Optimizing energy efficiency of CNN-based object detection with dynamic voltage and frequency scaling
On the one hand, accelerating convolution neural networks (CNNs) on FPGAs requires ever increasing high energy efficiency in the edge computing paradigm. On the other hand, unlike normal digital algorithms, CNNs maintain their high robustness even with limited timing errors. By taking advantage of t...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Chinese Institute of Electronics
2020
|
| Subjects: | |
| Online Access: | https://eprints.nottingham.ac.uk/60946/ |
| _version_ | 1848799824553443328 |
|---|---|
| author | Jiang, Weixiong Yu, Heng Zhang, Jiale Wu, Jiaxuan Luo, Shaobo Ha, Yajun |
| author_facet | Jiang, Weixiong Yu, Heng Zhang, Jiale Wu, Jiaxuan Luo, Shaobo Ha, Yajun |
| author_sort | Jiang, Weixiong |
| building | Nottingham Research Data Repository |
| collection | Online Access |
| description | On the one hand, accelerating convolution neural networks (CNNs) on FPGAs requires ever increasing high energy efficiency in the edge computing paradigm. On the other hand, unlike normal digital algorithms, CNNs maintain their high robustness even with limited timing errors. By taking advantage of this unique feature, we propose to use dynamic voltage and frequency scaling (DVFS) to further optimize the energy efficiency for CNNs. First, we have developed a DVFS framework on FPGAs. Second, we apply the DVFS to SkyNet, a state-of-the-art neural network targeting on object detection. Third, we analyze the impact of DVFS on CNNs in terms of performance, power, energy efficiency and accuracy. Compared to the state-of-the-art, experimental results show that we have achieved 38% improvement in energy efficiency without any loss in accuracy. Results also show that we can achieve 47% improvement in energy efficiency if we allow 0.11% relaxation in accuracy. |
| first_indexed | 2025-11-14T20:41:48Z |
| format | Article |
| id | nottingham-60946 |
| institution | University of Nottingham Malaysia Campus |
| institution_category | Local University |
| language | English |
| last_indexed | 2025-11-14T20:41:48Z |
| publishDate | 2020 |
| publisher | Chinese Institute of Electronics |
| recordtype | eprints |
| repository_type | Digital Repository |
| spelling | nottingham-609462020-06-22T03:09:16Z https://eprints.nottingham.ac.uk/60946/ Optimizing energy efficiency of CNN-based object detection with dynamic voltage and frequency scaling Jiang, Weixiong Yu, Heng Zhang, Jiale Wu, Jiaxuan Luo, Shaobo Ha, Yajun On the one hand, accelerating convolution neural networks (CNNs) on FPGAs requires ever increasing high energy efficiency in the edge computing paradigm. On the other hand, unlike normal digital algorithms, CNNs maintain their high robustness even with limited timing errors. By taking advantage of this unique feature, we propose to use dynamic voltage and frequency scaling (DVFS) to further optimize the energy efficiency for CNNs. First, we have developed a DVFS framework on FPGAs. Second, we apply the DVFS to SkyNet, a state-of-the-art neural network targeting on object detection. Third, we analyze the impact of DVFS on CNNs in terms of performance, power, energy efficiency and accuracy. Compared to the state-of-the-art, experimental results show that we have achieved 38% improvement in energy efficiency without any loss in accuracy. Results also show that we can achieve 47% improvement in energy efficiency if we allow 0.11% relaxation in accuracy. Chinese Institute of Electronics 2020-02-11 Article PeerReviewed application/pdf en cc_by https://eprints.nottingham.ac.uk/60946/1/heng.pdf Jiang, Weixiong, Yu, Heng, Zhang, Jiale, Wu, Jiaxuan, Luo, Shaobo and Ha, Yajun (2020) Optimizing energy efficiency of CNN-based object detection with dynamic voltage and frequency scaling. Journal of Semiconductors, 41 (2). 022406. ISSN 1674-4926 CNN; FPGA; DVFS; object detection http://dx.doi.org/10.1088/1674-4926/41/2/022406 doi:10.1088/1674-4926/41/2/022406 doi:10.1088/1674-4926/41/2/022406 |
| spellingShingle | CNN; FPGA; DVFS; object detection Jiang, Weixiong Yu, Heng Zhang, Jiale Wu, Jiaxuan Luo, Shaobo Ha, Yajun Optimizing energy efficiency of CNN-based object detection with dynamic voltage and frequency scaling |
| title | Optimizing energy efficiency of CNN-based object detection with dynamic voltage and frequency scaling |
| title_full | Optimizing energy efficiency of CNN-based object detection with dynamic voltage and frequency scaling |
| title_fullStr | Optimizing energy efficiency of CNN-based object detection with dynamic voltage and frequency scaling |
| title_full_unstemmed | Optimizing energy efficiency of CNN-based object detection with dynamic voltage and frequency scaling |
| title_short | Optimizing energy efficiency of CNN-based object detection with dynamic voltage and frequency scaling |
| title_sort | optimizing energy efficiency of cnn-based object detection with dynamic voltage and frequency scaling |
| topic | CNN; FPGA; DVFS; object detection |
| url | https://eprints.nottingham.ac.uk/60946/ https://eprints.nottingham.ac.uk/60946/ https://eprints.nottingham.ac.uk/60946/ |