Efficient Residue Number System Based Winograd Convolution

TL;DR

This paper extends the Winograd convolution algorithm to Residue Number System (RNS), enabling efficient low-precision CNN inference with reduced arithmetic complexity and significant performance gains without accuracy loss.

Contribution

It introduces an RNS-based Winograd convolution method that maintains accuracy while significantly reducing computational complexity for low-precision neural network inference.

Findings

Arithmetic complexity reduced by up to 7.03x

Performance improved by 2.30x to 4.69x

No degradation in network prediction accuracy

Abstract

Prior research has shown that Winograd algorithm can reduce the computational complexity of convolutional neural networks (CNN) with weights and activations represented in floating point. However it is difficult to apply the scheme to the inference of low-precision quantized (e.g. INT8) networks. Our work extends the Winograd algorithm to Residue Number System (RNS). The minimal complexity convolution is computed precisely over large transformation tile (e.g. 10 x 10 to 16 x 16) of filters and activation patches using the Winograd transformation and low cost (e.g. 8-bit) arithmetic without degrading the prediction accuracy of the networks during inference. The arithmetic complexity reduction is up to 7.03x while the performance improvement is up to 2.30x to 4.69x for 3 x 3 and 5 x 5 filters respectively.