A method and apparatus for determining the sector size and concomitant host metadata size to determine the difference between total size of the data block to be stored, and using the difference for parity data. This allows an increase in parity bits available for smaller sector sizes and/or data with smaller host metadata sizes. Because the amount of space available for additional parity bits is known, data with lower numbers of parity bits may be assigned to higher quality portions a memory array written with longer programming trim times, and/or written to memory dies with good redundant columns, further increasing performance and reliability.

The present disclosure provides a processing device including: a coarse-grained pruning unit configured to perform coarse-grained pruning on a weight of a neural network to obtain a pruned weight, an operation unit configured to train the neural network according to the pruned weight. The coarse-grained pruning unit is specifically configured to select M weights from the weights of the neural network through a sliding window, and when the M weights meet a preset condition, all or part of the M weights may be set to 0. The processing device can reduce the memory access while reducing the amount of computation, thereby obtaining an acceleration ratio and reducing energy consumption.

An electronic device includes a cache memory and a controller. The cache memory includes a set of cache blocks, each cache block having a number of locations usable for storing cache lines. The cache memory also includes a separate set of error correction code (ECC) bits for each of the locations. The controller stores a victim cache line, evicted from a first location in the cache block, in a second location in the cache block. The controller next stores victim reference information in a portion of the plurality of ECC bits for the first location, the victim reference information indicating that the victim cache line is stored in the second location.

In a server system, a host computing platform can have a processing unit separate from the host processor to detect and respond to failure of the host processor. The host computing platform includes a memory to store data for the host processor. The processing unit has an interface to the host processor and the memory and an interface to a network external to the host processor and has access to the memory. In response to detection of failure of the host processor, the processing unit migrates data from the memory to another memory or storage.

A cache read method of a nonvolatile memory device including a plurality of page buffer units and cache latches, each page buffer units having a sensing latch and a sensing node line is provided. The method comprises performing a first on-chip valley search (OVS) read on a selected memory cell using a first sensing node line and a first sensing latch of a first page buffer unit of the plurality of page buffer units; storing first data sensed from the selected memory cell in the first sensing latch, the first data based on a result of the first OVS read; dumping the first data to sensing node lines of at least one page buffer unit, excluding the first page buffer unit, from among the plurality of page buffer units; and performing a second OVS read on the selected memory cell using the first sensing latch.

There are provided methods and systems for correcting an error from a memory. For example, there is provided a system for mitigating an error in a memory. The system can include a memory controller communicatively coupled to a host. The memory controller may be configured to receive information associated with a memory location. The information can indicate the error at the memory location. The controller may be configured to perform, upon receiving the information, certain operations. The operations can include copying data around the memory location, placing the copied data in a reserved area. And the operations can further include outputting, to a central controller, a set of physical addresses associated with the reserved area, wherein the central controller is configured to modify the set of physical address to conduct a data recovery off-line.

An improved solid state drive (SSD). The SSD comprising a plurality of non-volatile memory dies, each configured to store at least one block of data associated with one of a plurality of superblocks each containing a plurality of blocks; a volatile memory; and a memory controller. The memory controller configured to store a bit map associated with a first superblock of the plurality of superblocks in the volatile memory, wherein the bit map is configured to indicate whether each of the plurality of blocks is a replacement block, store a block address list in the volatile memory, the block address list is configured to store an address of one or more replacement blocks, and store a replacement block index in the volatile memory associated with the first superblock of the plurality of superblocks, the replacement block index corresponding to the location of an address of a first replacement block of the first superblock in the block address list.

The present disclosure generally relates to methods of operating storage devices. The storage device comprises a controller comprising first random access memory (RAM1), second random access memory (RAM2), and a storage unit divided into a plurality of streams. When a write command is received to write data to a stream, change log data is generated and stored in the RAM1, the previous delta data for the stream is copied from the RAM2 to the RAM1 to be updated with the change log data, and the updated delta data is copied to the RAM2. The delta data stored in the RAM2 is copied to the storage unit periodically. The controller tracks which delta data has been copied to the RAM2 and to the storage unit. During a power failure, the delta data and the change log data are copied from the RAM1 or the RAM2 to the storage unit.

A method includes determining that a ratio of valid data portions to a total quantity of data portions of a block of memory cells is greater than or less than a valid data portion threshold and determining that health characteristics for the valid data portions of the block of memory cells are greater than or less than a valid data health characteristic threshold. The method further includes performing a first media management operation on the block of memory cells in response to determining that the ratio of valid data portions to the total quantity of data portions is greater than the valid data portion threshold and performing a second media management operation on at least a portion of the block of memory cells in response to determining that the ratio of valid data portions to the total quantity of data portions is less than the valid data portion threshold and the health characteristics for the valid data portions are greater than the valid data health characteristic threshold.

Read operations can be performed to read data stored at a data block. Parameters reflective of a separation between a pair of programming distributions associated with the data block can be determined based on the plurality of read operations. A read request to read the data stored at the data block can be received. In response to receiving the read request, a read operation can be performed to read the data stored at the data block based on the parameters that are reflective of the separation between the pair of programming distributions associated with the data block.