×

Auto-learning log likelihood ratio

  • US 10,291,263 B2
  • Filed: 07/24/2017
  • Issued: 05/14/2019
  • Est. Priority Date: 07/28/2016
  • Status: Active Grant
First Claim
Patent Images

1. A method comprising:

  • programming, at a nonvolatile memory controller, codewords into each of a plurality of nonvolatile memory devices in response to receiving host-requested write instructions at the nonvolatile memory controller;

    performing, at the nonvolatile memory controller, background reads of at least one codeword of the programmed codewords in a block of the nonvolatile memory devices at a default threshold voltage, at one or more threshold voltage offset that is less than the default threshold voltage and at one or more threshold voltage offset that is greater than the default threshold voltage;

    decoding, at a decoder circuit of the nonvolatile memory controller, at least one of the background reads to identify the at least one codeword;

    identifying, at the nonvolatile memory controller, a set of log likelihood ratio (LLR) values using the read results and the identified at least one codeword;

    continuing, at the nonvolatile memory controller, the performing background reads, the decoding at least one of the background reads and the identifying to identify a set of LLR values for each block of the nonvolatile memory devices;

    further continuing the performing background reads, the decoding at least one of the background reads and the identifying to identify updated sets of LLR values;

    performing, at the nonvolatile memory controller, a host-requested read of one of the nonvolatile memory devices; and

    decoding, at the nonvolatile memory controller, the results of the host-requested read using LLR values from the updated set of LLR values corresponding to the block that was read in the host-requested read.

View all claims
  • 5 Assignments
Timeline View
Assignment View
    ×
    ×