Fast Switching Serial and Parallel Paradigms of SNN Inference on Multi-core Heterogeneous Neuromorphic Platform SpiNNaker2

TL;DR

This paper introduces a fast-switching compilation system for SNN inference on SpiNNaker2, combining serial and parallel paradigms with a classifier to optimize performance and resource usage.

Contribution

It presents the first system that efficiently switches between serial and parallel SNN inference paradigms using a trained classifier, improving speed and resource efficiency.

Findings

Classifier achieves 91.69% accuracy in paradigm prediction

System reduces memory and processor usage compared to individual paradigms

First implementation of fast-switching SNN compilation system

Abstract

With serial and parallel processors introduced into Spiking Neural Networks (SNNs) execution, more and more researchers are dedicated to improving the performance of the computing paradigms by taking full advantage of the strengths of the available processor. In this paper, we compare and integrate serial and parallel paradigms into one SNN compiling system. For a faster switching between them in the layer granularity, we train the classifier to prejudge a better paradigm before compiling instead of making the decision afterward, saving a great amount of compiling time and RAM space on the host PC. The classifier Adaptive Boost, with the highest accuracy (91.69%) among 12 classifiers, is integrated into the switching system, which utilizes less memory and processors on the multi-core neuromorphic hardware backend SpiNNaker2 than two individual paradigms. To the best of our knowledge, it is the first fast-switching compiling system for SNN simulation.