⚡ 本页包含 AI 生成的分析内容，仅供参考
该论文提出了一种1Mb多比特ReRAM存内计算宏，用于基于CNN的AI边缘处理器。该宏实现了14.6ns的并行乘积累加计算时间，显著降低了边缘AI设备的延迟和能耗。解决了传统冯·诺依曼架构中数据搬移导致的性能瓶颈。
Wei-En Lin, Jing-Hong Wang, Wei-Chen Wei, Ting-Wei Chang, Tung-Cheng Chang, Tsung-Yuan Huang, Hui-Yao Kao, Shih-Ying Wei, Yen-Cheng Chiu, Chun-Ying Lee, Chung-Chuan Lo, Ya-Chin King, Chorng-Jung Lin, Ren-Shuo Liu, Chih-Cheng Hsieh, Kea-Tiong Tang, Meng-Fan Chang National Tsing Hua University, Hsinchu, Taiwan Embedded nonvolatile memory (NVM) and computing-in-memory (CIM) are significantly reducing the latency (tMAC) and energy consumption (EMAC) of multiplyand-accumulate (MAC) operations in artificial intelligence (AI) edge devices [1,2]. Previous ReRAM CIM macros demonstrated MAC operations for 1b-input, ternaryweighted, 3b-output CNNs [1] or 1b-input, 8b-weighted, 1b-output fully-connected networks with limited accuracy [2]. To support higher-accuracy convolution neural network heavy applications NVM-CIM should support multibit inputs/weights and multi-bit output (MAC-OUT) for CNN operations. One way to achieve multibit weights
Cheng-Xin Xue, Wei-Hao Chen, Je-Syu Liu, Jia-Fang Li, Wei-Yu Lin,