https://github.com/sfmth/OpenSpike
OpenSpike 是一种尖峰神经网络 (SNN) 加速器,使用完全开源的 EDA 工具、流程设计工具包 (PDK) 和使用OpenRAM合成的内存宏。该芯片采用 130 nm SkyWater 工艺流片,集成了超过 100 万个突触权重,并提供可重新编程的架构。它以40 MHz的时钟速度运行,1.8 V的电源,使用PicoRV32内核进行控制,占用面积为33.3 mm 2. 加速器的吞吐量为每秒 48,262 张图像,挂钟时间为 20.72 μs,速度为 56.8 GOPS/W。尖峰神经元使用滞后作用来提供可以减少状态不稳定性的自适应阈值(即,施密特触发器)。这导致了一系列基准测试中的高性能 SNN 与最先进的全精度 SNN 相比仍然具有竞争力。
OpenSpike a spiking neural network (SNN) accelerator made using fully open-source EDA tools, process design kit (PDK), and memory macros synthesized using OpenRAM. The chip is taped out in the 130 nm SkyWater process and integrates over 1 million synaptic weights, and offers a reprogrammable architecture. It operates at a clock speed of 40 MHz, a supply of 1.8 V, uses a PicoRV32 core for control, and occupies an area of 33.3 mm2. The throughput of the accelerator is 48,262 images per second with a wallclock time of 20.72 μs, at 56.8 GOPS/W. The spiking neurons use hysteresis to provide an adaptive threshold (i.e., a Schmitt trigger) which can reduce state instability. This results in high performing SNNs across a range of benchmarks that remain competitive with state-of-the- art, full precision SNNs.
论文: https://arxiv.org/pdf/2302.01015.pdf
阅读原文参考原论文。
相关推荐:
code:通过进化、可塑性和 元 元学习 获得认知能力(4个时间维度的学习迭代)
脑记忆产生和巩固建模研究总结(3假设3发现3创新符合13篇脑科学实验和假设)
代码:Learning to Learn and Forget (华为)
Self-building Neural Networks 代码