用于内存计算的磁电阻记忆器的横条阵列
Seungchul Jung1, Hyungwoo Lee1, Sungmeen Myung1
1Samsung Advanced Institute of Technology, Samsung Electronics, Suwon-si, South Korea.
Nature
|January 13, 2022
概括
研究人员为低功耗的人工神经网络开发了一种新的64x64 MRAM交叉阵列. 这种内存计算方法使用阻力总和,克服了MRAM
科学领域:
- * 材料科学与工程
- * 计算机工程
- * 人工智能
背景情况:
- * 人工神经网络 (ANN) 需要高效的多重积累操作,通常使用非易失性记忆的内存计算来实现.
- * 旋转转矩磁式随机访问存储器 (MRAM) 具有实用优势,但由于电阻低和功耗高,在传统的横杆阵列中面临挑战.
- *现有的模拟内存计算方法使用电阻,相变或闪存,但MRAM集成仍然很困难.
研究的目的:
- * 开发一种使用MRAM进行模拟人工神经网络加速的低功耗高性能交叉阵列.
- * 克服传统电流总和横杆架构中MRAM低电阻的局限性.
- 在各种机器学习任务中展示MRAM交叉阵列的有效性.
主要方法:
- * 设计和制造了使用MRAM单元的64x64横杆阵列.
- * 实现了一种新的架构,用于模拟多重积累操作的阻力总和.
- * 集成了MRAM阵列与28nm互补金属氧化物半导体 (CMOS) 读取电子.
主要成果:
- 在使用双层感知器 (软件基线:95.24%) 进行修改国家标准与技术研究所 (MNIST) 数字的分类时,获得了93.23%的准确性.
- * 在测量错误的八层视觉几何组-8 (VGG-8) 神经网络模拟中证明了98.86%的准确性 (软件基线:99.28%).
- 使用单层神经网络成功实现了93.4%的面部检测准确度.
结论:
- * 开发的MRAM横杆阵列通过阻力总和有效地解决了低阻力挑战,使得高效的模拟多重积累操作成为可能.
- 基于MRAM的内存计算架构显示出低功耗,高性能的人工神经网络实现的巨大潜力.
- 这项技术为MRAM在神经形态计算应用中的实用,大规模商业化铺平了道路.
相关概念视频
MOS Capacitor
1.0K
A Metal-Oxide-Semiconductor (MOS) capacitor is a fundamental structure used extensively in semiconductor device technology, particularly in the fabrication of integrated circuits and MOSFETs (metal-oxide-semiconductor field-effect transistors). The MOS capacitor consists of three layers: a metal gate, a dielectric oxide, and a semiconductor substrate.
The metal gate is typically made from highly conductive materials such as aluminum or polysilicon. Beneath the metal gate lies a thin layer of...
The metal gate is typically made from highly conductive materials such as aluminum or polysilicon. Beneath the metal gate lies a thin layer of...
1.0K
Mnemonic Devices
199
Mnemonic devices are cognitive tools that facilitate memory retention by linking new information to familiar patterns or organizational strategies. These techniques are beneficial for remembering complex or lengthy sets of information by simplifying and structuring them in easily retrievable ways.
Acronyms
Acronyms are created by using the initial letters of a series of words to form a new word or phrase. This approach condenses complex information into a single, memorable entity. For example,...
Acronyms
Acronyms are created by using the initial letters of a series of words to form a new word or phrase. This approach condenses complex information into a single, memorable entity. For example,...
199
System of Memory
6.6K
Memory is categorized into three major systems: sensory memory, short-term memory (STM), and long-term memory (LTM). These systems differ in their capacity and the duration for which they can hold information. Sensory memory captures raw sensory input from the environment, holding it for just a few seconds or less. For example, on hearing a brief, loud sound, like a car horn honking, the sound seems to linger in the mind for a moment even after it stops. This is an instance of sensory memory...
6.6K
Non-ohmic Devices
1.2K
In most substances, the current flow is proportional to the voltage applied to it. A simple relationship between the values of current, voltage, and resistance is known as Ohm's law. Nonohmic devices do not exhibit a linear relationship between voltage and current. One such device is the semiconducting circuit element known as a diode. A diode is a circuit device that allows current flow in only one direction.
Consider a simple circuit consisting of a battery, a diode, and a resistor. A...
Consider a simple circuit consisting of a battery, a diode, and a resistor. A...
1.2K


