MILC staggered conjugate gradient performance on Intel KNL

Ruizi Li, Carleton DeTar, Douglas Doerfler, Steven Gottlieb, Ashish Jha, Dhiraj Kalamkar, William D Toussaint

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

We review our work done to optimize the staggered conjugate gradient (CG) algorithm in the MILC code for use with the Intel Knights Landing (KNL) architecture. KNL is the second generation Intel Xeon Phi processor. It is capable of massive thread parallelism, data parallelism, and high on-board memory bandwidth and is being adopted in supercomputing centers for scientific research. The CG solver consumes the majority of time in production running, so we have spent most of our effort on it. We compare performance of an MPI+OpenMP baseline version of the MILC code with a version incorporating the QPhiX staggered CG solver, for both one-node and multi-node runs.

Original languageEnglish (US)
JournalProceedings of Science
VolumePart F128557
StatePublished - 2016

Fingerprint

Landing
Bandwidth
Data storage equipment

ASJC Scopus subject areas

  • General

Cite this

Li, R., DeTar, C., Doerfler, D., Gottlieb, S., Jha, A., Kalamkar, D., & Toussaint, W. D. (2016). MILC staggered conjugate gradient performance on Intel KNL. Proceedings of Science, Part F128557.

MILC staggered conjugate gradient performance on Intel KNL. / Li, Ruizi; DeTar, Carleton; Doerfler, Douglas; Gottlieb, Steven; Jha, Ashish; Kalamkar, Dhiraj; Toussaint, William D.

In: Proceedings of Science, Vol. Part F128557, 2016.

Research output: Contribution to journalArticle

Li, R, DeTar, C, Doerfler, D, Gottlieb, S, Jha, A, Kalamkar, D & Toussaint, WD 2016, 'MILC staggered conjugate gradient performance on Intel KNL', Proceedings of Science, vol. Part F128557.
Li R, DeTar C, Doerfler D, Gottlieb S, Jha A, Kalamkar D et al. MILC staggered conjugate gradient performance on Intel KNL. Proceedings of Science. 2016;Part F128557.
Li, Ruizi ; DeTar, Carleton ; Doerfler, Douglas ; Gottlieb, Steven ; Jha, Ashish ; Kalamkar, Dhiraj ; Toussaint, William D. / MILC staggered conjugate gradient performance on Intel KNL. In: Proceedings of Science. 2016 ; Vol. Part F128557.
@article{dd34b7c7d35340f4aeb065c092a16279,
title = "MILC staggered conjugate gradient performance on Intel KNL",
abstract = "We review our work done to optimize the staggered conjugate gradient (CG) algorithm in the MILC code for use with the Intel Knights Landing (KNL) architecture. KNL is the second generation Intel Xeon Phi processor. It is capable of massive thread parallelism, data parallelism, and high on-board memory bandwidth and is being adopted in supercomputing centers for scientific research. The CG solver consumes the majority of time in production running, so we have spent most of our effort on it. We compare performance of an MPI+OpenMP baseline version of the MILC code with a version incorporating the QPhiX staggered CG solver, for both one-node and multi-node runs.",
author = "Ruizi Li and Carleton DeTar and Douglas Doerfler and Steven Gottlieb and Ashish Jha and Dhiraj Kalamkar and Toussaint, {William D}",
year = "2016",
language = "English (US)",
volume = "Part F128557",
journal = "Proceedings of Science",
issn = "1824-8039",
publisher = "Sissa Medialab Srl",

}

TY - JOUR

T1 - MILC staggered conjugate gradient performance on Intel KNL

AU - Li, Ruizi

AU - DeTar, Carleton

AU - Doerfler, Douglas

AU - Gottlieb, Steven

AU - Jha, Ashish

AU - Kalamkar, Dhiraj

AU - Toussaint, William D

PY - 2016

Y1 - 2016

N2 - We review our work done to optimize the staggered conjugate gradient (CG) algorithm in the MILC code for use with the Intel Knights Landing (KNL) architecture. KNL is the second generation Intel Xeon Phi processor. It is capable of massive thread parallelism, data parallelism, and high on-board memory bandwidth and is being adopted in supercomputing centers for scientific research. The CG solver consumes the majority of time in production running, so we have spent most of our effort on it. We compare performance of an MPI+OpenMP baseline version of the MILC code with a version incorporating the QPhiX staggered CG solver, for both one-node and multi-node runs.

AB - We review our work done to optimize the staggered conjugate gradient (CG) algorithm in the MILC code for use with the Intel Knights Landing (KNL) architecture. KNL is the second generation Intel Xeon Phi processor. It is capable of massive thread parallelism, data parallelism, and high on-board memory bandwidth and is being adopted in supercomputing centers for scientific research. The CG solver consumes the majority of time in production running, so we have spent most of our effort on it. We compare performance of an MPI+OpenMP baseline version of the MILC code with a version incorporating the QPhiX staggered CG solver, for both one-node and multi-node runs.

UR - http://www.scopus.com/inward/record.url?scp=85025828626&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85025828626&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:85025828626

VL - Part F128557

JO - Proceedings of Science

JF - Proceedings of Science

SN - 1824-8039

ER -