Memristive devices as computational memory

Abu Sebastian, Damien Querlioz, Bipin Rajendran, Sabina Spiga

Research output: Chapter in Book/Report/Conference proceedingChapter


Traditional von Neumann computers have separate processing and memory units. This necessitates frequent shuttling of data between these units when performing computational tasks. The resulting latency and energy cost is a key challenge especially given the recent explosive growth in highly data-centric applications such as those related to artificial intelligence. In-memory computing is an emerging non-von Neumann paradigm where certain computational tasks are performed in the memory itself by exploiting the physical attributes of the memory devices. Memristive devices that store information in terms of their resistance values are particularly well suited for in-memory computing. These devices when organized within a computational memory unit can be used to perform a range of tasks from logical and arithmetic operations to stochastic computing. In this chapter we introduce this topic with an emphasis on the key physical attributes of memristive devices that facilitate in-memory computing. We also present a future outlook highlighting some of the challenges and device-level requirements.

Original languageEnglish (US)
Title of host publicationMemristive Devices for Brain-Inspired Computing
Subtitle of host publicationFrom Materials, Devices, and Circuits to Applications - Computational Memory, Deep Learning, and Spiking Neural Networks
Number of pages8
ISBN (Electronic)9780081027820
StatePublished - Jan 1 2020
Externally publishedYes

All Science Journal Classification (ASJC) codes

  • General Engineering


  • In-memory computing
  • Memristive devices


Dive into the research topics of 'Memristive devices as computational memory'. Together they form a unique fingerprint.

Cite this