Optimization of condensed matter physics application with openMP tasking model

Joel Criado; Marta Garcia-Gasulla; Jesús Labarta; Arghya Chatterjee; Oscar Hernandez; Raül Sirvent; Gonzalo Alvarez

doi:10.1007/978-3-030-28596-8_20

Optimization of condensed matter physics application with openMP tasking model

Joel Criado, Marta Garcia-Gasulla, Jesús Labarta, Arghya Chatterjee, Oscar Hernandez, Raül Sirvent, Gonzalo Alvarez

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

4 Scopus citations

Abstract

The Density Matrix Renormalization Group (DMRG++) is a condensed matter physics application used to study superconductivity properties of materials. It’s main computations consist of calculating hamiltonian matrix which requires sparse matrix-vector multiplications. This paper presents task-based parallelization and optimization strategies of the Hamiltonian algorithm. The algorithm is implemented as a mini-application in C++ and parallelized with OpenMP. The optimization leverages tasking features, such as dependencies or priorities included in the OpenMP standard 4.5. The code refactoring targets performance as much as programmability. The optimized version achieves a speedup of 8.0 × with 8 threads and 20.5 × with 40 threads on a Power9 computing node while reducing the memory consumption to 90 MB with respect to the original code, by adding less than ten OpenMP directives.

Original language	English
Title of host publication	OpenMP
Subtitle of host publication	Conquering the Full Hardware Spectrum - 15th International Workshop on OpenMP, IWOMP 2019, Proceedings
Editors	Xing Fan, Oliver Sinnen, Nasser Giacaman, Bronis R. de Supinski
Publisher	Springer Verlag
Pages	291-305
Number of pages	15
ISBN (Print)	9783030285951
DOIs	https://doi.org/10.1007/978-3-030-28596-8_20
State	Published - 2019
Event	15th International Workshop on OpenMP, IWOMP 2019 - Auckland, New Zealand Duration: Sep 11 2019 → Sep 13 2019

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	11718 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	15th International Workshop on OpenMP, IWOMP 2019
Country/Territory	New Zealand
City	Auckland
Period	09/11/19 → 09/13/19

Funding

This work is partially supported by the Spanish Government through Programa Severo Ochoa (SEV-2015-0493), by the Spanish Ministry of Science and Technology (project TIN2015-65316-P), by the Generalitat de Catalunya (contract 2017-SGR-1414) and by the BSC-IBM Deep Learning Research Agreement, under JSA “Application porting, analysis and optimization for POWER and POWER AI”. This work was partially supported by the Scientific Discovery through Advanced Computing (SciDAC) program funded by U.S. Department of Energy, Office of Science, Advanced Scientific Computing Research and Basic Energy Sciences, Division of Materials Sciences and Engineering. This research used resources of the Oak Ridge Leadership Computing Facility at the Oak Ridge National Laboratory, which is supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC05-00OR22725.

Keywords

Analysis
Dependencies
OpenMP
Optimization
Tasks

Access to Document

10.1007/978-3-030-28596-8_20

Cite this

Criado, J., Garcia-Gasulla, M., Labarta, J., Chatterjee, A., Hernandez, O., Sirvent, R., & Alvarez, G. (2019). Optimization of condensed matter physics application with openMP tasking model. In X. Fan, O. Sinnen, N. Giacaman, & B. R. de Supinski (Eds.), OpenMP: Conquering the Full Hardware Spectrum - 15th International Workshop on OpenMP, IWOMP 2019, Proceedings (pp. 291-305). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11718 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-030-28596-8_20

Criado, Joel ; Garcia-Gasulla, Marta ; Labarta, Jesús et al. / Optimization of condensed matter physics application with openMP tasking model. OpenMP: Conquering the Full Hardware Spectrum - 15th International Workshop on OpenMP, IWOMP 2019, Proceedings. editor / Xing Fan ; Oliver Sinnen ; Nasser Giacaman ; Bronis R. de Supinski. Springer Verlag, 2019. pp. 291-305 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{1886c7b50ee2424abdf78c7767a51743,

title = "Optimization of condensed matter physics application with openMP tasking model",

abstract = "The Density Matrix Renormalization Group (DMRG++) is a condensed matter physics application used to study superconductivity properties of materials. It{\textquoteright}s main computations consist of calculating hamiltonian matrix which requires sparse matrix-vector multiplications. This paper presents task-based parallelization and optimization strategies of the Hamiltonian algorithm. The algorithm is implemented as a mini-application in C++ and parallelized with OpenMP. The optimization leverages tasking features, such as dependencies or priorities included in the OpenMP standard 4.5. The code refactoring targets performance as much as programmability. The optimized version achieves a speedup of 8.0 × with 8 threads and 20.5 × with 40 threads on a Power9 computing node while reducing the memory consumption to 90 MB with respect to the original code, by adding less than ten OpenMP directives.",

