Evaluating the Performance of NVIDIA's A100 Ampere GPU for Sparse and Batched Computations

Hartwig Anzt; Yuhsiang M. Tsai; Ahmad Abdelfattah; Terry Cojean; Jack Dongarra

doi:10.1109/PMBS51919.2020.00009

Evaluating the Performance of NVIDIA's A100 Ampere GPU for Sparse and Batched Computations

Hartwig Anzt, Yuhsiang M. Tsai, Ahmad Abdelfattah, Terry Cojean, Jack Dongarra

Computer Science and Mathematics Div

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

18 Scopus citations

Abstract

GPU accelerators have become an Important backbone for scientific high performance-computing, and the performance advances obtained from adopting new GPU hardware are significant. In this paper we take a first look at NVIDIA's newest server-line GPU, the A100 architecture, part of the Ampere generation. Specifically, we assess its performance for sparse and batch computations, as these routines are relied upon in many scientific applications, and compare to the performance achieved on NVIDIA's previous server-line GPU.

Original language	English
Title of host publication	Proceedings of PMBS 2020
Subtitle of host publication	Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems, Held in conjunction with SC 2020: The International Conference for High Performance Computing, Networking, Storage and Analysis
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	26-38
Number of pages	13
ISBN (Electronic)	9781665422659
DOIs	https://doi.org/10.1109/PMBS51919.2020.00009
State	Published - Nov 2020
Event	2020 IEEE/ACM Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems, PMBS 2020 - Virtual, Atlanta, United States Duration: Nov 12 2020 → …

Publication series

Name	Proceedings of PMBS 2020: Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems, Held in conjunction with SC 2020: The International Conference for High Performance Computing, Networking, Storage and Analysis

Conference

Conference	2020 IEEE/ACM Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems, PMBS 2020
Country/Territory	United States
City	Virtual, Atlanta
Period	11/12/20 → …

Funding

ACKNOWLEDGMENTS This work was supported by the “Impuls und Vernetzungs-fond” of the Helmholtz Association under grant VH-NG-1241 and by the Exascale Computing Project (17-SC-20-SC), a collaborative effort of the U.S. Department of Energy Office of Science and the National Nuclear Security Administration. The authors would like to thank the Steinbuch Centre for Computing (SCC) of the Karlsruhe Institute of Technology for providing access to an NVIDIA A100 GPU.

Keywords

Batched Linear Algebra
NVIDIA A100 GPU
Sparse Linear Algebra
Sparse Matrix Vector Product

Access to Document

10.1109/PMBS51919.2020.00009

Cite this

Anzt, H., Tsai, Y. M., Abdelfattah, A., Cojean, T., & Dongarra, J. (2020). Evaluating the Performance of NVIDIA's A100 Ampere GPU for Sparse and Batched Computations. In Proceedings of PMBS 2020: Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems, Held in conjunction with SC 2020: The International Conference for High Performance Computing, Networking, Storage and Analysis (pp. 26-38). Article 9307857 (Proceedings of PMBS 2020: Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems, Held in conjunction with SC 2020: The International Conference for High Performance Computing, Networking, Storage and Analysis). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/PMBS51919.2020.00009

Anzt, Hartwig ; Tsai, Yuhsiang M. ; Abdelfattah, Ahmad et al. / Evaluating the Performance of NVIDIA's A100 Ampere GPU for Sparse and Batched Computations. Proceedings of PMBS 2020: Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems, Held in conjunction with SC 2020: The International Conference for High Performance Computing, Networking, Storage and Analysis. Institute of Electrical and Electronics Engineers Inc., 2020. pp. 26-38 (Proceedings of PMBS 2020: Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems, Held in conjunction with SC 2020: The International Conference for High Performance Computing, Networking, Storage and Analysis).

@inproceedings{943c92c92f7346ea937fd0425b423f48,

title = "Evaluating the Performance of NVIDIA's A100 Ampere GPU for Sparse and Batched Computations",

abstract = "GPU accelerators have become an Important backbone for scientific high performance-computing, and the performance advances obtained from adopting new GPU hardware are significant. In this paper we take a first look at NVIDIA's newest server-line GPU, the A100 architecture, part of the Ampere generation. Specifically, we assess its performance for sparse and batch computations, as these routines are relied upon in many scientific applications, and compare to the performance achieved on NVIDIA's previous server-line GPU.",

keywords = "Batched Linear Algebra, NVIDIA A100 GPU, Sparse Linear Algebra, Sparse Matrix Vector Product",

author = "Hartwig Anzt and Tsai, {Yuhsiang M.} and Ahmad Abdelfattah and Terry Cojean and Jack Dongarra",

