An HSAIL ISA conformed GPU platform

Heng Yi Chen; Chung Ho Chen; Yun Chi Huang; Kuan Chieh Hsu; Chen Chieh Wang

doi:10.1201/b21811-52

An HSAIL ISA conformed GPU platform

Heng Yi Chen, Chung Ho Chen, Yun Chi Huang, Kuan Chieh Hsu, Chen Chieh Wang

Institute of Computer and Communication Engineering

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

Abstract

This paper presents a GPU platform based on the HSA (Heterogeneous System Architecture). The platform contains the OpenCL programming interface, a cycle-level HSAIL GPU simulator, and a finalizer that is able to translate the HSAIL code to our custom binary ISA. The system platform is able to run OpenCL applications, and the results are verified with the real GPU hardware. To improve the scheduling efficiency, we explore warp scheduling policies on our platform including Fair Round- Robin, Loose Round- Robin, Two-Level, and Greedy-Then-Oldest. We found that these algorithms are not optimal in using the load/store units. We propose a Memory Access First (MAF) mechanism which can be adopted to the Two- Level and Greedy-Then-Oldest scheduling policies to reduce the idle time of the load/store units. With the proposed mechanism, the warp scheduler can remove 23 to 34 percent of load/store unit idle time and improve the GPU performance by about 16%.

Original language	English
Title of host publication	Applied System Innovation - Proceedings of the International Conference on Applied System Innovation, ICASI 2015
Editors	Teen-Hang Meen, Stephen D. Prior, Artde Donald Kin-Tak Lam
Publisher	CRC Press/Balkema
Pages	243-247
Number of pages	5
ISBN (Print)	9781138028937
DOIs	https://doi.org/10.1201/b21811-52
Publication status	Published - 2016
Event	International Conference on Applied System Innovation, ICASI 2015 - Osaka, Japan Duration: 2015 May 22 → 2015 May 27

Publication series

Name	Applied System Innovation - Proceedings of the International Conference on Applied System Innovation, ICASI 2015

Other

Other	International Conference on Applied System Innovation, ICASI 2015
Country/Territory	Japan
City	Osaka
Period	15-05-22 → 15-05-27

All Science Journal Classification (ASJC) codes

Business, Management and Accounting (miscellaneous)
Computer Networks and Communications
Computer Science Applications
Control and Optimization
Control and Systems Engineering
Social Sciences (miscellaneous)
Electrical and Electronic Engineering

Access to Document

10.1201/b21811-52

Cite this

Chen, H. Y., Chen, C. H., Huang, Y. C., Hsu, K. C., & Wang, C. C. (2016). An HSAIL ISA conformed GPU platform. In T-H. Meen, S. D. Prior, & A. D. K-T. Lam (Eds.), Applied System Innovation - Proceedings of the International Conference on Applied System Innovation, ICASI 2015 (pp. 243-247). (Applied System Innovation - Proceedings of the International Conference on Applied System Innovation, ICASI 2015). CRC Press/Balkema. https://doi.org/10.1201/b21811-52

Chen, Heng Yi ; Chen, Chung Ho ; Huang, Yun Chi ; Hsu, Kuan Chieh ; Wang, Chen Chieh. / An HSAIL ISA conformed GPU platform. Applied System Innovation - Proceedings of the International Conference on Applied System Innovation, ICASI 2015. editor / Teen-Hang Meen ; Stephen D. Prior ; Artde Donald Kin-Tak Lam. CRC Press/Balkema, 2016. pp. 243-247 (Applied System Innovation - Proceedings of the International Conference on Applied System Innovation, ICASI 2015).

@inproceedings{b20e2d2bd99f49af841e8b5e0aece20c,

title = "An HSAIL ISA conformed GPU platform",

abstract = "This paper presents a GPU platform based on the HSA (Heterogeneous System Architecture). The platform contains the OpenCL programming interface, a cycle-level HSAIL GPU simulator, and a finalizer that is able to translate the HSAIL code to our custom binary ISA. The system platform is able to run OpenCL applications, and the results are verified with the real GPU hardware. To improve the scheduling efficiency, we explore warp scheduling policies on our platform including Fair Round- Robin, Loose Round- Robin, Two-Level, and Greedy-Then-Oldest. We found that these algorithms are not optimal in using the load/store units. We propose a Memory Access First (MAF) mechanism which can be adopted to the Two- Level and Greedy-Then-Oldest scheduling policies to reduce the idle time of the load/store units. With the proposed mechanism, the warp scheduler can remove 23 to 34 percent of load/store unit idle time and improve the GPU performance by about 16%.",

author = "Chen, {Heng Yi} and Chen, {Chung Ho} and Huang, {Yun Chi} and Hsu, {Kuan Chieh} and Wang, {Chen Chieh}",

note = "Funding Information: This work was in part supported by the Ministry of Science and Technology, Taiwan, MOST 104-2220- E-006-013. Publisher Copyright: {\textcopyright} 2016 Taylor & Francis Group.; International Conference on Applied System Innovation, ICASI 2015 ; Conference date: 22-05-2015 Through 27-05-2015",

year = "2016",

doi = "10.1201/b21811-52",

language = "English",

isbn = "9781138028937",

series = "Applied System Innovation - Proceedings of the International Conference on Applied System Innovation, ICASI 2015",

publisher = "CRC Press/Balkema",

pages = "243--247",

editor = "Teen-Hang Meen and Prior, {Stephen D.} and Lam, {Artde Donald Kin-Tak}",

booktitle = "Applied System Innovation - Proceedings of the International Conference on Applied System Innovation, ICASI 2015",

