Run-time performance and power optimization of parallel disparity estimation on many-core platforms
Run-time performance and power optimization of parallel disparity estimation on many-core platforms
This paper investigates the use of many-core systems to execute the disparity estimation algorithm, used in stereo vision applications, as these systems can provide flexibility between performance scaling and power consumption. We present a learning-based run-time management approach which achieves a required performance threshold whilst minimizing power consumption through dynamic control of frequency and core allocation. Experimental results are obtained from a 61-core Intel Xeon Phi platform for the above investigation. The same performance can be achieved with an average reduction in power consumption of 27.8% and increased energy efficiency by 30.04% when compared to DVFS control alone without run-time management.
Leech, Charles
6ba70c54-3792-41cd-a8d6-9e8884ae004f
Vala, Charan Kumar
41279fa4-5cb5-469b-88ae-cc43a35c4325
Acharyya, Amit
1f8a0620-1c00-4306-a64c-5185ede71f38
Yang, Sheng
04b9848f-ddd4-4d8f-93b6-b91a2144d49c
Merrett, Geoffrey
89b3a696-41de-44c3-89aa-b0aa29f54020
Al-Hashimi, Bashir
0b29c671-a6d2-459c-af68-c4614dce3b5d
Leech, Charles
6ba70c54-3792-41cd-a8d6-9e8884ae004f
Vala, Charan Kumar
41279fa4-5cb5-469b-88ae-cc43a35c4325
Acharyya, Amit
1f8a0620-1c00-4306-a64c-5185ede71f38
Yang, Sheng
04b9848f-ddd4-4d8f-93b6-b91a2144d49c
Merrett, Geoffrey
89b3a696-41de-44c3-89aa-b0aa29f54020
Al-Hashimi, Bashir
0b29c671-a6d2-459c-af68-c4614dce3b5d
Leech, Charles, Vala, Charan Kumar, Acharyya, Amit, Yang, Sheng, Merrett, Geoffrey and Al-Hashimi, Bashir
(2017)
Run-time performance and power optimization of parallel disparity estimation on many-core platforms.
ACM Transactions on Embedded Computing Systems.
(In Press)
Abstract
This paper investigates the use of many-core systems to execute the disparity estimation algorithm, used in stereo vision applications, as these systems can provide flexibility between performance scaling and power consumption. We present a learning-based run-time management approach which achieves a required performance threshold whilst minimizing power consumption through dynamic control of frequency and core allocation. Experimental results are obtained from a 61-core Intel Xeon Phi platform for the above investigation. The same performance can be achieved with an average reduction in power consumption of 27.8% and increased energy efficiency by 30.04% when compared to DVFS control alone without run-time management.
Text
acmtecs
- Accepted Manuscript
More information
Accepted/In Press date: 11 August 2017
Identifiers
Local EPrints ID: 413390
URI: http://eprints.soton.ac.uk/id/eprint/413390
ISSN: 1539-9087
PURE UUID: 8faa1ae1-d311-4db8-9989-a66eff9d1854
Catalogue record
Date deposited: 23 Aug 2017 16:31
Last modified: 16 Mar 2024 03:46
Export record
Contributors
Author:
Charles Leech
Author:
Charan Kumar Vala
Author:
Amit Acharyya
Author:
Sheng Yang
Author:
Geoffrey Merrett
Author:
Bashir Al-Hashimi
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics