Fast Algorithms for Spiking Neural Network Simulation with FPGAs

TL;DR

This paper presents FPGA-based spiking neural network simulators that outperform GPU-based methods in speed and energy efficiency, achieving real-time simulation of complex cortical microcircuits with novel high-level synthesis techniques.

Contribution

The paper introduces the first FPGA-based simulator for the Potjans-Diesmann cortical microcircuit, demonstrating significant speed and energy efficiency improvements over existing GPU-based solutions.

Findings

Simulates the circuit 25% faster than real-time

Requires less than 21 nJ per synaptic event

Outperforms state-of-the-art GPU simulators in speed and energy efficiency

Abstract

Using OpenCL-based high-level synthesis, we create a number of spiking neural network (SNN) simulators for the Potjans-Diesmann cortical microcircuit for a high-end Field-Programmable Gate Array (FPGA). Our best simulators simulate the circuit 25\% faster than real-time, require less than 21 nJ per synaptic event, and are bottle-necked by the device's on-chip memory. Speed-wise they compare favorably to the state-of-the-art GPU-based simulators and their energy usage is lower than any other published result. This result is the first for simulating the circuit on a single hardware accelerator. We also extensively analyze the techniques and algorithms we implement our simulators with, many of which can be realized on other types of hardware. Thus, this article is of interest to any researcher or practitioner interested in efficient SNN simulation, whether they target FPGAs or not.