An FPGA-Based Hardware Accelerator for CNNs Using On-Chip Memories Only: Design and Benchmarking with Intel Movidius Neural Compute Stick
| Autor: | Luca Fanucci, Gionata Benelli, Emilio Rapuano, Gianmarco Dinelli, Gabriele Meoni |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2019 |
| Předmět: |
Hardware architecture
lcsh:Computer engineering. Computer hardware Article Subject Computer science business.industry Deep learning lcsh:TK7885-7895 02 engineering and technology Convolutional neural network Computer Science::Hardware Architecture Computer engineering Hardware and Architecture Gate array 020204 information systems Keyword spotting 0202 electrical engineering electronic engineering information engineering Memory footprint Hardware acceleration 020201 artificial intelligence & image processing Artificial intelligence Field-programmable gate array business |
| Zdroj: | International Journal of Reconfigurable Computing, Vol 2019 (2019) |
| ISSN: | 1687-7209 |
| Popis: | During the last years, convolutional neural networks have been used for different applications, thanks to their potentiality to carry out tasks by using a reduced number of parameters when compared with other deep learning approaches. However, power consumption and memory footprint constraints, typical of on the edge and portable applications, usually collide with accuracy and latency requirements. For such reasons, commercial hardware accelerators have become popular, thanks to their architecture designed for the inference of general convolutional neural network models. Nevertheless, field-programmable gate arrays represent an interesting perspective since they offer the possibility to implement a hardware architecture tailored to a specific convolutional neural network model, with promising results in terms of latency and power consumption. In this article, we propose a full on-chip field-programmable gate array hardware accelerator for a separable convolutional neural network, which was designed for a keyword spotting application. We started from the model implemented in a previous work for the Intel Movidius Neural Compute Stick. For our goals, we appropriately quantized such a model through a bit-true simulation, and we realized a dedicated architecture exclusively using on-chip memories. A benchmark comparing the results on different field-programmable gate array families by Xilinx and Intel with the implementation on the Neural Compute Stick was realized. The analysis shows that better inference time and energy per inference results can be obtained with comparable accuracy at expenses of a higher design effort and development time through the FPGA solution. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Plný text