A Neuromorphic Computing Platform with Compact Neuromorphic Core
A compact and versatile neuromorphic computing core that integrates 1K neurons and 1M synapses through 588 LUTs is built through the use of neuron multiplexing technology and weight clustering algorithm.
Abstract
With the rapid development of brain-like computing, large-scale neural computing platforms have received much attention. In order to reduce hardware overhead and build a large-scale neural computing platform, this work proposes a compact neuromorphic core model. Through the use of neuron multiplexing technology and weight clustering algorithm, we built a compact and versatile neuromorphic computing core that integrates 1K neurons and 1M synapses through 588 LUTs. Based on the core design, we propose a large-scale neuromorphic system. This neuromorphic computing platform integrates 64 neuromorphic cores and related control components. We successfully deployed this platform on Xilinx’s FPGA-Vertex-6 platform. We successfully deployed a three-layered spiking neural network (SNN) for image recognition and achieved 98.41% recognition accuracy through this platform.