Abstract: The present disclosure describes apparatuses and methods for machine learning-enabled (ML-enabled) management of storage media access. In some aspects, an ML-enabled storage controller obtains features of available blocks of storage media of a storage media system. The controller can receive, from a host system, a request to write data and determine features of the data to be written to the storage media. The controller provides the respective features of the available blocks and the data to a neural network and receives, from the neural network, a selected block of the available blocks for writing of the data. The selected block may include an ML-optimized selection from the available blocks based on the features of both the available blocks and the data. The controller then writes the data of the request to the ML-selected block of storage media of the storage media system, which may improve storage media performance.

Abstract: An apparatus includes a memory and a controller. The memory may be configured to store data. The controller may be configured to process a plurality of input/output requests to a plurality of blocks of the memory that are not marked as bad on a block list, perform a code rate test that programs the plurality of blocks of the memory at three or more code rates of an error correction code scheme, and mark any of the plurality of blocks identified as bad during the code rate test on the block list.

Abstract: A method for capacitance coupling parameter estimation includes determining a plurality of mean voltages among a plurality of memory cells of the memory in each of a plurality of cases related to inter-cell interference, generating a plurality of middle state mean voltages in response to the mean voltages, and adjusting one or more threshold voltages used to read from the memory based on the middle state mean voltages to operate independently of knowledge of middle state distributions in the memory cells.

Abstract: Channel information and channel conditions determined by an Offline Tracking process are used to determine whether or not an adjustment to the read reference voltage can be avoided altogether without detrimentally affecting performance, or, alternatively, to determine a precision with which a read reference voltage adjustment should be made. If it is determined based on the channel conditions that a read reference voltage adjustment can be avoided altogether, read performance is improved by reducing the probability that a read reference voltage adjustment needs to be made during normal read operations. If it is determined based on the channel conditions that a read reference voltage adjustment needs to be made with a particular precision, the read reference voltage is so adjusted. This latter approach is advantageous in that relatively fewer adjustments will be made during normal read operations.

Abstract: Methods, systems, and machine-readable storage medium for multi- tier data recovery utilizing a series of progressively more complex detection and decoding modes based on data from additional pages or wordlines. In one aspect, read data is obtained from at least one cell comprising a given page of a flash memory, and reliability values are generated for the cell from the read data. The reliability values are utilized to decode the read data for the given page. If the decoding of the read data fails, a series of successive decoding steps is performed, with each successive decoding step utilizing additional read data to generate reliability values for the decoding. In one example, reads of one or more additional pages in the same wordline are performed. In a second example, several read retries (soft reads) of the same wordline are performed. In a third example, one or more additional neighboring wordlines are read.

Abstract: Techniques are provided for adaptive read threshold voltage tracking with gap estimation between default read threshold voltages. A read threshold voltage for a memory is adjusted by estimating a gap between two adjacent default read threshold voltages using binary data from the memory, wherein the gap is estimated using statistical characteristics of at least one of two adjacent memory levels of the memory; computing an adjusted read threshold voltage associated with the two adjacent memory levels by using the statistical characteristics of the two adjacent memory levels and the gap; and updating the read threshold voltage with the adjusted read threshold voltage. Pages of the memory are optionally read at multiple read threshold offset locations to obtain disparity statistics, which can be used to estimate mean and/or standard deviation values for a given memory level. The gap is optionally estimated using the mean and/or standard deviation values.

Abstract: A data processing system includes a likelihood input operable to receive encoded data, a decoder operable to apply a decoding algorithm to likelihood values for the received encoded data and to yield a decoded output, and a decoder input initialization circuit operable to generate new decoder input values based in part on the likelihood values for the received encoded data after the likelihood values for the received encoded data have failed to converge in the decoder.

Abstract: An apparatus includes an interface and a control circuit. The interface may be configured to process transfers to/from a medium. The control circuit may be configured to generate a trapping set list of trapping sets of a low-density parity check code, classify bit positions of the trapping sets as belonging to either a user bits field or a parity bits field of a codeword, encode data using the low-density parity check code to generate the codeword, and present the codeword to the interface to transfer the codeword to the medium. The generation of the codeword may include at least one of a shortening or a puncturing of bit locations in the codeword in response to the classifying of the bit positions of the trapping sets. All of the data may be held in the bit locations of the codeword other than the bit locations that are shortened or punctured.

Abstract: Data compression techniques are provided that remove redundancy across the boundary of compression search engines. An illustrative method comprises splitting the data frame into a plurality of sub-chunks; comparing at least two of the plurality of sub-chunks to one another to remove at least one sub-chunk from the plurality of sub-chunks that substantially matches at least one other sub-chunk to generate a remaining plurality of sub-chunks; generating matching sub-chunk information for data reconstruction identifying the at least one removed sub-chunk and the corresponding substantially matched at least one other sub-chunk; grouping the remaining plurality of sub-chunks into sub-units; removing substantially repeated patterns within the sub-units to generate corresponding compressed sub-units; and combining the compressed sub-units with the matching sub-chunk information to generate a compressed data frame.

Abstract: An apparatus includes a memory and a controller. The memory may be configured to store data. The controller may be configured to process a plurality of input/output requests to a plurality of blocks of the memory that are not marked as bad on a block list, perform a code rate test that programs the plurality of blocks of the memory at three or more code rates of an error correction code scheme, and mark any of the plurality of blocks identified as bad during the code rate test on the block list.

