论文标题
使用基于模拟MRAM的神经元和突触的单周期MLP分类器
A Single-Cycle MLP Classifier Using Analog MRAM-based Neurons and Synapses
论文作者
论文摘要
在本文中，杠杆化旋转轨道扭矩（SOT）磁磁性随机访问记忆（MRAM）设备可用于实现单周期类似物中的内存计算（IMC）结构的sigmoidal神经元和二元突触。首先，提出了一个基于模拟SOT-MRAM的神经元比特电池，与以前的功率最高和面积和区域效率高的模拟sigmoidal神经元设计相比，功率区域的功率降低12倍。接下来，在存储子阵列中使用了建议的神经元和突触位细胞，以形成基于模拟IMC的多层式构建（MLP）体系结构，以用于MNIST模式识别应用。体系结构级别的结果表明，与混合信号模拟/数字IMC架构和数字GPU实现相比，我们的模拟IMC架构至少具有两个和四个数量级的性能提高，同时实现了可比的分类精度。
In this paper, spin-orbit torque (SOT) magnetoresistive random-access memory (MRAM) devices are leveraged to realize sigmoidal neurons and binarized synapses for a single-cycle analog in-memory computing (IMC) architecture. First, an analog SOT-MRAM-based neuron bitcell is proposed which achieves a 12x reduction in power-area-product compared to the previous most power- and area-efficient analog sigmoidal neuron design. Next, proposed neuron and synapse bit cells are used within memory subarrays to form an analog IMC-based multilayer perceptron (MLP) architecture for the MNIST pattern recognition application. The architecture-level results exhibit that our analog IMC architecture achieves at least two and four orders of magnitude performance improvement compared to a mixed-signal analog/digital IMC architecture and a digital GPU implementation, respectively while realizing a comparable classification accuracy.
