Abstract: Systems and methods are disclosed for providing port matching features for storage devices and cables. In certain embodiments, a data storage device includes a non-volatile memory, a controller configured to process data storage requests, a plurality of ports associated with different protocols, wherein the plurality of ports have the same connector type, and each port includes a port matching feature indicative of a protocol associated with the port, and a plurality of cables associated with the different protocols, wherein the plurality of cables have the same connector type and are configured to connect to the plurality of ports, and each cable includes a port matching feature indicative of a protocol associated with the cable, wherein the port matching feature of the cable corresponds to the port matching feature of a port of the plurality of ports that is associated with the same protocol.

Abstract: A data storage device processes a mixed workload including a plurality of superblocks to be written to and read from a plurality of memory dies, where each of the plurality of superblocks to be apportioned among the plurality of memory dies. The data storage device writes a first data stripe associated with a first superblock to the plurality of memory dies according to a sequential write pattern, and reads the first data stripe associated with the first superblock from the plurality of memory dies according to a sequential read pattern. The sequential write pattern causes the controller to write to the plurality of memory dies in a first order of memory dies. The sequential read pattern causes the controller to read from the plurality of memory dies in a second order of memory dies different from the first order of memory dies, thereby reducing read collisions.

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: Technology is disclosed for testing a 3D memory structure. The 3D memory structure has blocks with layers of word lines. Each word line is connected to control gates of NAND memory cells. The 3D memory structure may be tested while concurrently applying a set of layer dependent voltages to a corresponding set of word lines. The magnitude of each layer dependent voltage may depend on which layer the word line to which the voltage is applied resides. There may be physical differences between the different layers such as differences in the diameters of the memory holes in which NAND string are formed. The layer dependent voltages provide for a more accurate test in view of these and other physical differences between the different layers.

Abstract: In NAND memory, data sanitization allows a relatively small unit of data (e.g., less than a block) to be effectively destroyed by increasing threshold voltages of memory cells from their programmed threshold voltage to the highest threshold state. To reduce the amount of disturb on memory cells not selected for data sanitization, prior to applying a program voltage to a target word line, a hole based pre-charge operation is performed. More specifically, for NAND strings having a memory cell selected for data sanitation, prior to applying a programming pulse to the corresponding word line, a soft erase operation is performed. After biasing the memory cells and select gates of the NAND strings to a low voltage, a soft erase voltage pulse is applied to the source lines and bit line to pre-charge the NAND string channels with holes.

Abstract: An apparatus includes a semiconductor wafer including a first die and a second die, each including, a selector circuit including a first input terminal coupled to a first die bond pad, a second input terminal, and an output terminal, the selector circuit configured to selectively couple the first input terminal and the second input terminal to the output terminal, a conductor selectively configured to couple the second input terminal to either a first power supply or a second power supply, and an address determined based on a signal value at the output terminal of the selector circuit.

Abstract: Technology is disclosed herein for multi-stage data compaction. In a first data compaction stage valid data fragments from source erase block(s) are programmed into a destination erase block at two bits per memory cell. In a second data compaction stage additional valid data from the source erase block(s) is programmed into the destination erase block at two bits per memory cell. In this second stage, the same physical pages of memory cells in the destination erase block may be programmed such that each memory cell in the destination erase block is programmed to four bits.

Abstract: A device includes arrays of Non-Volatile Memory (NVM) cells. Reference sequences representing portions of a genome are stored in respective groups of NVM cells. Exact matching phase substring sequences representing portions of at least one sample read are loaded into groups of NVM cells. One or more groups of NVM cells are identified where the stored reference sequence matches the loaded exact matching phase substring sequence using the arrays at Content Addressable Memories (CAMs). Approximate matching phase substring sequences are loaded into groups of NVM cells. One or more groups of NVM cells are identified where the stored reference sequence approximately matches the loaded approximate matching phase substring sequence using the arrays as Ternary CAMs (TCAMs). At least one of the reference sequence and the approximate matching phase substring sequence for each group of NVM cells includes at least one wildcard value when the arrays are used as TCAMs.

Abstract: Disclosed herein is a data storage device. A data port transmits data between a host computer system and the data storage device. A non-volatile storage medium stores encrypted user content data and a cryptography engine connected between the data port and the storage medium uses a cryptographic key to decrypt the encrypted user content data. Multiple manager device records each comprise a first key identical for each of the records, and a second key that different for each of the records. The controller generates an authorization request using the first key and receives a response to the request generated by a manager device. The response is specific to that manager device. The controller uses the response to locate the record; decrypts the located manager device record to obtain key data; and generates configuration data based on the key data to register the device.

Abstract: A semiconductor wafer includes semiconductor dies and laser grooves formed in the scribe lines along the long edges of the semiconductor dies. A laser groove extends between the long edges of two adjacent semiconductor dies to encompass the corners of the two adjacent semiconductor dies. When diced, the resulting semiconductor dies have portions of the corners removed.