note = "Publisher Copyright: {\textcopyright} 2020 IEEE.; 2020 IEEE/ACM Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems, PMBS 2020 ; Conference date: 12-11-2020",

year = "2020",

month = nov,

doi = "10.1109/PMBS51919.2020.00009",

language = "English",

series = "Proceedings of PMBS 2020: Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems, Held in conjunction with SC 2020: The International Conference for High Performance Computing, Networking, Storage and Analysis",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "26--38",

booktitle = "Proceedings of PMBS 2020",

}

Anzt, H, Tsai, YM, Abdelfattah, A, Cojean, T & Dongarra, J 2020, Evaluating the Performance of NVIDIA's A100 Ampere GPU for Sparse and Batched Computations. in Proceedings of PMBS 2020: Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems, Held in conjunction with SC 2020: The International Conference for High Performance Computing, Networking, Storage and Analysis., 9307857, Proceedings of PMBS 2020: Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems, Held in conjunction with SC 2020: The International Conference for High Performance Computing, Networking, Storage and Analysis, Institute of Electrical and Electronics Engineers Inc., pp. 26-38, 2020 IEEE/ACM Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems, PMBS 2020, Virtual, Atlanta, United States, 11/12/20. https://doi.org/10.1109/PMBS51919.2020.00009

Evaluating the Performance of NVIDIA's A100 Ampere GPU for Sparse and Batched Computations. / Anzt, Hartwig; Tsai, Yuhsiang M.; Abdelfattah, Ahmad et al.
Proceedings of PMBS 2020: Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems, Held in conjunction with SC 2020: The International Conference for High Performance Computing, Networking, Storage and Analysis. Institute of Electrical and Electronics Engineers Inc., 2020. p. 26-38 9307857 (Proceedings of PMBS 2020: Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems, Held in conjunction with SC 2020: The International Conference for High Performance Computing, Networking, Storage and Analysis).

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

TY - GEN

T1 - Evaluating the Performance of NVIDIA's A100 Ampere GPU for Sparse and Batched Computations

AU - Anzt, Hartwig

AU - Tsai, Yuhsiang M.

AU - Abdelfattah, Ahmad

AU - Cojean, Terry

AU - Dongarra, Jack

PY - 2020/11

Y1 - 2020/11

N2 - GPU accelerators have become an Important backbone for scientific high performance-computing, and the performance advances obtained from adopting new GPU hardware are significant. In this paper we take a first look at NVIDIA's newest server-line GPU, the A100 architecture, part of the Ampere generation. Specifically, we assess its performance for sparse and batch computations, as these routines are relied upon in many scientific applications, and compare to the performance achieved on NVIDIA's previous server-line GPU.

AB - GPU accelerators have become an Important backbone for scientific high performance-computing, and the performance advances obtained from adopting new GPU hardware are significant. In this paper we take a first look at NVIDIA's newest server-line GPU, the A100 architecture, part of the Ampere generation. Specifically, we assess its performance for sparse and batch computations, as these routines are relied upon in many scientific applications, and compare to the performance achieved on NVIDIA's previous server-line GPU.

KW - Batched Linear Algebra

KW - NVIDIA A100 GPU

KW - Sparse Linear Algebra

KW - Sparse Matrix Vector Product

UR - http://www.scopus.com/inward/record.url?scp=85099532233&partnerID=8YFLogxK

U2 - 10.1109/PMBS51919.2020.00009

DO - 10.1109/PMBS51919.2020.00009

M3 - Conference contribution

AN - SCOPUS:85099532233

T3 - Proceedings of PMBS 2020: Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems, Held in conjunction with SC 2020: The International Conference for High Performance Computing, Networking, Storage and Analysis

SP - 26

EP - 38

BT - Proceedings of PMBS 2020

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2020 IEEE/ACM Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems, PMBS 2020

Y2 - 12 November 2020

ER -

Anzt H, Tsai YM, Abdelfattah A, Cojean T, Dongarra J. Evaluating the Performance of NVIDIA's A100 Ampere GPU for Sparse and Batched Computations. In Proceedings of PMBS 2020: Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems, Held in conjunction with SC 2020: The International Conference for High Performance Computing, Networking, Storage and Analysis. Institute of Electrical and Electronics Engineers Inc. 2020. p. 26-38. 9307857. (Proceedings of PMBS 2020: Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems, Held in conjunction with SC 2020: The International Conference for High Performance Computing, Networking, Storage and Analysis). doi: 10.1109/PMBS51919.2020.00009

Evaluating the Performance of NVIDIA's A100 Ampere GPU for Sparse and Batched Computations

Abstract

Publication series

Conference

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this