RRAM-enabled AI Accelerator Architecture

2021 IEEE International Electron Devices Meeting (IEDM)（2021）

引用 1|浏览9

暂无评分

摘要

Resistive random-access memory (RRAM) offers high-density non-volatile storage and potential for efficient in-memory computing (IMC). RRAM-enabled accelerators can solve the von Neumann bottleneck and meet the ever-growing computing needs of applications such as Artificial Intelligence (AI). In this paper, we discuss progress and challenges in RRAM-based accelerators for AI inference, training and...

查看译文

关键词

Training,Performance evaluation,Nonvolatile memory,Heuristic algorithms,Neural networks,Memory management,Systems architecture

AI 理解论文

溯源树

样例

生成溯源树，研究论文发展脉络

Chat Paper

正在生成论文摘要