keywords = "Analysis, Dependencies, OpenMP, Optimization, Tasks",

author = "Joel Criado and Marta Garcia-Gasulla and Jes{\'u}s Labarta and Arghya Chatterjee and Oscar Hernandez and Ra{\"u}l Sirvent and Gonzalo Alvarez",

note = "Publisher Copyright: {\textcopyright} Springer Nature Switzerland AG 2019.; 15th International Workshop on OpenMP, IWOMP 2019 ; Conference date: 11-09-2019 Through 13-09-2019",

year = "2019",

doi = "10.1007/978-3-030-28596-8_20",

language = "English",

isbn = "9783030285951",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "291--305",

editor = "Xing Fan and Oliver Sinnen and Nasser Giacaman and {de Supinski}, {Bronis R.}",

booktitle = "OpenMP",

}

Criado, J, Garcia-Gasulla, M, Labarta, J, Chatterjee, A, Hernandez, O, Sirvent, R & Alvarez, G 2019, Optimization of condensed matter physics application with openMP tasking model. in X Fan, O Sinnen, N Giacaman & BR de Supinski (eds), OpenMP: Conquering the Full Hardware Spectrum - 15th International Workshop on OpenMP, IWOMP 2019, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11718 LNCS, Springer Verlag, pp. 291-305, 15th International Workshop on OpenMP, IWOMP 2019, Auckland, New Zealand, 09/11/19. https://doi.org/10.1007/978-3-030-28596-8_20

Optimization of condensed matter physics application with openMP tasking model. / Criado, Joel; Garcia-Gasulla, Marta; Labarta, Jesús et al.
OpenMP: Conquering the Full Hardware Spectrum - 15th International Workshop on OpenMP, IWOMP 2019, Proceedings. ed. / Xing Fan; Oliver Sinnen; Nasser Giacaman; Bronis R. de Supinski. Springer Verlag, 2019. p. 291-305 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11718 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Optimization of condensed matter physics application with openMP tasking model

AU - Criado, Joel

AU - Garcia-Gasulla, Marta

AU - Labarta, Jesús

AU - Chatterjee, Arghya

AU - Hernandez, Oscar

AU - Sirvent, Raül

AU - Alvarez, Gonzalo

N1 - Publisher Copyright: © Springer Nature Switzerland AG 2019.

PY - 2019

Y1 - 2019

N2 - The Density Matrix Renormalization Group (DMRG++) is a condensed matter physics application used to study superconductivity properties of materials. It’s main computations consist of calculating hamiltonian matrix which requires sparse matrix-vector multiplications. This paper presents task-based parallelization and optimization strategies of the Hamiltonian algorithm. The algorithm is implemented as a mini-application in C++ and parallelized with OpenMP. The optimization leverages tasking features, such as dependencies or priorities included in the OpenMP standard 4.5. The code refactoring targets performance as much as programmability. The optimized version achieves a speedup of 8.0 × with 8 threads and 20.5 × with 40 threads on a Power9 computing node while reducing the memory consumption to 90 MB with respect to the original code, by adding less than ten OpenMP directives.

AB - The Density Matrix Renormalization Group (DMRG++) is a condensed matter physics application used to study superconductivity properties of materials. It’s main computations consist of calculating hamiltonian matrix which requires sparse matrix-vector multiplications. This paper presents task-based parallelization and optimization strategies of the Hamiltonian algorithm. The algorithm is implemented as a mini-application in C++ and parallelized with OpenMP. The optimization leverages tasking features, such as dependencies or priorities included in the OpenMP standard 4.5. The code refactoring targets performance as much as programmability. The optimized version achieves a speedup of 8.0 × with 8 threads and 20.5 × with 40 threads on a Power9 computing node while reducing the memory consumption to 90 MB with respect to the original code, by adding less than ten OpenMP directives.

KW - Analysis

KW - Dependencies

KW - OpenMP

KW - Optimization

KW - Tasks

UR - http://www.scopus.com/inward/record.url?scp=85072851381&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-28596-8_20

DO - 10.1007/978-3-030-28596-8_20

M3 - Conference contribution

AN - SCOPUS:85072851381

SN - 9783030285951

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 291

EP - 305

BT - OpenMP

A2 - Fan, Xing

A2 - Sinnen, Oliver

A2 - Giacaman, Nasser

A2 - de Supinski, Bronis R.

PB - Springer Verlag

T2 - 15th International Workshop on OpenMP, IWOMP 2019

Y2 - 11 September 2019 through 13 September 2019

ER -

Criado J, Garcia-Gasulla M, Labarta J, Chatterjee A, Hernandez O, Sirvent R et al. Optimization of condensed matter physics application with openMP tasking model. In Fan X, Sinnen O, Giacaman N, de Supinski BR, editors, OpenMP: Conquering the Full Hardware Spectrum - 15th International Workshop on OpenMP, IWOMP 2019, Proceedings. Springer Verlag. 2019. p. 291-305. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-28596-8_20

Optimization of condensed matter physics application with openMP tasking model

Abstract

Publication series

Conference

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this