Enabling a Uniform OpenCL Device View for Heterogeneous Platforms
Summary :
Aiming to ease the parallel programming for heterogeneous architectures, we propose and implement a high-level OpenCL runtime that conceptually merges multiple heterogeneous hardware devices into one virtual heterogeneous compute device (VHCD). Moreover, automated workload distribution among the devices is based on offline profiling, together with new programming directives that define the device-independent data access range per work-group. Therefore, an OpenCL program originally written for a single compute device can, after inserting a small number of programming directives, run efficiently on a platform consisting of heterogeneous compute devices. Performance is ensured by introducing the technique of virtual cache management, which minimizes the amount of host-device data transfer. Our new OpenCL runtime is evaluated by a diverse set of OpenCL benchmarks, demonstrating good performance on various configurations of a heterogeneous system.
- Publication
- IEICE TRANSACTIONS on Information Vol.E98-D No.4 pp.812-823
- Publication Date
- 2015/04/01
- Publicized
- 2015/01/20
- Online ISSN
- 1745-1361
- DOI
- 10.1587/transinf.2014EDP7244
- Type of Manuscript
- PAPER
- Category
- Fundamentals of Information Systems
Authors
Dafei HUANG
National University of Defense Technology
Changqing XUN
National University of Defense Technology
Nan WU
National University of Defense Technology
Mei WEN
National University of Defense Technology
Chunyuan ZHANG
National University of Defense Technology
Xing CAI
University of Oslo
Qianming YANG
National University of Defense Technology
Keyword
Latest Issue
Copyrights notice of machine-translated contents
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Cite this
Copy
Dafei HUANG, Changqing XUN, Nan WU, Mei WEN, Chunyuan ZHANG, Xing CAI, Qianming YANG, "Enabling a Uniform OpenCL Device View for Heterogeneous Platforms" in IEICE TRANSACTIONS on Information,
vol. E98-D, no. 4, pp. 812-823, April 2015, doi: 10.1587/transinf.2014EDP7244.
Abstract: Aiming to ease the parallel programming for heterogeneous architectures, we propose and implement a high-level OpenCL runtime that conceptually merges multiple heterogeneous hardware devices into one virtual heterogeneous compute device (VHCD). Moreover, automated workload distribution among the devices is based on offline profiling, together with new programming directives that define the device-independent data access range per work-group. Therefore, an OpenCL program originally written for a single compute device can, after inserting a small number of programming directives, run efficiently on a platform consisting of heterogeneous compute devices. Performance is ensured by introducing the technique of virtual cache management, which minimizes the amount of host-device data transfer. Our new OpenCL runtime is evaluated by a diverse set of OpenCL benchmarks, demonstrating good performance on various configurations of a heterogeneous system.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2014EDP7244/_p
Copy
@ARTICLE{e98-d_4_812,
author={Dafei HUANG, Changqing XUN, Nan WU, Mei WEN, Chunyuan ZHANG, Xing CAI, Qianming YANG, },
journal={IEICE TRANSACTIONS on Information},
title={Enabling a Uniform OpenCL Device View for Heterogeneous Platforms},
year={2015},
volume={E98-D},
number={4},
pages={812-823},
abstract={Aiming to ease the parallel programming for heterogeneous architectures, we propose and implement a high-level OpenCL runtime that conceptually merges multiple heterogeneous hardware devices into one virtual heterogeneous compute device (VHCD). Moreover, automated workload distribution among the devices is based on offline profiling, together with new programming directives that define the device-independent data access range per work-group. Therefore, an OpenCL program originally written for a single compute device can, after inserting a small number of programming directives, run efficiently on a platform consisting of heterogeneous compute devices. Performance is ensured by introducing the technique of virtual cache management, which minimizes the amount of host-device data transfer. Our new OpenCL runtime is evaluated by a diverse set of OpenCL benchmarks, demonstrating good performance on various configurations of a heterogeneous system.},
keywords={},
doi={10.1587/transinf.2014EDP7244},
ISSN={1745-1361},
month={April},}
Copy
TY - JOUR
TI - Enabling a Uniform OpenCL Device View for Heterogeneous Platforms
T2 - IEICE TRANSACTIONS on Information
SP - 812
EP - 823
AU - Dafei HUANG
AU - Changqing XUN
AU - Nan WU
AU - Mei WEN
AU - Chunyuan ZHANG
AU - Xing CAI
AU - Qianming YANG
PY - 2015
DO - 10.1587/transinf.2014EDP7244
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E98-D
IS - 4
JA - IEICE TRANSACTIONS on Information
Y1 - April 2015
AB - Aiming to ease the parallel programming for heterogeneous architectures, we propose and implement a high-level OpenCL runtime that conceptually merges multiple heterogeneous hardware devices into one virtual heterogeneous compute device (VHCD). Moreover, automated workload distribution among the devices is based on offline profiling, together with new programming directives that define the device-independent data access range per work-group. Therefore, an OpenCL program originally written for a single compute device can, after inserting a small number of programming directives, run efficiently on a platform consisting of heterogeneous compute devices. Performance is ensured by introducing the technique of virtual cache management, which minimizes the amount of host-device data transfer. Our new OpenCL runtime is evaluated by a diverse set of OpenCL benchmarks, demonstrating good performance on various configurations of a heterogeneous system.
ER -
English
