Evaluation of autoparallelization toolkits for commodity GPUs

David Williams; Valeriu Codreanu; Po Yang; Baoquan Liu; Feng Dong; Burhan Yasar; Babak Mahdian; Alessandro Chiarini; Xia Zhao; Jos B.T.M. Roerdink

doi:10.1007/978-3-642-55224-3_42

Evaluation of autoparallelization toolkits for commodity GPUs

David Williams
, Valeriu Codreanu
, Po Yang
, Baoquan Liu
, Feng Dong
, Burhan Yasar
, Babak Mahdian
, Alessandro Chiarini
, Xia Zhao
, Jos B.T.M. Roerdink

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

4 Citations (Scopus)

Abstract

In this paper we evaluate the performance of the OpenACC and Mint toolkits against C and CUDA implementations of the standard PolyBench test suite. Our analysis reveals that performance is similar in many cases, but that a certain set of code constructs impede the ability of Mint to generate optimal code. We then present some small improvements which we integrate into our own GPSME toolkit (which is derived from Mint) and show that our toolkit now out-performs OpenACC in the majority of tests.

Original language	English
Title of host publication	Parallel Processing and Applied Mathematics: 10th International Conference, PPAM 2013, Warsaw, Poland, September 8-11, 2013, Revised Selected Papers, Part I
Publisher	Springer
Pages	447-457
ISBN (Print)	9783642552236
DOIs	https://doi.org/10.1007/978-3-642-55224-3_42
Publication status	Published - 31 Dec 2014

Keywords

GPU computing
Autoparallelization

Access to Document

10.1007/978-3-642-55224-3_42

https://link.springer.com/chapter/10.1007/978-3-642-55224-3_42

Handle.net

Persistent link

Cite this

Williams, D., Codreanu, V., Yang, P., Liu, B., Dong, F., Yasar, B., Mahdian, B., Chiarini, A., Zhao, X., & Roerdink, J. B. T. M. (2014). Evaluation of autoparallelization toolkits for commodity GPUs. In Parallel Processing and Applied Mathematics: 10th International Conference, PPAM 2013, Warsaw, Poland, September 8-11, 2013, Revised Selected Papers, Part I (pp. 447-457). Springer. https://doi.org/10.1007/978-3-642-55224-3_42

@inbook{e7fca53358d54c3dafef4e8e9a0bd3ae,

title = "Evaluation of autoparallelization toolkits for commodity GPUs",

abstract = "In this paper we evaluate the performance of the OpenACC and Mint toolkits against C and CUDA implementations of the standard PolyBench test suite. Our analysis reveals that performance is similar in many cases, but that a certain set of code constructs impede the ability of Mint to generate optimal code. We then present some small improvements which we integrate into our own GPSME toolkit (which is derived from Mint) and show that our toolkit now out-performs OpenACC in the majority of tests.",

keywords = "GPU computing, Autoparallelization",

author = "David Williams and Valeriu Codreanu and Po Yang and Baoquan Liu and Feng Dong and Burhan Yasar and Babak Mahdian and Alessandro Chiarini and Xia Zhao and Roerdink, \{Jos B.T.M.\}",

year = "2014",

month = dec,

day = "31",

doi = "10.1007/978-3-642-55224-3\_42",

language = "English",

isbn = "9783642552236",

pages = "447--457",

booktitle = "Parallel Processing and Applied Mathematics: 10th International Conference, PPAM 2013, Warsaw, Poland, September 8-11, 2013, Revised Selected Papers, Part I",

publisher = "Springer",

address = "Germany",

}

Williams, D, Codreanu, V, Yang, P, Liu, B, Dong, F, Yasar, B, Mahdian, B, Chiarini, A, Zhao, X & Roerdink, JBTM 2014, Evaluation of autoparallelization toolkits for commodity GPUs. in Parallel Processing and Applied Mathematics: 10th International Conference, PPAM 2013, Warsaw, Poland, September 8-11, 2013, Revised Selected Papers, Part I. Springer, pp. 447-457. https://doi.org/10.1007/978-3-642-55224-3_42

Evaluation of autoparallelization toolkits for commodity GPUs. / Williams, David; Codreanu, Valeriu; Yang, Po et al.
Parallel Processing and Applied Mathematics: 10th International Conference, PPAM 2013, Warsaw, Poland, September 8-11, 2013, Revised Selected Papers, Part I. Springer, 2014. p. 447-457.

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

TY - CHAP

T1 - Evaluation of autoparallelization toolkits for commodity GPUs

AU - Williams, David

AU - Codreanu, Valeriu

AU - Yang, Po

AU - Liu, Baoquan

AU - Dong, Feng

AU - Yasar, Burhan

AU - Mahdian, Babak

AU - Chiarini, Alessandro

AU - Zhao, Xia

AU - Roerdink, Jos B.T.M.

PY - 2014/12/31

Y1 - 2014/12/31

N2 - In this paper we evaluate the performance of the OpenACC and Mint toolkits against C and CUDA implementations of the standard PolyBench test suite. Our analysis reveals that performance is similar in many cases, but that a certain set of code constructs impede the ability of Mint to generate optimal code. We then present some small improvements which we integrate into our own GPSME toolkit (which is derived from Mint) and show that our toolkit now out-performs OpenACC in the majority of tests.

AB - In this paper we evaluate the performance of the OpenACC and Mint toolkits against C and CUDA implementations of the standard PolyBench test suite. Our analysis reveals that performance is similar in many cases, but that a certain set of code constructs impede the ability of Mint to generate optimal code. We then present some small improvements which we integrate into our own GPSME toolkit (which is derived from Mint) and show that our toolkit now out-performs OpenACC in the majority of tests.

KW - GPU computing

KW - Autoparallelization

U2 - 10.1007/978-3-642-55224-3_42

DO - 10.1007/978-3-642-55224-3_42

M3 - Chapter

SN - 9783642552236

SP - 447

EP - 457

BT - Parallel Processing and Applied Mathematics: 10th International Conference, PPAM 2013, Warsaw, Poland, September 8-11, 2013, Revised Selected Papers, Part I

PB - Springer

ER -

Evaluation of autoparallelization toolkits for commodity GPUs

Abstract

Keywords

Access to Document

Handle.net

Fingerprint

Cite this