Abstract:We present a massively-parallel scalable multi-purpose neuromorphic engine. All existing neuromorphic hardware systems suffer from Liebig's law (that the performance of the system is limited by the component in shortest supply) as they have fixed numbers of dedicated neurons and synapses for specific types of plasticity. For any application, it is always the availability of one of these components that limits the size of the model, leaving the others unused. To overcome this problem, our engine adopts a unique novel architecture: an array of identical components, each of which can be configured as a leaky-integrate-and-fire (LIF) neuron, a learning-synapse, or an axon with trainable delay. Spike timing dependent plasticity (STDP) and spike timing dependent delay plasticity (STDDP) are the two supported learning rules. All the parameters are stored in the SRAMs such that runtime reconfiguration is supported. As a proof of concept, we have implemented a prototype system with 16 neural engines, each of which consists of 32768 (32k) components, yielding half a million components, on an entry level FPGA (Altera Cyclone V). We verified the prototype system with measurement results. To demonstrate that our neuromorphic engine is a high performance and scalable digital design, we implemented it using TSMC 28nm HPC technology. Place and route results using Cadence Innovus with a clock frequency of 2.5 GHz show that this engine achieves an excellent area efficiency of 1.68 μm2 per component: 256k (218) components in a silicon area of 650 μm × 680 μm (∼0.44 mm2, the utilization of the silicon area is 98.7%). The power consumption of this engine is 37 mW, yielding a power efficiency of 0.92 pJ per synaptic operation (SOP).

Asynchronous Neuromorphic Optimization with Lava

Parallelized Multi-Agent Bayesian Optimization in Lava

Neuromorphic Bayesian Optimization in Lava

Transductive Spiking Graph Neural Networks for Loihi

Implementing and Benchmarking the Locally Competitive Algorithm on the Loihi 2 Neuromorphic Processor

Neuromorphic on-chip reservoir computing with spiking neural network architectures

Mapping and Validating a Point Neuron Model on Intel's Neuromorphic Hardware Loihi

Efficient Video and Audio processing with Loihi 2

Fully Asynchronous Neuromorphic Perception for Mobile Robot Dodging with Loihi Chips

Benchmarking Keyword Spotting Efficiency on Neuromorphic Hardware

Asynchronous Multi-Information Source Bayesian Optimization

Exploring Liquid Neural Networks on Loihi-2

Asynchronous Complex Analytics in a Distributed Dataflow Architecture

ANAS: Asynchronous Neuromorphic Hardware Architecture Search Based on a System-Level Simulator.

Large-Batch, Iteration-Efficient Neural Bayesian Design Optimization

RANC: Reconfigurable Architecture for Neuromorphic Computing

Simplified Klinokinesis using Spiking Neural Networks for Resource-Constrained Navigation on the Neuromorphic Processor Loihi

Design Technology Co-Optimization for Neuromorphic Computing

New Paradigms for Exploiting Parallel Experiments in Bayesian Optimization

Breaking Liebig's Law: An Advanced Multipurpose Neuromorphic Engine

Neuromorphic Nearest-Neighbor Search Using Intel's Pohoiki Springs