Logo image
Back
A Design Flow for Mapping Spiking Neural Networks to Many-Core Neuromorphic Hardware
Conference proceeding   Open access

A Design Flow for Mapping Spiking Neural Networks to Many-Core Neuromorphic Hardware

Shihao Song, M. Lakshmi Varshika, Anup Das, Nagarajan Kandasamy and IEEE
2021 IEEE/ACM International Conference On Computer Aided Design (ICCAD)
01 Nov 2021
DOI: https://doi.org/10.1109/ICCAD51958.2021.9643500
url
http://arxiv.org/abs/2108.12444View
Submitted Open Open Access (License Unspecified)

Abstract

Computational modeling dataflow design-space exploration (DSE) Hardware Machine learning Neural networks neuromorphic computing Neuromorphics oxide-based resistive random access memory (OxRRAM) Space exploration spiking neural network (SNN) Throughput
The design of many-core neuromorphic hardware is becoming increasingly complex as these systems are now expected to execute large machine-learning models. A predictable design flow is needed to guarantee real-time performance such as latency and throughput without significantly increasing the buffer requirement of computing cores. Synchronous Data Flow Graphs (SDFGs) have been previously used for predictable mapping of streaming applications to multiprocessor systems. We propose an SDFG-based design flow to map spiking neural networks (SNNs) to many-core neuromorphic hardware with the objective of exploring the tradeoff between throughput and buffer-size requirements. The proposed design flow integrates an iterative partitioning approach based on Kernighan-Lin graph partitioning heuristic to create SNN clusters such that each cluster can be mapped to a core of the hardware. The partitioning approach minimizes inter-cluster spike communication, which improves latency on the shared interconnect of the hardware. Next, the design flow maps clusters to cores using Particle Swarm Optimization (PSO), an evolutionary algorithm, while exploring the design space of throughput and buffer size. Pareto-optimal mappings are retained from the design flow, allowing system designers to select a Pareto mapping that satisfies throughput and buffer-size requirements of the design. We evaluated the developed design flow using five large-scale convolutional neural network (CNN) models. Results demonstrate 63% higher maximum throughput and 10% lower buffer-size requirement compared to state-of-the-art dataflow-based mapping solutions.

Metrics

28 Record Views
13 citations in Web of Science
31 citations in Scopus

Details

InCites Highlights

Data related to this publication, from InCites Benchmarking & Analytics tool:

Web of Science research areas
Engineering, Manufacturing
Logo image