Subtractor-Based CNN Inference Accelerator
This work addresses efficiency issues in hardware accelerators for deep learning, offering a trade-off between performance and accuracy, but it is incremental as it builds on existing CNN acceleration methods.
The paper tackles the problem of high power and area costs in CNN inference accelerators by replacing multiplication operations with subtractions through weight preprocessing, achieving 32.03% power savings and 24.59% area reduction with only 0.1% accuracy loss on LeNet-5 with MNIST.
This paper presents a novel method to boost the performance of CNN inference accelerators by utilizing subtractors. The proposed CNN preprocessing accelerator relies on sorting, grouping, and rounding the weights to create combinations that allow for the replacement of one multiplication operation and addition operation by a single subtraction operation when applying convolution during inference. Given the high cost of multiplication in terms of power and area, replacing it with subtraction allows for a performance boost by reducing power and area. The proposed method allows for controlling the trade-off between performance gains and accuracy loss through increasing or decreasing the usage of subtractors. With a rounding size of 0.05 and by utilizing LeNet-5 with the MNIST dataset, the proposed design can achieve 32.03% power savings and a 24.59% reduction in area at the cost of only 0.1% in terms of accuracy loss.