Abstract: Instead of having all zones open across all dies, optimizing caching of non-direct write active zones using a host append point (HAP) for maximum write bandwidth is sufficient. The controller will calculate the write rate for a jumbo device (JD). Based on the JD with lowest total write rate, the controller will assign the JD a new zone. The controller will then determine whether the write rate is either appropriate for a cache write or for a direct write. Based on the determination, the controller will write data to the new zone. If the controller direct writes to the new zone, then the controller will close the zone when the zone is full. If the controller cache writes the data to the new zone, then the controller will copy the zone to the direct write area and return the cache write zone to a zone pool. The controller will reclassify the zone if the initial classification is incorrect.

Abstract: Clamped semiconductor wafers and clamped semiconductor devices include reservoirs filled with a flowable metal which hardens to allow the wafers/devices to be shipped or stored. The hardened metal may also be reflowed to a liquid to allow clamping of the semiconductor wafers together and to allow clamping of the semiconductor packages together. The flowable metal may be filled into the reservoirs as a liquid or paste. Thereafter, the flowable metal may be cooled to harden the flowable metal into a clamping member.

Abstract: A non-volatile memory system includes a control circuit connected to non-volatile memory cells. The control circuit is configured to concurrently program memory cells connected to different word lines that are in different sub-blocks of different blocks in different planes of a die.

Abstract: A data storage device and method for dynamic controller memory buffer allocation are disclosed. In one embodiment, a data storage device is provided comprising a memory and a controller with a controller memory buffer. The controller is configured to communicate with the non-volatile memory and is further configured to configure a size of the controller memory buffer; receive a request from the host to modify the size of the controller memory buffer during operation of the data storage device; and determine whether to grant the request to modify the size of the controller memory buffer. Other embodiments are possible, and each of the embodiments can be used alone or together in combination.

Abstract: A memory device includes a first electrode, a second electrode, and a magnetic tunnel junction located between the first electrode and the second electrode. The magnetic tunnel junction includes a reference layer, a free layer, a nonmagnetic tunnel barrier layer located between the reference layer and the free layer, and a platinum-containing layer containing platinum and at least one element selected from iridium, hafnium or ruthenium. The platinum-containing layer contacts the free layer.

Abstract: In some situations, the programming of one memory die can be suspended in favor of the programming of another memory die. This can lead to a delay in certain programming operations. To avoid this problem, a data storage device can perform dynamic logical page write ordering by determining an availability of each memory die of a plurality of memory dies and changing a programing order of the plurality of memory dies in response to the determined availability.

Abstract: Bit-flip object insertion techniques are provided for use with a non-volatile memory (NVM) wherein an object is inserted into a background image by flipping or inverting one or more bits within the pixels of the background image that correspond to the shape and insertion location of an object being inserted. In an illustrative example, pixels within the background image that correspond to the shape and insertion location of the object are XORed with binary 1s. This flips the bits of those pixels to change the color (hue) and/or intensity (brightness) of the pixels so the object appears in the background image. In other examples, only the most significant bits of pixels in the background image are inverted (flipped). Exemplary latch-based procedures are described herein for high-speed processing on an NVM die. Multiple plane NVM die implementations are also described for massive processing.

Abstract: Memory devices, or memory systems, described herein may include a controller (e.g., SSD controller) and a NAND memory device for storing inflight data. When the power loss event occurs, a memory system maintains (i.e., not un-select) the existing memory block being programmed at the time of power loss. The existing program operation at the event of power loss can be suspended by controller. The inflight data can be re-sent by controller directly to NAND latches, when power loss event was detected. The memory system can select a next, immediate available erased page and begin one-pulse programming to store the inflight data, without ramping down the program pump and program pulse, which was in use before the power loss event. The existing programming voltage is used to store/program the inflight data via single pulse programming. When power is restored, the inflight data is moved/programmed to another block for good data reliability.

Abstract: Technology is disclosed herein for smart verify in a memory system that has a four bit per cell program mode (or X4 mode) and also a three bit per cell program mode (or X3 mode). The X3 mode uses a three-bit gray code that is based on a four-bit gray code of the X4 mode. The memory system skips verify of states in the X3 mode, while using a considerable portion of the programming logic from the X4 mode. In one X3 mode the memory system skips B-state verify while the number of memory cells having a Vt above an A-state verify voltage is below a threshold. In one X3 mode the memory system determines whether to skip verify for a first set of data states based on a first test and determines whether to skip verify for a second set of data states based on a second test.

Abstract: The present disclosure generally relates to improved unaligned deallocated logical block transfer. Rather than stalling the data-path in unaligned deallocated LBA scenarios, the data-path will work regularly while ignoring the unaligned deallocated indication. The old and non-valid data received for the unaligned deallocated LBA will be written to the host. The device controller will detect the unaligned deallocated LBA and overwrite the data with other values such as 0's or 1's as specified in the standard. The implementation increases the performance of unaligned deallocated commands and the endurance of the NVM. The implementation also simplifies the logic implemented in the device controller.