}

Chen, HY, Chen, CH, Huang, YC, Hsu, KC & Wang, CC 2016, An HSAIL ISA conformed GPU platform. in T-H Meen, SD Prior & ADK-T Lam (eds), Applied System Innovation - Proceedings of the International Conference on Applied System Innovation, ICASI 2015. Applied System Innovation - Proceedings of the International Conference on Applied System Innovation, ICASI 2015, CRC Press/Balkema, pp. 243-247, International Conference on Applied System Innovation, ICASI 2015, Osaka, Japan, 15-05-22. https://doi.org/10.1201/b21811-52

An HSAIL ISA conformed GPU platform. / Chen, Heng Yi; Chen, Chung Ho; Huang, Yun Chi; Hsu, Kuan Chieh; Wang, Chen Chieh.

Applied System Innovation - Proceedings of the International Conference on Applied System Innovation, ICASI 2015. ed. / Teen-Hang Meen; Stephen D. Prior; Artde Donald Kin-Tak Lam. CRC Press/Balkema, 2016. p. 243-247 (Applied System Innovation - Proceedings of the International Conference on Applied System Innovation, ICASI 2015).

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

TY - GEN

T1 - An HSAIL ISA conformed GPU platform

AU - Chen, Heng Yi

AU - Chen, Chung Ho

AU - Huang, Yun Chi

AU - Hsu, Kuan Chieh

AU - Wang, Chen Chieh

PY - 2016

Y1 - 2016

N2 - This paper presents a GPU platform based on the HSA (Heterogeneous System Architecture). The platform contains the OpenCL programming interface, a cycle-level HSAIL GPU simulator, and a finalizer that is able to translate the HSAIL code to our custom binary ISA. The system platform is able to run OpenCL applications, and the results are verified with the real GPU hardware. To improve the scheduling efficiency, we explore warp scheduling policies on our platform including Fair Round- Robin, Loose Round- Robin, Two-Level, and Greedy-Then-Oldest. We found that these algorithms are not optimal in using the load/store units. We propose a Memory Access First (MAF) mechanism which can be adopted to the Two- Level and Greedy-Then-Oldest scheduling policies to reduce the idle time of the load/store units. With the proposed mechanism, the warp scheduler can remove 23 to 34 percent of load/store unit idle time and improve the GPU performance by about 16%.

AB - This paper presents a GPU platform based on the HSA (Heterogeneous System Architecture). The platform contains the OpenCL programming interface, a cycle-level HSAIL GPU simulator, and a finalizer that is able to translate the HSAIL code to our custom binary ISA. The system platform is able to run OpenCL applications, and the results are verified with the real GPU hardware. To improve the scheduling efficiency, we explore warp scheduling policies on our platform including Fair Round- Robin, Loose Round- Robin, Two-Level, and Greedy-Then-Oldest. We found that these algorithms are not optimal in using the load/store units. We propose a Memory Access First (MAF) mechanism which can be adopted to the Two- Level and Greedy-Then-Oldest scheduling policies to reduce the idle time of the load/store units. With the proposed mechanism, the warp scheduler can remove 23 to 34 percent of load/store unit idle time and improve the GPU performance by about 16%.

UR - http://www.scopus.com/inward/record.url?scp=85016782946&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85016782946&partnerID=8YFLogxK

U2 - 10.1201/b21811-52

DO - 10.1201/b21811-52

M3 - Conference contribution

AN - SCOPUS:85016782946

SN - 9781138028937

T3 - Applied System Innovation - Proceedings of the International Conference on Applied System Innovation, ICASI 2015

SP - 243

EP - 247

BT - Applied System Innovation - Proceedings of the International Conference on Applied System Innovation, ICASI 2015

A2 - Meen, Teen-Hang

A2 - Prior, Stephen D.

A2 - Lam, Artde Donald Kin-Tak

PB - CRC Press/Balkema

T2 - International Conference on Applied System Innovation, ICASI 2015

Y2 - 22 May 2015 through 27 May 2015

ER -

Chen HY, Chen CH, Huang YC, Hsu KC, Wang CC. An HSAIL ISA conformed GPU platform. In Meen T-H, Prior SD, Lam ADK-T, editors, Applied System Innovation - Proceedings of the International Conference on Applied System Innovation, ICASI 2015. CRC Press/Balkema. 2016. p. 243-247. (Applied System Innovation - Proceedings of the International Conference on Applied System Innovation, ICASI 2015). https://doi.org/10.1201/b21811-52

An HSAIL ISA conformed GPU platform

Abstract

Publication series

Other

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this