Compile-time GPU memory access optimizations

G.J.W. Braak, van den; B. Mesman; H. Corporaal

doi:10.1109/ICSAMOS.2010.5642066

Compile-time GPU memory access optimizations

G.J.W. Braak, van den, B. Mesman, H. Corporaal

Electronic Systems

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

17 Citations (Scopus)

272 Downloads (Pure)

Abstract

In the last three years, GPUs are more and more being used for general purpose applications instead of only for computer graphics. Programming these GPUs is a big challenge; in current GPUs the main bottleneck for many applications is not the computing power, but the memory access bandwidth. Two compile-time optimizations are presented in this paper to deal with the two most important memory access issues. To describe these optimizations, a new notation of the parallel execution of GPU programs is introduced. An implementation of the optimizations shows that performance improvements of up to 40 times are possible.

Original language	English
Title of host publication	Proceedings of the 2010 International Conference on Embedded Computer Systems (SAMOS), 19-22 July , 2010, Samos Greece
Editors	F.J. Kurdahi, J. Takala
Place of Publication	Piscataway
Publisher	Institute of Electrical and Electronics Engineers
Pages	200-207
ISBN (Print)	978-1-4244-7937-5
DOIs	https://doi.org/10.1109/ICSAMOS.2010.5642066
Publication status	Published - 2010

Access to Document

10.1109/ICSAMOS.2010.5642066

Metis244490Final published version, 411 KB

Cite this

Braak, van den, G. J. W., Mesman, B., & Corporaal, H. (2010). Compile-time GPU memory access optimizations. In F. J. Kurdahi, & J. Takala (Eds.), Proceedings of the 2010 International Conference on Embedded Computer Systems (SAMOS), 19-22 July , 2010, Samos Greece (pp. 200-207). Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/ICSAMOS.2010.5642066

@inproceedings{f1aca295da2846da8af23b2bb05ab170,

title = "Compile-time GPU memory access optimizations",

abstract = "In the last three years, GPUs are more and more being used for general purpose applications instead of only for computer graphics. Programming these GPUs is a big challenge; in current GPUs the main bottleneck for many applications is not the computing power, but the memory access bandwidth. Two compile-time optimizations are presented in this paper to deal with the two most important memory access issues. To describe these optimizations, a new notation of the parallel execution of GPU programs is introduced. An implementation of the optimizations shows that performance improvements of up to 40 times are possible.",

author = "{Braak, van den}, G.J.W. and B. Mesman and H. Corporaal",

year = "2010",

doi = "10.1109/ICSAMOS.2010.5642066",

language = "English",

isbn = "978-1-4244-7937-5",

pages = "200--207",

editor = "F.J. Kurdahi and J. Takala",

booktitle = "Proceedings of the 2010 International Conference on Embedded Computer Systems (SAMOS), 19-22 July , 2010, Samos Greece",

publisher = "Institute of Electrical and Electronics Engineers",

address = "United States",

}

Braak, van den, GJW, Mesman, B & Corporaal, H 2010, Compile-time GPU memory access optimizations. in FJ Kurdahi & J Takala (eds), Proceedings of the 2010 International Conference on Embedded Computer Systems (SAMOS), 19-22 July , 2010, Samos Greece. Institute of Electrical and Electronics Engineers, Piscataway, pp. 200-207. https://doi.org/10.1109/ICSAMOS.2010.5642066

Compile-time GPU memory access optimizations. / Braak, van den, G.J.W.; Mesman, B.; Corporaal, H.
Proceedings of the 2010 International Conference on Embedded Computer Systems (SAMOS), 19-22 July , 2010, Samos Greece. ed. / F.J. Kurdahi; J. Takala. Piscataway: Institute of Electrical and Electronics Engineers, 2010. p. 200-207.

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

TY - GEN

T1 - Compile-time GPU memory access optimizations

AU - Braak, van den, G.J.W.

AU - Mesman, B.

AU - Corporaal, H.

PY - 2010

Y1 - 2010

N2 - In the last three years, GPUs are more and more being used for general purpose applications instead of only for computer graphics. Programming these GPUs is a big challenge; in current GPUs the main bottleneck for many applications is not the computing power, but the memory access bandwidth. Two compile-time optimizations are presented in this paper to deal with the two most important memory access issues. To describe these optimizations, a new notation of the parallel execution of GPU programs is introduced. An implementation of the optimizations shows that performance improvements of up to 40 times are possible.

AB - In the last three years, GPUs are more and more being used for general purpose applications instead of only for computer graphics. Programming these GPUs is a big challenge; in current GPUs the main bottleneck for many applications is not the computing power, but the memory access bandwidth. Two compile-time optimizations are presented in this paper to deal with the two most important memory access issues. To describe these optimizations, a new notation of the parallel execution of GPU programs is introduced. An implementation of the optimizations shows that performance improvements of up to 40 times are possible.

U2 - 10.1109/ICSAMOS.2010.5642066

DO - 10.1109/ICSAMOS.2010.5642066

M3 - Conference contribution

SN - 978-1-4244-7937-5

SP - 200

EP - 207

BT - Proceedings of the 2010 International Conference on Embedded Computer Systems (SAMOS), 19-22 July , 2010, Samos Greece

A2 - Kurdahi, F.J.

A2 - Takala, J.

PB - Institute of Electrical and Electronics Engineers

CY - Piscataway

ER -

Compile-time GPU memory access optimizations

Abstract

Access to Document

Fingerprint

Cite this