A Novel FPGA Accelerator Design for Real-Time and Ultra-Low Power Deep Convolutional Neural Networks Compared With Titan X GPU
| Autor: | Nandakishor Yadav, Yukui Luo, Shuai Li, Kyuwon Ken Choi, Kuangyuan Sun |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2020 |
| Předmět: |
General Computer Science
Computer science 02 engineering and technology Convolutional neural network pipeline architecture 0202 electrical engineering electronic engineering information engineering General Materials Science Field-programmable gate array FPGA Deep neural network accelerator business.industry parallel computing Deep learning General Engineering 020206 networking & telecommunications object detection Object detection Data flow diagram Kernel (image processing) 020201 artificial intelligence & image processing Artificial intelligence lcsh:Electrical engineering. Electronics. Nuclear engineering business mixed fixed-point Electrical efficiency lcsh:TK1-9971 Computer hardware |
| Zdroj: | IEEE Access, Vol 8, Pp 105455-105471 (2020) |
| ISSN: | 2169-3536 |
| Popis: | Convolutional neural networks (CNNs) based deep learning algorithms require high data flow and computational intensity. For real-time industrial applications, they need to overcome challenges such as high data bandwidth requirement and power consumption on hardware platforms. In this work, we have analyzed in detail the data dependency in the CNN accelerator and propose specific pipelined operations and data organized manner to design a high throughput CNN accelerator on FPGA. Besides, we have optimized the kernel operations to obtain a high power efficiency. The proposed CNN accelerator supports image classification and real-time object detection with high accuracy. The evaluation results show that our CNN-based FPGA accelerator can achieve 740 Giga operations per second (GOPS) at 200 MHz with kernel power of 12.2 watts on Intel Arria 10 FPGA. For object detection tasks, our system can achieve 105 fps with 56.5 mAP or 25 fps with 73.6 mAP on VOC dataset. Since we use the mixed fixed-point data representation, the detection accuracy is comparable with the GPU-based YOLO V2 framework. The power efficiency of our system is 3.3× better than Titan X GPU and 418× better than Intel E5-2620 V4 CPU. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|