video: Broadly speaking, the goal of neuromorphic engineering is to build computer systems that mimic the brain. Spiking Neural Network (SNN) is a type of biologically-inspired neural networks that perform information processing based on discrete-time spikes instead of floating point or integer numbers as in traditional Artificial Neural Network (ANN). Hardware implementation of SNNs is necessary for achieving high-performance and low-power. Recently, scientists from Zhejiang University and hangzhou dianzi university developed the Darwin Neural Processing Unit (NPU), a neuromorphic hardware co-processor for SNN based on digital logic implementation. It supports a maximum of 2048 neurons, 15 different synaptic delays and 20482 = 4,194,304 synapses. The Darwin NPU was fabricated by standard180nm CMOS technology with area size of 5x5 mm2 and 70MHz clock frequency at the worst case. It consumes 0.84mW/MHz with 1.8 V power supply for typical applications. Two applications are built to demonstrate the configurability and efficiency of the hardware implementation. The relevant paper is going to publish in SCIENCE CHINA Information Sciences 2016(2) issue. view more
Credit: ©Science China Press