基于FPGA的卷积神经网络加速模块设计
梅志伟,王维东

Design of Convolutional Neural Network acceleration module based on FPGA
Zhiwei Mei,Weidong Wang
表5 与其他加速器在AlexNet卷积层加速效果对比
Table 5 The acceleration performance of convolutional layers and other accelerators in AlexNet
AlexNet加速[12][13][8]本文方案
年份2015201620172020
FPGA

Virtex7

VX485T

Stratix?V GSD8Arria10 GT 1150

Virtex7

VX485T

时钟100 MHz120 MHz270.8 MHz100 MHz
数据32 bits16 bits32 bits16 bits
DSP224019631290588(21%)
BRAM102425672360288(28%)
性能(GOPS)61.6272.4406.10113.55

DSP性能效率

(GOPS/DSP)

0.0270.0370.3150.193