Redundancy Analysis Optimization with Clustered Known Solutions for High Speed Repair

As the probability of fault occurrence increases with the advance of memory density and capacity, redundancy analysis (RA) is widely used for memory yield. However, the conventional RAs progress unnecessary solution search stages since they are applied per memory bank. It results in increase of the repair time. In this paper, redundancy analysis optimization with clustered known solutions (ROCK) is proposed for high speed repair. It progresses repair solution search considering multiple memory banks. During the repair solution search, RA using ROCK finds duplicated solution search stages based on a fault grouping method. After then, RA using ROCK enrolls the duplicated solution search stages as library and utilizes the library instead of progressing the duplicated solution search stages. It can highly reduce the repair time without any repair rate degradation.