Hybrid Redundancy for Defect Tolerance in Molecular Crossbar Memory

  • Authors:
  • Shuo Wang;Jianwei Dai;Lei Wang

  • Affiliations:
  • University of Connecticut;University of Connecticut;University of Connecticut

  • Venue:
  • ACM Journal on Emerging Technologies in Computing Systems (JETC)
  • Year:
  • 2013

Quantified Score

Hi-index 0.00

Visualization

Abstract

Nano/molecular technologies have emerged as the potential fabrics for building future integrated systems. However, due to the imperfect fabrication process, these extremely scaled devices are vulnerable to a large number of defects and transient faults. Memory systems, which are the primary application targeted by these technologies, are particularly exposed to this problem due to the ultra-high integration density and elevated error sensitivity. In this article, we propose a defect-tolerant technique, referred to as hybrid redundancy allocation, for the design of molecular crossbar memory systems. By using soft redundancy (runtime exploitation of memory spatial/temporal locality) in combination with hardware redundancy (spare memory cells), the proposed technique can achieve better error management at a low cost as compared with conventional techniques. Simulation results demonstrate the significant improvement in defect tolerance, efficiency, and scalability of the proposed technique.