Enabling execution of a legacy CFD mini application on accelerators using OpenMP

Ioannis Nompelis; Gabriele Jost; Alice Koniges; Christopher Daley; David Eder; Christopher Stone

doi:10.1007/978-3-030-50743-5_14

Enabling execution of a legacy CFD mini application on accelerators using OpenMP

Ioannis Nompelis, Gabriele Jost, Alice Koniges, Christopher Daley, David Eder, Christopher Stone

Aerospace Engineering and Mechanics

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

Abstract

We describe the process and outcome of our efforts to port a legacy Fortran benchmark code to heterogeneous GPU-accelerated computing architectures using OpenMP. The benchmark code is one of the multi-zone NAS Parallel Benchmarks (NPB-MZ) called SP-MZ. This “mini-app” mimics the computation and data movement that is found in popular legacy and modern implicit computational fluid dynamics (CFD) solvers. Our objective was to examine how efficiently legacy Fortran codes can be ported to accelerators by leveraging OpenMP directives. We describe the development and optimization process and demonstrate the performance impact of various code modifications. We show select profiling results from the NVIDIA Visual Profiler (nvpp) to help others diagnose and overcome performance issues in their own applications. We present results for two compute systems endowed with NVIDIA V100 accelerators.

Original language	English (US)
Title of host publication	High Performance Computing - 35th International Conference, ISC High Performance 2020, Proceedings
Editors	Ponnuswamy Sadayappan, Bradford L. Chamberlain, Guido Juckeland, Hatem Ltaief
Publisher	Springer
Pages	270-287
Number of pages	18
ISBN (Print)	9783030507428
DOIs	https://doi.org/10.1007/978-3-030-50743-5_14
State	Published - 2020
Event	35th International Conference on High Performance Computing, ISC High Performance 2020 - Frankfurt, Germany Duration: Jun 22 2020 → Jun 25 2020

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	12151 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	35th International Conference on High Performance Computing, ISC High Performance 2020
Country/Territory	Germany
City	Frankfurt
Period	6/22/20 → 6/25/20

Bibliographical note

Publisher Copyright:
© Springer Nature Switzerland AG 2020.

Keywords

Accelerator
Fortran
GPU
Implicit CFD
OpenMP

Publisher link

10.1007/978-3-030-50743-5_14

Find It

Find It at U of M Libraries

Cite this

Nompelis, I., Jost, G., Koniges, A., Daley, C., Eder, D., & Stone, C. (2020). Enabling execution of a legacy CFD mini application on accelerators using OpenMP. In P. Sadayappan, B. L. Chamberlain, G. Juckeland, & H. Ltaief (Eds.), High Performance Computing - 35th International Conference, ISC High Performance 2020, Proceedings (pp. 270-287). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12151 LNCS). Springer. https://doi.org/10.1007/978-3-030-50743-5_14

Enabling execution of a legacy CFD mini application on accelerators using OpenMP. / Nompelis, Ioannis; Jost, Gabriele; Koniges, Alice et al.
High Performance Computing - 35th International Conference, ISC High Performance 2020, Proceedings. ed. / Ponnuswamy Sadayappan; Bradford L. Chamberlain; Guido Juckeland; Hatem Ltaief. Springer, 2020. p. 270-287 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12151 LNCS).

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

Nompelis, I, Jost, G, Koniges, A, Daley, C, Eder, D & Stone, C 2020, Enabling execution of a legacy CFD mini application on accelerators using OpenMP. in P Sadayappan, BL Chamberlain, G Juckeland & H Ltaief (eds), High Performance Computing - 35th International Conference, ISC High Performance 2020, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 12151 LNCS, Springer, pp. 270-287, 35th International Conference on High Performance Computing, ISC High Performance 2020, Frankfurt, Germany, 6/22/20. https://doi.org/10.1007/978-3-030-50743-5_14

Nompelis I, Jost G, Koniges A, Daley C, Eder D, Stone C. Enabling execution of a legacy CFD mini application on accelerators using OpenMP. In Sadayappan P, Chamberlain BL, Juckeland G, Ltaief H, editors, High Performance Computing - 35th International Conference, ISC High Performance 2020, Proceedings. Springer. 2020. p. 270-287. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-50743-5_14

