Fan, Hongxiang;
Ferianc, Martin;
Que, Zhiqiang;
Li, He;
Liu, Shuanglong;
Niu, Xinyu;
Luk, Wayne;
(2022)
Algorithm and Hardware Co-design for Reconfigurable CNN Accelerator.
In:
2022 27th Asia and South Pacific Design Automation Conference (ASP-DAC).
IEEE: Taipei, Taiwan.
Preview |
Text
ASP_DAC_21_NAS_FPGA.pdf - Accepted Version Download (994kB) | Preview |
Abstract
Recent advances in algorithm-hardware co-design for deep neural networks (DNNs) have demonstrated their potential in automatically designing neural architectures and hardware designs. Nevertheless, it is still a challenging optimization problem due to the expensive training cost and the time-consuming hardware implementation, which makes the exploration on the vast design space of neural architecture and hardware design intractable. In this paper, we demonstrate that our proposed approach is capable of locating designs on the Pareto frontier. This capability is enabled by a novel three-phase co-design framework, with the following new features: (a) decoupling DNN training from the design space exploration of hardware architecture and neural architecture, (b) providing a hardware-friendly neural architecture space by considering hardware characteristics in constructing the search cells, (c) adopting Gaussian process to predict accuracy, latency and power consumption to avoid time-consuming synthesis and place-and-route processes. In comparison with the manually-designed ResNet101, InceptionV2 and MobileNetV2, we can achieve up to 5% higher accuracy with up to 3× speed up on the ImageNet dataset. Compared with other state-of-the-art co-design frameworks, our found network and hardware configuration can achieve 2% (~ 6% higher accuracy, 2×∼26× smaller latency and 8.5× higher energy efficiency.
Type: | Proceedings paper |
---|---|
Title: | Algorithm and Hardware Co-design for Reconfigurable CNN Accelerator |
Event: | 2022 27th Asia and South Pacific Design Automation Conference (ASP-DAC) |
Dates: | 17 Jan 2022 - 20 Jan 2022 |
Open access status: | An open access version is available from UCL Discovery |
DOI: | 10.1109/asp-dac52403.2022.9712541 |
Publisher version: | https://doi.org/10.1109/ASP-DAC52403.2022.9712541 |
Language: | English |
Additional information: | This version is the author accepted manuscript. For information on re-use, please refer to the publisher’s terms and conditions. |
Keywords: | Training, Costs, Recurrent neural networks, Manuals, Gaussian processes, Hardware, Energy efficiency |
UCL classification: | UCL > Provost and Vice Provost Offices > UCL SLASH > Faculty of Arts and Humanities UCL > Provost and Vice Provost Offices > UCL SLASH > Faculty of Arts and Humanities > Dept of Information Studies UCL > Provost and Vice Provost Offices > UCL SLASH UCL |
URI: | https://discovery.ucl.ac.uk/id/eprint/10150064 |
Archive Staff Only
View Item |