DYNAMICALLY MODIFYING A BOUNDARY OF A DEEP LEARNING NETWORK
    11.
    发明申请
    DYNAMICALLY MODIFYING A BOUNDARY OF A DEEP LEARNING NETWORK 审中-公开
    动态修改深度学习网络的边界

    公开(公告)号:US20160098646A1

    公开(公告)日:2016-04-07

    申请号:US14506972

    申请日:2014-10-06

    CPC classification number: G06N20/00 G06N3/0454 G06N7/005 H04L67/10 H04L67/141

    Abstract: A connection between a user device and a network server is established. Via the connection, a deep learning network is formed for a processing task. A first portion of the deep learning network operates on the user device and a second portion of the deep learning network operates on the network server. Based on cooperation between the user device and the network server, a boundary between the first portion and the second portion of the deep learning network is dynamically modified based on a change in a performance indicator that could affect the processing task

    Abstract translation: 建立用户设备与网络服务器之间的连接。 通过连接,形成了用于处理任务的深度学习网络。 深度学习网络的第一部分在用户设备上操作,并且深度学习网络的第二部分在网络服务器上操作。 基于用户设备和网络服务器之间的协作,基于可能影响处理任务的性能指标的变化来动态地修改深度学习网络的第一部分和第二部分之间的边界

    MANAGING RESOURCE COLLISIONS IN A STORAGE COMPUTE DEVICE
    12.
    发明申请
    MANAGING RESOURCE COLLISIONS IN A STORAGE COMPUTE DEVICE 有权
    在存储计算机设备中管理资源冲突

    公开(公告)号:US20160077885A1

    公开(公告)日:2016-03-17

    申请号:US14486234

    申请日:2014-09-15

    CPC classification number: G06F9/52 G06F9/505

    Abstract: A storage compute device includes a data storage section that facilitates persistently storing host data as data objects. The storage compute device also includes two or more compute sections that perform computations on the data objects. A controller monitors resource collisions affecting a first of the compute sections. The controller creates a copy of at least one of the data objects to be processed in parallel at a second of the compute sections in response to the resource collisions

    Abstract translation: 存储计算设备包括有助于持久存储主机数据作为数据对象的数据存储部分。 存储计算设备还包括对数据对象执行计算的两个或更多个计算部分。 控制器监视影响第一个计算部分的资源冲突。 响应于资源冲突,控制器在计算部分的第二个并行处创建要并行处理的至少一个数据对象的副本

    Transfer unit management
    13.
    发明授权
    Transfer unit management 有权
    转运单位管理

    公开(公告)号:US09164837B2

    公开(公告)日:2015-10-20

    申请号:US14025277

    申请日:2013-09-12

    Abstract: Method and apparatus for managing data in a memory, such as a flash memory. In accordance with some embodiments, a non-volatile memory is arranged into a plurality of blocks, with each of the blocks constituting an integral plural number N of fixed-sized, multi-bit transfer units. A processing circuit retrieves at least a portion of the data stored in a selected block to a volatile memory buffer in response to a transfer unit (TU) bit map. The TU bit map is stored in a memory and provides a multi-bit sequence of bits corresponding to the N transfer units of the selected block. The values of the bits in the multi-bit sequence of bits indicate whether the corresponding transfer units are to be retrieved.

    Abstract translation: 用于管理诸如闪存之类的存储器中的数据的方法和装置。 根据一些实施例,非易失性存储器被布置成多个块,其中每个块构成固定大小的多位传输单元的整数个数量N。 处理电路响应于传送单元(TU)位图,将存储在所选块中的数据的至少一部分检索到易失性存储器缓冲器。 TU位图被存储在存储器中并且提供与所选块的N个传送单元相对应的位的多位序列。 多比特序列中的比特值表示是否检索相应的传送单元。

    Using ECC data for write deduplication processing
    14.
    发明授权
    Using ECC data for write deduplication processing 有权
    使用ECC数据进行重复数据删除处理

    公开(公告)号:US09043668B2

    公开(公告)日:2015-05-26

    申请号:US13762436

    申请日:2013-02-08

    CPC classification number: H03M13/2906 G06F11/1048

    Abstract: Method and apparatus for managing data in a memory. In accordance with some embodiments, a first data object and an associated first ECC data set are generated and stored in a non-volatile (NV) main memory responsive to a first set of data blocks having a selected logical address. A second data object and an associated second ECC data set are generated responsive to receipt of a second set of data blocks having the selected logical address. The second data object and the second ECC data set are subsequently stored in the in the NV main memory responsive to a mismatch between the first ECC data set and the second ECC data set.

    Abstract translation: 用于管理存储器中的数据的方法和装置。 根据一些实施例,响应于具有所选逻辑地址的第一组数据块,生成第一数据对象和相关联的第一ECC数据集并将其存储在非易失性(NV)主存储器中。 响应于接收到具有所选逻辑地址的第二组数据块,产生第二数据对象和相关联的第二ECC数据集。 随后,第二数据对象和第二ECC数据集随后存储在NV主存储器中,以响应第一ECC数据集和第二ECC数据集之间的不匹配。

    Restoring Virtualized GCU State Information
    15.
    发明申请
    Restoring Virtualized GCU State Information 审中-公开
    恢复虚拟GCU状态信息

    公开(公告)号:US20150143039A1

    公开(公告)日:2015-05-21

    申请号:US14599725

    申请日:2015-01-19

    Abstract: Method and apparatus for managing a memory, such as but not limited to a flash memory. In accordance with some embodiments, initial state information is stored which identifies an actual state of a garbage collection unit (GCU) of a memory during a normal operational mode. During a restoration mode after a memory power cycle event, a virtualized state of the GCU is determined responsive to the initial state information and to data read from the GCU. The memory is transitioned from the restoration mode to the normal operational mode once the virtualized state for the GCU is determined.

