PCNNA: A Photonic Convolutional Neural Network Accelerator

TL;DR

This paper introduces PCNNA, a photonic CNN accelerator leveraging silicon photonic microring weight banks and WDM to significantly speed up convolution operations, achieving over 1000x faster performance than electronic systems.

Contribution

The paper presents a novel photonic accelerator design for CNNs that exploits photonic parallelism and sparsity, demonstrating substantial speedup over electronic counterparts.

Findings

Achieves over 1000x speedup in CNN convolution operations.

Utilizes silicon photonic microring weight banks with broadcast-and-weight protocol.

Exploits WDM and sparsity for high-performance neural network acceleration.

Abstract

Convolutional Neural Networks (CNN) have been the centerpiece of many applications including but not limited to computer vision, speech processing, and Natural Language Processing (NLP). However, the computationally expensive convolution operations impose many challenges to the performance and scalability of CNNs. In parallel, photonic systems, which are traditionally employed for data communication, have enjoyed recent popularity for data processing due to their high bandwidth, low power consumption, and reconfigurability. Here we propose a Photonic Convolutional Neural Network Accelerator (PCNNA) as a proof of concept design to speedup the convolution operation for CNNs. Our design is based on the recently introduced silicon photonic microring weight banks, which use broadcast-and-weight protocol to perform Multiply And Accumulate (MAC) operation and move data through layers of a neural network. Here, we aim to exploit the synergy between the inherent parallelism of photonics in the form of Wavelength Division Multiplexing (WDM) and sparsity of connections between input feature maps and kernels in CNNs. While our full system design offers up to more than 3 orders of magnitude speedup in execution time, its optical core potentially offers more than 5 order of magnitude speedup compared to state-of-the-art electronic counterparts.