Maestro: Data orchestration and tuning for OpenCL devices

Kyle Spafford; Jeremy Meredith; Jeffrey Vetter

doi:10.1007/978-3-642-15291-7_26

Maestro: Data orchestration and tuning for OpenCL devices

Kyle Spafford, Jeremy Meredith, Jeffrey Vetter

Computer Science and Mathematics Div

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

43 Scopus citations

Abstract

As heterogeneous computing platforms become more prevalent, the programmer must account for complex memory hierarchies in addition to the difficulties of parallel programming. OpenCL is an open standard for parallel computing that helps alleviate this difficulty by providing a portable set of abstractions for device memory hierarchies. However, OpenCL requires that the programmer explicitly controls data transfer and device synchronization, two tedious and error-prone tasks. This paper introduces Maestro, an open source library for data orchestration on OpenCL devices. Maestro provides automatic data transfer, task decomposition across multiple devices, and autotuning of dynamic execution parameters for some types of problems.

Original language	English
Title of host publication	Euro-Par 2010 Parallel Processing - 16th International Euro-Par Conference, Proceedings
Pages	275-286
Number of pages	12
Edition	PART 2
DOIs	https://doi.org/10.1007/978-3-642-15291-7_26
State	Published - 2010
Event	16th International Euro-Par Conference on Parallel Processing, Euro-Par 2010 - Ischia, Italy Duration: Aug 31 2010 → Sep 3 2010

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Number	PART 2
Volume	6272 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	16th International Euro-Par Conference on Parallel Processing, Euro-Par 2010
Country/Territory	Italy
City	Ischia
Period	08/31/10 → 09/3/10

Funding

This manuscript has been authored by a contractor of the U.S. Government under Contract No. DE-AC05-00OR22725. Accordingly, the U.S. Government retains a non-exclusive, royalty-free license to publish or reproduce the published form of this contribution, or allow others to do so, for U.S. Government purposes.

Funders	Funder number
U.S. Government	DE-AC05-00OR22725

Access to Document

10.1007/978-3-642-15291-7_26

Cite this

Spafford, K., Meredith, J., & Vetter, J. (2010). Maestro: Data orchestration and tuning for OpenCL devices. In Euro-Par 2010 Parallel Processing - 16th International Euro-Par Conference, Proceedings (PART 2 ed., pp. 275-286). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6272 LNCS, No. PART 2). https://doi.org/10.1007/978-3-642-15291-7_26

Spafford, Kyle ; Meredith, Jeremy ; Vetter, Jeffrey. / Maestro : Data orchestration and tuning for OpenCL devices. Euro-Par 2010 Parallel Processing - 16th International Euro-Par Conference, Proceedings. PART 2. ed. 2010. pp. 275-286 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2).

@inproceedings{351615fe22f74170b84cf57366b2018c,

title = "Maestro: Data orchestration and tuning for OpenCL devices",

abstract = "As heterogeneous computing platforms become more prevalent, the programmer must account for complex memory hierarchies in addition to the difficulties of parallel programming. OpenCL is an open standard for parallel computing that helps alleviate this difficulty by providing a portable set of abstractions for device memory hierarchies. However, OpenCL requires that the programmer explicitly controls data transfer and device synchronization, two tedious and error-prone tasks. This paper introduces Maestro, an open source library for data orchestration on OpenCL devices. Maestro provides automatic data transfer, task decomposition across multiple devices, and autotuning of dynamic execution parameters for some types of problems.",

author = "Kyle Spafford and Jeremy Meredith and Jeffrey Vetter",

year = "2010",

doi = "10.1007/978-3-642-15291-7_26",

language = "English",

isbn = "3642152902",

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

number = "PART 2",

pages = "275--286",

booktitle = "Euro-Par 2010 Parallel Processing - 16th International Euro-Par Conference, Proceedings",

edition = "PART 2",

note = "16th International Euro-Par Conference on Parallel Processing, Euro-Par 2010 ; Conference date: 31-08-2010 Through 03-09-2010",

}

Spafford, K, Meredith, J & Vetter, J 2010, Maestro: Data orchestration and tuning for OpenCL devices. in Euro-Par 2010 Parallel Processing - 16th International Euro-Par Conference, Proceedings. PART 2 edn, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 2, vol. 6272 LNCS, pp. 275-286, 16th International Euro-Par Conference on Parallel Processing, Euro-Par 2010, Ischia, Italy, 08/31/10. https://doi.org/10.1007/978-3-642-15291-7_26

Maestro: Data orchestration and tuning for OpenCL devices. / Spafford, Kyle; Meredith, Jeremy; Vetter, Jeffrey.
Euro-Par 2010 Parallel Processing - 16th International Euro-Par Conference, Proceedings. PART 2. ed. 2010. p. 275-286 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6272 LNCS, No. PART 2).

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

TY - GEN

T1 - Maestro

T2 - 16th International Euro-Par Conference on Parallel Processing, Euro-Par 2010

AU - Spafford, Kyle

AU - Meredith, Jeremy

AU - Vetter, Jeffrey

PY - 2010

Y1 - 2010

N2 - As heterogeneous computing platforms become more prevalent, the programmer must account for complex memory hierarchies in addition to the difficulties of parallel programming. OpenCL is an open standard for parallel computing that helps alleviate this difficulty by providing a portable set of abstractions for device memory hierarchies. However, OpenCL requires that the programmer explicitly controls data transfer and device synchronization, two tedious and error-prone tasks. This paper introduces Maestro, an open source library for data orchestration on OpenCL devices. Maestro provides automatic data transfer, task decomposition across multiple devices, and autotuning of dynamic execution parameters for some types of problems.

AB - As heterogeneous computing platforms become more prevalent, the programmer must account for complex memory hierarchies in addition to the difficulties of parallel programming. OpenCL is an open standard for parallel computing that helps alleviate this difficulty by providing a portable set of abstractions for device memory hierarchies. However, OpenCL requires that the programmer explicitly controls data transfer and device synchronization, two tedious and error-prone tasks. This paper introduces Maestro, an open source library for data orchestration on OpenCL devices. Maestro provides automatic data transfer, task decomposition across multiple devices, and autotuning of dynamic execution parameters for some types of problems.

UR - http://www.scopus.com/inward/record.url?scp=78249234704&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-15291-7_26

DO - 10.1007/978-3-642-15291-7_26

M3 - Conference contribution

AN - SCOPUS:78249234704

SN - 3642152902

SN - 9783642152900

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

SP - 275

EP - 286

BT - Euro-Par 2010 Parallel Processing - 16th International Euro-Par Conference, Proceedings

Y2 - 31 August 2010 through 3 September 2010

ER -

Spafford K, Meredith J, Vetter J. Maestro: Data orchestration and tuning for OpenCL devices. In Euro-Par 2010 Parallel Processing - 16th International Euro-Par Conference, Proceedings. PART 2 ed. 2010. p. 275-286. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2). doi: 10.1007/978-3-642-15291-7_26

Maestro: Data orchestration and tuning for OpenCL devices

Abstract

Publication series

Conference

Funding

Access to Document

Other files and links

Fingerprint

Cite this