Abstract: A system and method for calibrating read threshold voltages includes performing a plurality of read operations, determining to perform a read level tracking method, and performing the read level tracking method. The determining may be based on a temperature change or a bit error rate (BER). The read level tracking method includes determining the BER of an indicative word line, determining an adjusted read threshold level based on the BER, and adjusting read threshold levels according to the adjusted read threshold level.

Abstract: A storage system includes two or more data storage devices and a controller coupled to the two or more data storage devices. Each data storage device of the two or more data storage devices includes zoned namespace (ZNS) architecture. The controller is configured to collect thermal statistics for each data storage device of the two or more data storage devices, analyze the collected thermal statistics, and designate a zone by selecting one or more dies within at least one data storage device of the two or more data storage devices based on the analyzed collected thermal statistics. The data storage device includes a memory device having a plurality of dies and a controller coupled to the memory device. The controller is configured to collect thermal statistics for each die of the plurality of dies, analyze the collected statistics, and allocate one or more dies to form a zone.

Abstract: The present disclosure generally relates to determining host device read patterns and then matching autonomous defragmentation to the read pattern to reduce latency impact and avoid unnecessary write amplification (WA). Host devices tend to read data in similar sized chunks. Additionally, host devices tend to read certain data sequentially. Based upon the typical chunk size and data read, the data can be defragmented in sizes to match the typical host device read chunks, and the data defragmented can then be read sequentially within a same plane even if the defragmented data is on different dies. The data is defragmented without relying upon a host command to be presented. Background operation time is used to move updated data such that a future sequential read is supported.

Abstract: A data storage device includes a memory device, a random access memory (RAM) device, and a controller coupled to the memory device and the RAM device. The controller is configured to determine a workload type of the data storage device, determine to store at least a portion of a compressed logical to physical translation table (ZCAT) in the RAM device based on the workload type, and utilize a remaining portion of the RAM device to perform background operations. The controller is further configured to determine a persistence or temperature of the ZCAT, where the portion of the ZCAT stored to the RAM device is based on the persistence and/or the temperature of the ZCAT. The remaining portion of the ZCAT is stored in the HMB, where the remaining portion may update any out-of-date ZCAT entries.

Abstract: Interleaved ECC coding for key-value data storage devices. In one embodiment, a controller includes a memory interface including a namespace database; an ECC engine; a controller memory; and an electronic processor. The electronic processor is configured to receive a host write command, determine whether write access was setup as a key-value (KV) namespace in the namespace database and is associated with the host write command, and control the ECC engine and the memory interface to perform one or more program operations on the data in the memory using the interleaved ECC coding and based on the host write command in response to determining that the write access was setup as the KV namespace in the namespace database and the KV namespace is associated with the host write command.

Abstract: Storage devices are configured to be utilized in a variety of blockchain related activities that rely on a proof of space consensus model. These storage devices are required to process a lot of read and write cycles on their memory devices to generate the desired proof of space consensus data. The generation and storing of this generated data require very different types of memory device usage. As these operations increase in popularity, the need to segment storage devices meant for proof of space usage and those which are not becomes more important. Storage devices may be configured to throttle these different usage types upon detecting these proof of space blockchain activities. Throttling may include reducing clock frequencies, selecting slower performing trim parameters, and programming memory devices with a reduced voltage window, among other processes. Detecting whether throttling should commence, or end can be done via a deployed machine learning classifier.

Abstract: A data storage device includes a memory device and a controller coupled to the memory device. The controller is configured to determine that a power loss event has occurred, determine that one or more blocks are in an erased state, examine a block of the one or more blocks to determine whether the block is a SLC erased block or a TLC erased block, and place the block in a SLC pre-erase heap if the block is the SLC erased block or in a TLC pre-erase heap if the block is the TLC erased block. The controller is further configured to determine a first bit count of page0 for a SLC voltage for the block, determine a second bit count of page1 for a TLC voltage for the block, and classify the block as either a SLC erased block or a TLC erased block.

Abstract: Recurrent Neural Networks (RNNs) wherein a non-volatile memory (NVM) array provides a memory bank for the RNN. The RNN may include a Neural Turning Machine (NTM) and the memory bank may be an NTM matrix stored in the NVM array. In some examples, a data storage device (DSD) that controls the NVM array includes both a data storage controller and a separate NTM controller. The separate NTM controller accesses the NTM matrix of the NVM array directly while bypassing flash translation layer (FTL) components of the data storage controller. Additionally, various majority wins error detection and correction procedures are described, as well as various disparity count-based procedures.

Abstract: A data storage device includes a memory device and a controller coupled to the memory device. The controller is configured to store a plurality of codewords in the memory device. Each codeword of the plurality of codewords includes host data and parity data corresponding to the host data. Less than all of the plurality of codewords further includes statistics corresponding to the host data. Each statistic of the plurality of codewords is the same or different as another statistic of the plurality of codewords. The statistics are either incremental statistics, adaptive statistics, or both incremental statistics and adaptive statistics.

Abstract: Storage devices are configured to be utilized in a variety of blockchain related activities that rely on a proof of space consensus model. These storage devices are required to process a lot of read and write cycles on their memory devices to generate the desired proof of space consensus data. The generation and storing of this generated data requires very different types of memory device usage. Storage devices may be configured to optimize these different usage types upon detecting these proof of space blockchain activities. These optimizations can include suspending one or more background or other garbage collection activities. Additional optimizations can further include configuring partitions or namespaces to comprise single-level-cell majority or single-level-cell only memory devices to increase writing speeds. Further optimizations can include interleaving or extending the length of error correction codes.