    Abstract translation: 用于管理存储器的方法和装置,例如但不限于闪速存储器。 根据一些实施例,存储在正常操作模式期间识别存储器的垃圾回收单元(GCU)的实际状态的初始状态信息。 在存储器电源循环事件之后的恢复模式期间,响应于初始状态信息和从GCU读取的数据确定GCU的虚拟化状态。 一旦确定了GCU的虚拟化状态,存储器将从恢复模式转换到正常操作模式。

    VARIABLE DATA RECOVERY SCHEME HIERARCHY
    16.
    发明申请
    VARIABLE DATA RECOVERY SCHEME HIERARCHY 有权
    可变数据恢复方案

    公开(公告)号:US20150089278A1

    公开(公告)日:2015-03-26

    申请号:US14034251

    申请日:2013-09-23

    CPC classification number: G06F11/1048

    Abstract: Method and apparatus for managing data in a memory, such as a flash memory. In accordance with some embodiments, a memory has a plurality of solid-state non-volatile memory cells. A processing circuit is connected to the memory and configured to direct the execution of a plurality of read error recovery routines in response to at least one uncorrectable read error in a data set retrieved from the memory. The recovery routines are executed in a selected order based on an elapsed recovery time parameter for each of the recovery routines and an estimated probability of success of each of the recovery routines.

    Abstract translation: 用于管理诸如闪存之类的存储器中的数据的方法和装置。 根据一些实施例,存储器具有多个固态非易失性存储单元。 处理电路连接到存储器并且被配置为响应于从存储器检索的数据集中的至少一个不可校正的读取错误来引导执行多个读取错误恢复例程。 基于每个恢复例程的经过恢复时间参数和每个恢复例程的估计的成功概率,以所选择的顺序执行恢复例程。

    Memory Device with Variable Code Rate
    17.
    发明申请
    Memory Device with Variable Code Rate 有权
    具有变码率的存储器

    公开(公告)号:US20150074487A1

    公开(公告)日:2015-03-12

    申请号:US14025327

    申请日:2013-09-12

    CPC classification number: G06F11/1012

    Abstract: Method and apparatus for managing data in a memory, such as a flash memory. In accordance with some embodiments, the apparatus has a solid-state non-volatile memory and a processing circuit configured to write data to a selected location of the memory. The data are arranged in the form of multi-bit code words each comprising a user data payload and associated parity data configured to correct one or more bit errors in the user data payload. The processing circuit adjusts at least a selected one of a size of the code words, a size of the user data payloads or a size of the parity data responsive to at least a selected one of an accumulated count of access operations upon the selected location or an error rate associated with the selected location.

    Abstract translation: 用于管理诸如闪存之类的存储器中的数据的方法和装置。 根据一些实施例,该装置具有固态非易失性存储器和被配置为将数据写入存储器的选定位置的处理电路。 数据以多位码字的形式排列,每个码字包括用户数据有效载荷和被配置为校正用户数据有效载荷中的一个或多个位错误的相关联的奇偶校验数据。 所述处理电路至少根据所选择的位置的访问操作的累积计数中的至少一个来选择所述码字的大小,所述用户数据有效载荷的大小或所述奇偶校验数据的大小中的一个, 与所选位置相关联的错误率。

    Restoring virtualized GCU state information
    18.
    发明授权
    Restoring virtualized GCU state information 有权
    恢复虚拟化GCU状态信息

    公开(公告)号:US08938597B2

    公开(公告)日:2015-01-20

    申请号:US13658673

    申请日:2012-10-23

    Abstract: Method and apparatus for managing a memory, such as but not limited to a flash memory. In accordance with some embodiments, initial state information is stored which identifies an actual state of a garbage collection unit (GCU) of a memory during a normal operational mode. During a restoration mode after a memory power cycle event, a virtualized state of the GCU is determined responsive to the initial state information and to data read from the GCU. The memory is transitioned from the restoration mode to the normal operational mode once the virtualized state for the GCU is determined.

    Abstract translation: 用于管理存储器的方法和装置,例如但不限于闪速存储器。 根据一些实施例,存储在正常操作模式期间识别存储器的垃圾回收单元(GCU)的实际状态的初始状态信息。 在存储器电源循环事件之后的恢复模式期间,响应于初始状态信息和从GCU读取的数据确定GCU的虚拟化状态。 一旦确定了GCU的虚拟化状态,存储器将从恢复模式转换到正常操作模式。

    ECC Management for Variable Resistance Memory Cells
    19.
    发明申请
    ECC Management for Variable Resistance Memory Cells 有权
    可变电阻存储单元的ECC管理

    公开(公告)号:US20140245108A1

    公开(公告)日:2014-08-28

    申请号:US13779434

    申请日:2013-02-27

    Abstract: A data storage device may generally be constructed and operated with at least a controller configured to identify a variance from a predetermined threshold in at least one variable resistance memory cell and upgrade a first error correction code (ECC) level to a second ECC level for the at least one variable resistance memory cell.

    Abstract translation: 一般来说,数据存储设备可被构造和操作至少一个控制器,该控制器被配置为在至少一个可变电阻存储器单元中识别来自预定阈值的方差,并将第一纠错码(ECC)级别升级到第二ECC级别 至少一个可变电阻存储单元。

Patent Agency Ranking