Hybrid memory system with configurable error thresholds and failure analysis capability
Abstract
A system and method for configuring fault tolerance in nonvolatile memory (NVM) are operative to set a first threshold value, declare one or more portions of NVM invalid based on an error criterion, track the number of declared invalid NVM portions, determine if the tracked number exceeds the first threshold value, and if the tracked number exceeds the first threshold value, perform one or more remediation actions, such as issue a warning or prevent backup of volatile memory data in a hybrid memory system. In the event of backup failure, an extent of the backup can still be assessed by determining the amount of erased NVM that has remained erased after the backup, or by comparing a predicted backup end point with an actual endpoint.
Claims
exact text as granted — not AI-modifiedWe claim:
1 . In a hybrid memory system having a volatile memory subsystem and a nonvolatile memory (NVM) subsystem, a method for configuring fault tolerance in the non-volatile memory subsystem, the method comprising:
setting a first threshold register with a first threshold value corresponding to a first number of correctable symbol errors in a portion of the NVM subsystem, wherein the portion of the NVM subsystem includes a first number of symbols, wherein the first number of correctable symbol errors is less than the first number of symbols, wherein a symbol is specified by a number of bits, wherein the first threshold register comprises a portion of a register set coupled to an NVM controller, and wherein the NVM controller is operable to communicate with the threshold register set and the NVM subsystem; detecting a second number of correctable symbol errors in the portion of the NVM subsystem; determining a first symbol error value based on the detected second number of correctable symbol errors; and if the first symbol error value matches or exceeds the first threshold value, performing one or more first remediation actions, wherein the first number of correctable symbol errors is mere-less than a maximum number of symbol errors that can be corrected using an error correction algorithm, such that setting the first threshold register with the threshold value corresponding to the first number of correctable symbol errors achieves a selectable error tolerance level.Join the waitlist — get patent alerts
Track US2022206905A1 — get alerts on status changes and closely related new filings.
We store only your email — no account needed. See our privacy policy.