Abstract: Data storage devices, such as solid state drives (SSDs), are disclosed. A read threshold calibration operation is utilized to generate a calibrated read threshold for one or more voltage states of a cell of a MLC memory. A single-level cell (SLC) read is then executed to sense the ratio of bit values at the read thresholds of the voltage states, where SLC read refers to reading at a single read threshold, rather than to the cell type. The sensing results in a binary page with certain statistics of 1's and 0's. The ratio of 1's (or 0's) in the binary page is used to determine a deviation from the expected ratio, where the deviation is used to adjust the calibrated read threshold to match the voltage states of the MLC memory.

Abstract: A data storage device includes a memory and a controller coupled to the memory device. The controller is configured to be coupled to a host device. The controller is further configured to receive a plurality of commands, generate logical block address (LBA) to physical block address (PBA) (L2P) mappings for each of the plurality of commands, and store data of the plurality of commands to a respective PBA according to the generated L2P mappings. Each of the L2P mappings are generated based on a result of a deep learning (DL) training model using a neural network (NN) structure. The controller includes a NN command interpretation unit and a L2P mapping generator coupled to the NN command interpretation unit. The controller is configured to fetch training data and NN parameters from the memory device.

Abstract: The present disclosure generally relates to utilizing improved DL training models stored in non-volatile memory to optimize data transfer and storage. The proposed system would identify workloads of DNN training and occasionally check the difference rate between successive data transfers (representing successive training iterations of the model). Comparing the difference rate to given thresholds could indicate “recommendation-system” typical use case. In such a case the NAND operating system would apply systematic compression of the data by saving only the changed parameters between successive iteration cycles (“batches”). The host may indicate the checkpoint storage configuration of the training model (every iteration, every several iterations etc. . . . ) and other elements. The system may be efficiently utilized combining the NAND based DNN training interface, adding the checkpoint configuration information to the dedicated interface.

Abstract: Storage devices are configured to be utilized in a variety of blockchain related activities that rely on a proof of space consensus model. These storage devices are required to process a lot of read and write cycles on their memory devices to generate the desired proof of space consensus data. The generation and storing of this generated data requires very different types of memory device usage. Storage devices may be configured with a first partition for high-speed access for generating the data, while a second partition is also configured for long-term storage of the generated data. As memory devices reach their estimated end-of-life, they can be dynamically reassigned to the second partition. Likewise, some storage devices may be equipped with multiple memory arrays of different types of memory devices. One set of memory devices can be used for generation, while cheaper, write-once or few memory devices are provided for storage.

Abstract: Exemplary methods and apparatus are provided for implementing a deep learning accelerator (DLA) or other neural network components within the die of a non-volatile memory (NVM) apparatus using, for example, under-the-array circuit components within the die. Some aspects disclosed herein relate to configuring the under-the-array components to implement feedforward DLA operations. Other aspects relate to backpropagation operations. Still other aspects relate to using an NAND-based on-chip copy with update function to facilitate updating synaptic weights of a neural network stored on a die. Other aspects disclosed herein relate to configuring a solid state device (SSD) controller for use with the NVM. In some aspects, the SSD controller includes flash translation layer (FTL) tables configured specifically for use with neural network data stored in the NVM.

Abstract: A data storage device includes a housing and an indicator coupled to the housing. The indicator is configured to indicate a health and/or life stage of the data storage device and operate in the absence of an external power source. The indicator is an electrophoretic display or includes a thermochromic material. The electrophoretic display includes a single indication. The electrophoretic display is a scaling bar. The indicator is coupled to a controller. The controller is configured to calculate a health parameter of the data storage device, determine that that the health parameter has exceeded a threshold, and cause the indicator change from a first state to a second state.

Abstract: A data storage device includes a memory device and a controller coupled to the memory device. The controller is configured to store a plurality of codewords in the memory device. Each codeword of the plurality of codewords includes host data and parity data corresponding to the host data. Less than all of the plurality of codewords further includes statistics corresponding to the host data. Each statistic of the plurality of codewords is the same or different as another statistic of the plurality of codewords. The statistics are either incremental statistics, adaptive statistics, or both incremental statistics and adaptive statistics.

Abstract: A data storage device includes a memory device and a controller coupled to the memory device. The controller is configured to determine an error correction code (ECC) code length for KV pair data and/or an ECC code rate for the KV pair data, where the ECC code length and the ECC code rate are selected according to a value length and decoding capability of the KV pair data, generate ECC parity based on the selecting, and program the KV pair data and the generated ECC parity to the memory device.

Abstract: A data storage device includes a memory device and a controller coupled to the memory device. The controller is configured to determine a health of a plurality of wordlines of a block of a plurality of blocks, receive key value (KV) pair data, select a wordline of the plurality of wordlines based on the health, and program the KV pair data to the selected wordline. The KV pair data includes a value length and a relative performance indicator. The controller is further configured to mark a block of the plurality of blocks due to a high bit error rate (BER) indication, where the marked block is KV operable only. The non-KV pair data stored in the marked block is relocated to a non-marked block.

Abstract: A method and apparatus for cache management and eviction polices using unsupervised reinforcement learning schemes is disclosed.