Repair Analysis for Embedded Memories Using Block-Based Redundancy Architecture

Capacity and density of embedded memories have rapidly increased therefore they have higher probability of faults. As a result, yield of system-on-a-chip designs with embedded memories drops. Built-in self-repair is widely used to improve manufacturing yield by replacing faulty memory cells with redundant elements. Most approaches perform reconfiguration on the row/column level. Block-based redundancy architectures divide memory and redundancies into blocks, performing reconfiguration on the block level and offering more efficient usage of the redundant elements.