Accession Number:

AD1043032

Title:

New Content Addressable Memory (CAM) Technologies for Big Data and Intelligent Electronics Enabled by Magneto-Electric Ternary CAM

Descriptive Note:

Technical Report,09 Aug 2016,09 Aug 2017

Corporate Author:

University of California, Los Angeles Los Angeles United States

Personal Author(s):

Report Date:

2017-12-11

Pagination or Media Count:

27.0

Abstract:

The Magneto-Electric Ternary Content Addressable Memory MeTCAM is a new associative memory based on the propagation of spikes. By removing redundant searches and only passing spikes where a match is found, this approach provides ultra-low energy search operations. To improve throughput, the in-array pipeline scheme has been developed, allowing the MeTCAM to operate at a higher frequency than its intrinsic search delay. Furthermore, the MeTCAM incorporates the state-of-the art nonvolatile memory device, the voltage-controlled magnetic tunnel junction VC-MTJ, which not only reduces cell area thus achieving higher density but also eliminates standby energy. This report covers the technical details of the MeTCAM concept and supports the discussions by presenting simulation and analysis results that were developed under a DARPA funded effort. During the past year, advancements have been made on the design, verification, and implementation of the MeTCAM cell as well as the physical VC-MTJ device. Design and implementation challenges of the MeTCAM array are investigated and specific design guidelines are provided. Variations of the cell design are presented and evaluated. The results indicated a potential 90x improvement in the energy efficiency and a 50x improvement in the energy efficiency-throughput merit. Physical layout of the MeTCAM cell displays an area reduction of approximate 3x as compared to the conventional TCAM.

Subject Categories:

  • Computer Hardware
  • Electrical and Electronic Equipment

Distribution Statement:

APPROVED FOR PUBLIC RELEASE