Neural Network Exemplar Parallelization with Go

TL;DR

This paper demonstrates that exemplar parallelism of neural networks implemented in Go can significantly accelerate processing even on limited hardware like notebooks and single-board computers, achieving over 300% speedup with multiple goroutines.

Contribution

It introduces a method for neural network parallelization using Go goroutines on limited hardware, showing substantial speedup in MNIST digit recognition tasks.

Findings

Notebooks achieved 252% speedup with 4 goroutines.

Single-board computer achieved 320% with 4 goroutines.

Single-board computer achieved 432% with 8 goroutines.

Abstract

This paper presents a case for exemplar parallelism of neural networks using Go as parallelization framework. Further it is shown that also limited multi-core hardware systems are feasible for these parallelization tasks, as notebooks and single board computer systems. The main question was how much speedup can be generated when using concurrent Go goroutines specifically. A simple concurrent feedforward network for MNIST digit recognition with the programming language Go was created to find the answer. The first findings when using a notebook (Lenovo Yoga 2) showed a speedup of 252% when utilizing 4 goroutines. Testing a single board computer (Banana Pi M3) delivered more convincing results: 320% with 4 goroutines, and 432% with 8 goroutines.