithm-on the Cell Broadband Engine processor. Fifty-four individual computational kernels responsible for 98.8{\%} of the overall execution time were ported to the Cell's Synergistic Processing Elements (SPEs). The remaining application framework, including MPI-based distributed code execution, was left to the Cell's PowerPC processor. We observe that we only infrequently achieve more than 10 GFLOPS with any of the kernels, which is just over 4{\%} of the Cell's peak performance. At the same time, many of the kernels are sustaining a bandwidth close to 20 GB/s, which is 78{\%} of the Cell's peak. This indicates that the application performance is limited by the bandwidth between the main memory and the SPEs. In spite of this limitation, speedups of 8.7× (for 8×8×16×16 lattice) and 9.6× (for 16×16×16×16 lattice) were achieved when comparing a 3.2 GHz Cell processor to a single core of a 2.33 GHz Intel Xeon processor. When comparing the code scaled up to execute on a dual-Cell blade and a quad-core dual-chip Intel Xeon blade, the speedups are 1.5× (8×8×16×16 lattice) and 4.1× (16×16×16×16 lattice).",
author = "Guochun Shi and Volodymyr Kindratenko and Steven Gottlieb",
year = "2008",
month = "1",
day = "1",
language = "English (US)",
volume = "66",
journal = "Proceedings of Science",
issn = "1824-8039",
publisher = "Sissa Medialab Srl",

}

TY - JOUR

T1 - Cell processor implementation of a MILC lattice QCD application

AU - Shi, Guochun

AU - Kindratenko, Volodymyr

AU - Gottlieb, Steven

PY - 2008/1/1

Y1 - 2008/1/1

N2 - We present results of the implementation of one MILC lattice QCD application-simulation with dynamical clover fermions using the hybrid-molecular dynamics R algorithm-on the Cell Broadband Engine processor. Fifty-four individual computational kernels responsible for 98.8% of the overall execution time were ported to the Cell's Synergistic Processing Elements (SPEs). The remaining application framework, including MPI-based distributed code execution, was left to the Cell's PowerPC processor. We observe that we only infrequently achieve more than 10 GFLOPS with any of the kernels, which is just over 4% of the Cell's peak performance. At the same time, many of the kernels are sustaining a bandwidth close to 20 GB/s, which is 78% of the Cell's peak. This indicates that the application performance is limited by the bandwidth between the main memory and the SPEs. In spite of this limitation, speedups of 8.7× (for 8×8×16×16 lattice) and 9.6× (for 16×16×16×16 lattice) were achieved when comparing a 3.2 GHz Cell processor to a single core of a 2.33 GHz Intel Xeon processor. When comparing the code scaled up to execute on a dual-Cell blade and a quad-core dual-chip Intel Xeon blade, the speedups are 1.5× (8×8×16×16 lattice) and 4.1× (16×16×16×16 lattice).

AB - We present results of the implementation of one MILC lattice QCD application-simulation with dynamical clover fermions using the hybrid-molecular dynamics R algorithm-on the Cell Broadband Engine processor. Fifty-four individual computational kernels responsible for 98.8% of the overall execution time were ported to the Cell's Synergistic Processing Elements (SPEs). The remaining application framework, including MPI-based distributed code execution, was left to the Cell's PowerPC processor. We observe that we only infrequently achieve more than 10 GFLOPS with any of the kernels, which is just over 4% of the Cell's peak performance. At the same time, many of the kernels are sustaining a bandwidth close to 20 GB/s, which is 78% of the Cell's peak. This indicates that the application performance is limited by the bandwidth between the main memory and the SPEs. In spite of this limitation, speedups of 8.7× (for 8×8×16×16 lattice) and 9.6× (for 16×16×16×16 lattice) were achieved when comparing a 3.2 GHz Cell processor to a single core of a 2.33 GHz Intel Xeon processor. When comparing the code scaled up to execute on a dual-Cell blade and a quad-core dual-chip Intel Xeon blade, the speedups are 1.5× (8×8×16×16 lattice) and 4.1× (16×16×16×16 lattice).

UR - http://www.scopus.com/inward/record.url?scp=78650841958&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=78650841958&partnerID=8YFLogxK

M3 - Conference article

AN - SCOPUS:78650841958

VL - 66

JO - Proceedings of Science

JF - Proceedings of Science

SN - 1824-8039

ER -