Abstract: Channel information and channel conditions determined by an Offline Tracking process are used to determine whether or not an adjustment to the read reference voltage can be avoided altogether without detrimentally affecting performance, or, alternatively, to determine a precision with which a read reference voltage adjustment should be made. If it is determined based on the channel conditions that a read reference voltage adjustment can be avoided altogether, read performance is improved by reducing the probability that a read reference voltage adjustment needs to be made during normal read operations. If it is determined based on the channel conditions that a read reference voltage adjustment needs to be made with a particular precision, the read reference voltage is so adjusted. This latter approach is advantageous in that relatively fewer adjustments will be made during normal read operations.

Abstract: Systems and method relating generally to solid state memory, and more particularly to systems and methods for recycling data in a solid state memory. The systems and methods include receiving a data set maintained in a memory device, applying at least one iteration of a data decoding algorithm to the data set by a data decoder circuit to yield a decoded output, counting the number of iterations of the data decoding algorithm applied to the data set to yield an iteration count, and recycling the data set to the memory device. The recycling is triggered based at least in part on the iteration count.

Abstract: An apparatus includes a memory and a controller. The memory may be configured to store data. The controller may be configured to process a plurality of input/output requests to a plurality of blocks of the memory that are not marked as bad on a block list, perform a code rate test that programs the plurality of blocks of the memory at three or more code rates of an error correction code scheme, and mark any of the plurality of blocks identified as bad during the code rate test on the block list.

Abstract: A low-density parity-check decoder utilizes information about hard errors in a storage medium to identify bit locations to flip log-likelihood ratios while attempting to decode codewords. The decoder iteratively flips and saturates log-likelihood ratios for bits at hard error locations and re-decodes until a valid codeword is produced. The decoder also identifies variable nodes associated with trapping sets for iterative log-likelihood ratio bit flipping.

Abstract: Adaptive read reference voltage tracking techniques are provided that employ charge leakage mitigation. An exemplary device for use with multi-level memory cells, comprises a controller configured to: after a predefined time interval that approximates a settling time after a programming of the multi-level memory cells until a charge leakage of one or more of the multi-level memory cells has settled, determine a plurality of read reference voltages for the multi-level memory cells using a post-programming adaptive tracking algorithm; and employ the plurality of read reference voltages to read data from the multi-level memory cells. The reference voltage offsets are optionally determined based on a shift in the read reference voltages after the predefined time interval since the programming of the multi-level memory cells.

Abstract: Channel information and channel conditions determined by an Offline Tracking process are used to determine whether or not an adjustment to the read reference voltage can be avoided altogether without detrimentally affecting performance, or, alternatively, to determine a precision with which a read reference voltage adjustment should be made. If it is determined based on the channel conditions that a read reference voltage adjustment can be avoided altogether, read performance is improved by reducing the probability that a read reference voltage adjustment needs to be made during normal read operations. If it is determined based on the channel conditions that a read reference voltage adjustment needs to be made with a particular precision, the read reference voltage is so adjusted. This latter approach is advantageous in that relatively fewer adjustments will be made during normal read operations.

Abstract: A data processing system includes a likelihood input operable to receive encoded data, a decoder operable to apply a decoding algorithm to likelihood values for the received encoded data and to yield a decoded output, and a decoder input initialization circuit operable to generate new decoder input values based in part on the likelihood values for the received encoded data after the likelihood values for the received encoded data have failed to converge in the decoder.

Abstract: Systems and methods for internal copy-back with read-verify are described. In one embodiment, a storage device includes a controller to select a first single level cell (SLC) page of a plurality of SLC pages on the storage device to transfer to a triple level cell (TLC) page. The controller, in conjunction with an error correcting code (ECC) decoder, read-verifies the first SLC page. Read-verifying the first SLC page includes reading the first SLC page to an internal page buffer, decoding the first SLC page read into the internal page buffer, determining a number of errors contained in the first SLC page based at least in part on the decoding, and verifying whether the number of errors contained in the first SLC page satisfies an error threshold. The controller transfers the first SLC page to the TLC page according to a result of read-verifying the first SLC page.

Abstract: Methods and apparatus are provided for adaptive read threshold voltage tracking with separate characterization on each side of a voltage distribution about a distribution mean. A read threshold voltage for a memory is adjusted by determining statistical characteristics of two adjacent memory levels based at least in part on a type of statistical distribution of the memory levels and a distribution of data values read from cells using a plurality of read threshold voltages, wherein the statistical characteristics of the two adjacent memory levels are characterized independently on two sides about at least one mean of the statistical distribution; computing an adjusted read threshold voltage associated with the two adjacent memory levels by using the statistical characteristics of the two adjacent memory levels; and updating the read threshold voltage based on the adjusted read threshold voltage. The adjustment is optionally performed responsive to one or more read errors.

Abstract: A data storage device includes at least one data storage medium and a controller that is operably coupled to the at least one data storage medium. The controller receives the bit stream in a data storage device and performs a first level of compression on the received bit stream to obtain a symbol frame including a plurality of symbols. The controller encodes an initial portion of the plurality of symbols contained in the symbol frame by a fixed encoding scheme. The controller also collects statistics for the initial portion of the symbol frame. The controller then selects at least one data compression algorithm based on the collected statistics. The controller then performs compression encoding on a remaining portion of the symbol frame with the selected at least one data compression algorithm.