Metric-based reactive reading to screen for fault-prone memory blocks

摘要

Various applications may include apparatus and/or methods for preemptively detecting fault-prone memory blocks in a memory device and handling these memory blocks before they fail and trigger a data loss event. Metrics based on memory operations may be used to facilitate inspection of memory blocks. One or more metrics related to memory operation for the memory block may be tracked and a Z-score may be generated for each metric. In response to the comparison of the Z-score for the metric to the Z-score for the metric threshold, actions may be performed to control possible discarding of the memory block, beginning with the comparison. Additional devices, systems and methods are disclosed.