ARTIFICIAL NEURAL NETWORK TRAINING FOR MEAN TIME TO FAILURE PREDICTIONS

    公开(公告)号:US20240346321A1

    公开(公告)日:2024-10-17

    申请号:US18632708

    申请日:2024-04-11

    CPC classification number: G06N3/084

    Abstract: Training an artificial neural network (ANN) can include receiving device design parameters corresponding to a device and operation parameters corresponding to the device. Device throughput characteristics can also be received from a physics solver. Device throughput predictions can be generated utilizing the device design parameters, the operation parameters, and an artificial neural network. A loss gradient can be generated utilizing the device throughput characteristics and the device throughput predictions. The ANN can be trained, utilizing the loss gradient, to generate different device throughput predictions.

    PARALLEL RAS CHANNELS IN CXL MEMORY DEVICE
    423.
    发明公开

    公开(公告)号:US20240345982A1

    公开(公告)日:2024-10-17

    申请号:US18444550

    申请日:2024-02-16

    CPC classification number: G06F13/4221 G06F13/1642

    Abstract: A memory device comprises a plurality of backend interfaces that are each configured to connect to a respective media device (e.g., DDR, LPDDR DRAM), a plurality of reliability, availability and serviceability (RAS) channel datapaths with each RAS channel datapath connected to one or more backend interfaces, one or more caches that are each connected to at least one RAS channel datapath, and control circuitry. The control circuitry is configured to: in response to receiving a memory access request for a memory location, search a cache of the one or more caches for the requested memory location; and when, in response to the search, the requested memory location is not found in the cache, obtain the requested memory location from a media device connected to the at least one RAS channel datapath. Corresponding methods are also described.

    CACHE WITH SET ASSOCIATIVITY HAVING DATA DEFINED CACHE SETS

    公开(公告)号:US20240345958A1

    公开(公告)日:2024-10-17

    申请号:US18748668

    申请日:2024-06-20

    Abstract: A cache system, having: a first cache set; a second cache set; and a logic circuit coupled to a processor to control the caches based on at least respective first and second registers. When a connection to an address bus receives a memory address from the processor, the logic circuit is configured to: generate a set index from at least the address; and determine whether the generated set index matches with a content stored in the first register or with a content stored in the second register. And, the logic circuit is configured to implement a command via the first cache set in response to the generated set index matching with the content stored in the first register and via the second cache set in response to the generated set index matching with the content stored in the second register.

    WEAR LEVELING TECHNIQUES USING DATA CHARACTERISTICS

    公开(公告)号:US20240345951A1

    公开(公告)日:2024-10-17

    申请号:US18630915

    申请日:2024-04-09

    CPC classification number: G06F12/0246 G06F2212/7211

    Abstract: Methods, systems, and devices for wear leveling techniques using data characteristics are described. The described techniques provide for wear leveling across blocks of a memory system. A controller of the memory system may include additional criteria for determining a destination block for an operation, which may include a characteristic of data associated with the operation. The controller may select a destination block according to both an age of the block and the characteristic of the data. For example, the controller may select a relatively young block for data having a first characteristic and may select a relatively old block for data having a second characteristic. In some cases, the controller may partition free blocks into sub-pools based on an average age of virtual blocks (VBs) associated with each free block.

    MEMORY DEVICE HEALTH MONITORING LOGIC
    426.
    发明公开

    公开(公告)号:US20240345932A1

    公开(公告)日:2024-10-17

    申请号:US18630614

    申请日:2024-04-09

    CPC classification number: G06F11/3034 G06F11/3062 G06F11/3075

    Abstract: Methods, systems, and devices for memory device health monitoring logic are described. In accordance with examples as disclosed herein, a memory device may include health monitoring logic configured to monitor a degradation level of the memory device. Further, the health monitoring logic may include a self-check logic to monitor the degradation level of the health monitoring logic. Using the health monitoring logic, the memory device may evaluate and store a health state of the memory device, which may be used to flag a fault in the memory device, among other responsive operations. Additionally, using the self-check logic, the memory device may evaluate and store a health state of the health monitoring logic, which may be used to flag a fault of the previously evaluated health state of the memory device. Based on the self-check flag, a host device may halt or adjust the response operations associated with the memory device.

    MANAGING DATA INTEGRITY USING A CHANGE IN A NUMBER OF DATA ERRORS AND AN AMOUNT OF TIME IN WHICH THE CHANGE OCCURRED

    公开(公告)号:US20240345919A1

    公开(公告)日:2024-10-17

    申请号:US18755592

    申请日:2024-06-26

    Inventor: Ryan G. Fisher

    CPC classification number: G06F11/1048 G06F11/0757 G06F11/0793

    Abstract: Exemplary methods, apparatuses, and systems include performing an initial data integrity scan of a subset of memory at an initial time to determine an initial error rate for the subset of memory. The initial error rate and the initial time are stored. A subsequent integrity scan of the subset of memory is performed at a second time to determine a subsequent error rate for the subset of memory. A difference between the initial error rate and the subsequent error rate is determined. A difference between the initial time and the subsequent time is determined. A remedial action is selected using the difference between the initial error rate and the subsequent error rate and the difference between the initial time and the subsequent time and the remedial action is performed.

    DYNAMICALLY ADJUSTING DATA READ SIZE
    428.
    发明公开

    公开(公告)号:US20240345755A1

    公开(公告)日:2024-10-17

    申请号:US18630609

    申请日:2024-04-09

    Inventor: Wenjun Wu Feng Xu

    CPC classification number: G06F3/0647 G06F3/0625 G06F3/0656 G06F3/0683

    Abstract: Methods, systems, and devices for dynamically adjusting data read size are described. A memory system controller may be configured to dynamically adjust a size of data transmitted to a host system based on determining a quantity of data requested by the host system. The memory system controller may support receiving a read command of a first data size, incrementing a counter by the first data size, requesting the data from a memory device according to a second data size, and transmitting the data from the memory device according to a third data size based on the counter. The memory system controller may determine the counter does not satisfy a threshold and configure the third data size as a relatively small quantity of data. However, the memory system controller may determine the counter satisfies the threshold and configure the third data size as a relatively large quantity of data.

    INTELLIGENT PERFORMANCE MODE SELECTION TO MANAGE QUALITY OF SERVICE

    公开(公告)号:US20240345754A1

    公开(公告)日:2024-10-17

    申请号:US18629767

    申请日:2024-04-08

    Inventor: Lei Pan

    CPC classification number: G06F3/0647 G06F3/0604 G06F3/0679

    Abstract: Exemplary methods, apparatuses, and systems include a performance mode manager for controlling performance of a wireless update by selecting a performance mode using rations of allocation. The performance mode manager receives a request to initialize a file transfer from a host using wireless communication. In response to the request, the performance mode manager identifies a size of the file transfer by the memory subsystem. The performance mode manager selects a performance mode from a plurality of performance modes and allocates the available set of memory pages using the performance mode. The performance mode manager receives a file of the file transfer. The performance mode manager programs a first portion of the file at the default bit density to the first portion of memory and a second portion of the file at the reduced bit density to the second portion of memory.

Patent Agency Ranking