Nompelis, Ioannis ; Jost, Gabriele ; Koniges, Alice et al. / Enabling execution of a legacy CFD mini application on accelerators using OpenMP. High Performance Computing - 35th International Conference, ISC High Performance 2020, Proceedings. editor / Ponnuswamy Sadayappan ; Bradford L. Chamberlain ; Guido Juckeland ; Hatem Ltaief. Springer, 2020. pp. 270-287 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{d8b0cdb390f145338aa4756d87fd71c4,

title = "Enabling execution of a legacy CFD mini application on accelerators using OpenMP",

abstract = "We describe the process and outcome of our efforts to port a legacy Fortran benchmark code to heterogeneous GPU-accelerated computing architectures using OpenMP. The benchmark code is one of the multi-zone NAS Parallel Benchmarks (NPB-MZ) called SP-MZ. This “mini-app” mimics the computation and data movement that is found in popular legacy and modern implicit computational fluid dynamics (CFD) solvers. Our objective was to examine how efficiently legacy Fortran codes can be ported to accelerators by leveraging OpenMP directives. We describe the development and optimization process and demonstrate the performance impact of various code modifications. We show select profiling results from the NVIDIA Visual Profiler (nvpp) to help others diagnose and overcome performance issues in their own applications. We present results for two compute systems endowed with NVIDIA V100 accelerators.",

keywords = "Accelerator, Fortran, GPU, Implicit CFD, OpenMP",

author = "Ioannis Nompelis and Gabriele Jost and Alice Koniges and Christopher Daley and David Eder and Christopher Stone",

note = "Publisher Copyright: {\textcopyright} Springer Nature Switzerland AG 2020.; 35th International Conference on High Performance Computing, ISC High Performance 2020 ; Conference date: 22-06-2020 Through 25-06-2020",

year = "2020",

doi = "10.1007/978-3-030-50743-5_14",

language = "English (US)",

isbn = "9783030507428",

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

publisher = "Springer",

pages = "270--287",

editor = "Ponnuswamy Sadayappan and Chamberlain, {Bradford L.} and Guido Juckeland and Hatem Ltaief",

booktitle = "High Performance Computing - 35th International Conference, ISC High Performance 2020, Proceedings",

}

TY - GEN

T1 - Enabling execution of a legacy CFD mini application on accelerators using OpenMP

AU - Nompelis, Ioannis

AU - Jost, Gabriele

AU - Koniges, Alice

AU - Daley, Christopher

AU - Eder, David

AU - Stone, Christopher

N1 - Publisher Copyright: © Springer Nature Switzerland AG 2020.

PY - 2020

Y1 - 2020

N2 - We describe the process and outcome of our efforts to port a legacy Fortran benchmark code to heterogeneous GPU-accelerated computing architectures using OpenMP. The benchmark code is one of the multi-zone NAS Parallel Benchmarks (NPB-MZ) called SP-MZ. This “mini-app” mimics the computation and data movement that is found in popular legacy and modern implicit computational fluid dynamics (CFD) solvers. Our objective was to examine how efficiently legacy Fortran codes can be ported to accelerators by leveraging OpenMP directives. We describe the development and optimization process and demonstrate the performance impact of various code modifications. We show select profiling results from the NVIDIA Visual Profiler (nvpp) to help others diagnose and overcome performance issues in their own applications. We present results for two compute systems endowed with NVIDIA V100 accelerators.

AB - We describe the process and outcome of our efforts to port a legacy Fortran benchmark code to heterogeneous GPU-accelerated computing architectures using OpenMP. The benchmark code is one of the multi-zone NAS Parallel Benchmarks (NPB-MZ) called SP-MZ. This “mini-app” mimics the computation and data movement that is found in popular legacy and modern implicit computational fluid dynamics (CFD) solvers. Our objective was to examine how efficiently legacy Fortran codes can be ported to accelerators by leveraging OpenMP directives. We describe the development and optimization process and demonstrate the performance impact of various code modifications. We show select profiling results from the NVIDIA Visual Profiler (nvpp) to help others diagnose and overcome performance issues in their own applications. We present results for two compute systems endowed with NVIDIA V100 accelerators.

KW - Accelerator

KW - Fortran

KW - GPU

KW - Implicit CFD

KW - OpenMP

UR - http://www.scopus.com/inward/record.url?scp=85087016723&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85087016723&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-50743-5_14

DO - 10.1007/978-3-030-50743-5_14

M3 - Conference contribution

AN - SCOPUS:85087016723

SN - 9783030507428

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

SP - 270

EP - 287

BT - High Performance Computing - 35th International Conference, ISC High Performance 2020, Proceedings

A2 - Sadayappan, Ponnuswamy

A2 - Chamberlain, Bradford L.

A2 - Juckeland, Guido

A2 - Ltaief, Hatem

PB - Springer

T2 - 35th International Conference on High Performance Computing, ISC High Performance 2020

Y2 - 22 June 2020 through 25 June 2020

ER -

Enabling execution of a legacy CFD mini application on accelerators using OpenMP

Abstract

Publication series

Conference

Bibliographical note

Keywords

Publisher link

Find It

Other files and links

Fingerprint

Cite this