Scholarlink

Computing in Memory (CIM) is a promising architecture that accelerates applications by eliminating the data movement between memory and processing units. Memristor-Aided Logic (MAGIC) offers massive parallelism, flexible computing, and non-volatility. We propose a novel algorithm to find the maximum value in a set of numbers, leveraging the MAGIC NOR gate and the parallel structure of memristor arrays. Our experiment result shows that the proposed algorithm is 3.2 \(\times \) faster than AritPIM for finding the maximum of 8-bit value in 512 numbers. Additionally, this algorithm can accelerate the accumulation of floating-point vector multiplication.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Novel Efficient Maximum Searching Algorithm in ReRAM Array

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Wenqing Wang,
  • Ziming Chen,
  • Quan Deng,
  • Liang Fang

摘要

Computing in Memory (CIM) is a promising architecture that accelerates applications by eliminating the data movement between memory and processing units. Memristor-Aided Logic (MAGIC) offers massive parallelism, flexible computing, and non-volatility. We propose a novel algorithm to find the maximum value in a set of numbers, leveraging the MAGIC NOR gate and the parallel structure of memristor arrays. Our experiment result shows that the proposed algorithm is 3.2 \(\times \) faster than AritPIM for finding the maximum of 8-bit value in 512 numbers. Additionally, this algorithm can accelerate the accumulation of floating-point vector multiplication.