Scaling Analog Photonic Accelerators for Byte-Size, Integer General Matrix Multiply (GEMM) Kernels

TL;DR

This paper introduces SPOGA, a scalable analog photonic accelerator that supports byte-size integer GEMM kernels, significantly improving speed and energy efficiency for DNN training compared to existing photonic solutions.

Contribution

The paper presents SPOGA, a novel photonic GEMM accelerator that overcomes previous bit-width limitations, enabling efficient byte-size integer operations for neural network acceleration.

Findings

Achieves up to 14.4× FPS improvement

Doubles energy efficiency (FPS/Watt)

Increases spatial efficiency (FPS/Watt/mm²) by 28.5×

Abstract

Deep Neural Networks (DNNs) predominantly rely on General Matrix Multiply (GEMM) kernels, which are often accelerated using specialized hardware architectures. Recently, analog photonic GEMM accelerators have emerged as a promising alternative, offering vastly superior speed and energy efficiency compared to traditional electronic accelerators. However, these photonic cannot support wider than 4-bit integer operands due to their inherent trade-offs between analog dynamic range and parallelism. This is often inadequate for DNN training as at least 8-bit wide operands are deemed necessary to prevent significant accuracy drops. To address these limitations, we introduce a scalable photonic GEMM accelerator named SPOGA. SPOGA utilizes enhanced features such as analog summation of homodyne optical signals and in-transduction positional weighting of operands. By employing an extended optical-analog dataflow that minimizes overheads associated with bit-sliced integer arithmetic, SPOGA supports byte-size integer GEMM kernels, achieving significant improvements in throughput, latency, and energy efficiency. Specifically, SPOGA demonstrates up to 14.4$\times$, 2$\times$, and 28.5$\times$ improvements in frames-per-second (FPS), FPS/Watt, and FPS/Watt/mm$^2$ respectively, compared to existing state-of-the-art photonic solutions.