Abstract: The technology described herein is directed towards balancing workload between cluster nodes via redistribution of metadata data structures (e.g., memory tables corresponding to directory table partitions). Workload-related information of a node and its partitions' primary memory tables usage is measured, and if sufficiently high, causes a move of a highly-accessed memory table (corresponding to high workload on a first node) from the first node to a second node that has less workload. The second node can contain a backup (e.g., shallow) memory table to the primary node, whereby the move can be a logical move that transforms the backup memory table into a new instance of the primary memory table on the second node. The first node's primary memory table can be deflated into a backup table on the first node that backs up the new instance of the primary table on the second node.

Abstract: Circuitry comprises a transaction interface to receive a data handling transaction from an upstream device, the data handling transaction defining a target virtual memory address in a virtual memory address space; translation circuitry to access a set of address mappings between virtual memory addresses and physical memory addresses in a physical memory address space; the translation circuitry being configured to initiate handling of the data handling transaction by a downstream device according to a target physical memory address mapped from the target virtual memory address when the set of address mappings includes an address mapping for the target virtual memory address, and to provide a transaction response to the transaction interface indicating a fault condition when the set of address mappings fails to provide an address mapping for the target virtual memory address; and control circuitry to receive a memory region request from the upstream device, requesting that a memory region in the virtual memory

Abstract: A data storage service obtains derivation code and data. The derivation code is executable to generate derived data from the data. The data storage service stores the derivation code and the data in a logical data container. In response to receiving a request to obtain the derived data, the data storage service uses the derivation code to regenerate the derived data from the data and transmits the derived data to fulfill the request.

Abstract: Three new software instructions assist a processor in performing indirect prefetching, and managing a next-to-prefetch address list. The software instructions populate hardware register locations according to a hardware register description comprising a data structure of at least seven fields. Multiple instances of the data structure, shared across multiple respectively corresponding threads running concurrently, comprise an indirect-prefetch-tracker table. The indirect-prefetch-tracker table assists the processor to efficiently perform indirect prefetching, from random (not necessarily contiguous) memory locations, and reduces processor core real estate dedicated to control and management of data prefetch and loading operations.

Abstract: An operation method of a memory system includes: searching for a valid physical address in memory map segments stored in the memory system, based on a read request from a host, a logical address corresponding to the read requests, and a physical address corresponding to the logical address and performing a read operation corresponding to the read request; caching some of the memory map segments in the host as host map segments based on a read count threshold indicating the number of receptions of the read request for the logical address; and adjusting the read count threshold based on a miss count indicating the number of receptions of the read request with no physical address, and a provision count indicating the number of times the memory map segment is cached in the host.

Abstract: According to one embodiment, a memory system includes a nonvolatile memory and a memory controller. The memory controller is configured: to store, in a buffer, a data set read from a cell unit, and an expected data set generated by an error correction on the data set; to count a number of first and second memory cells corresponding to a first and a second combination of data in the data set and the expected data set, respectively, among the memory cells in the cell unit; to calculate a shift amount of a read voltage used in a read operation from the cell unit, based on the number of the first and second memory cells; and to apply the shift amount to a next read operation from the first cell unit.

Abstract: A storage array controller may receive data to be programmed to a solid-state storage device of a plurality of solid-state storage devices. The storage array controller may identify a type of the data and determine whether to program the data to a low latency portion of the solid-state storage device based on the type of the data. In response to determining to program the data to the low latency portion of the solid-state storage device, the storage array controller may program the data to the low latency portion of the solid-state storage device.

Abstract: A memory system having multiple memory layers includes a first memory layer comprising a volatile memory, a second memory layer comprising a first sub-memory and a second sub-memory. In response to a reference failure that occurred in the first memory layer, to which a read reference failed data and a write reference failed data are respectively loaded from a lower level memory layer.

Abstract: A memory system is provided to include memory devices and a controller including cores controlling the memory devices, respectively. The controller determines whether to perform a global wear-leveling operation based on a write count of the plurality of memory devices corresponding to each of the plurality of cores, performs a barrier operation for a request from a host when the global wear-leveling operation is determined to be performed, updates mapping information for mapping a core to memory device information by swapping the mapping information between different cores based on the write count of each of the plurality of cores and closes an open block assigned to each of the plurality of cores and then assigning a new open block to each of the plurality of cores based on the updated mapping information.

Abstract: An apparatus, according to one embodiment, includes non-volatile memory configured to store data, and a controller and logic integrated with and/or executable by the controller, the logic being configured to: determine, by the controller, that at least one block of the non-volatile memory and/or portion of a block of the non-volatile memory meets a retirement condition, re-evaluate, by the controller, the at least one block and/or the portion of a block to determine whether to retire the at least one block and/or the portion of a block, indicate, by the controller, that the at least one block and/or the portion of a block remains usable when a result of the re-evaluation is not to retire the block, and indicate, by the controller, that the at least one block and/or the portion of a block is retired when the result of the re-evaluation is to retire the block.

Abstract: The present invention relates to a graph-computing-oriented heterogeneous in-memory computing apparatus, comprising a memory control unit, a digital signal processing unit, and a plurality of analog signal processing units using the memory control unit.

Abstract: In an embodiment, a processor includes a plurality of prefetch circuits configured to prefetch data into a data cache. A primary prefetch circuit may be configured to generate first prefetch requests in response to a demand access, and may be configured to invoke a second prefetch circuit in response to the demand access. The second prefetch circuit may implement a different prefetch mechanism than the first prefetch circuit. If the second prefetch circuit reaches a threshold confidence level in prefetching for the demand access, the second prefetch circuit may communicate an indication to the primary prefetch circuit. The primary prefetch circuit may reduce a number of prefetch requests generated for the demand access responsive to the communication from the second prefetch circuit.

Abstract: The present disclosure includes apparatuses and methods for buffer reset commands for write buffers. An example apparatus includes a memory and a controller coupled to the memory. The memory can include an array of resistance variable memory cells configured to store data corresponding to a managed unit across multiple partitions each having a respective write buffer corresponding thereto. The controller can be configured to update the managed unit by providing, to the memory, a write buffer reset command followed by a write command. The memory can be configured to execute the write buffer reset command to place the write buffers in a reset state. The memory can be further configured to execute the write command to modify the content of the write buffers based on data corresponding to the write command and write the modified content of the write buffers to an updated location in the array.

Abstract: A method and system for controlling an MBC configured flash memory device to store data in an SBC storage mode, or a partial MBC storage mode. In a full MBC storage mode, pages of data are programmed sequentially from a first page to an Nth page for each physical row of memory cells. Up to N virtual page addresses per row of memory cells accompany each page to be programmed for designating the virtual position of the page in the row. For SBC or partial MBC data storage, a flash memory controller issues program command(s) to the MBC memory device using less than the maximum N virtual page addresses for each row. The MBC memory device sequentially executes programming operations up to the last received virtual page address for the row.

Abstract: Systems, apparatuses, and methods related to predictive memory management are described. Error correction operations can be performed on a memory system and can include a latency associated with performing various error correction techniques on data and the health of physical addresses used to store the data can be predicted based on that latency information. In an example, a method can include determining, by a controller, latency information corresponding to one or more error correction operations performed on data received by the controller, and assigning, based on the latency information corresponding to a health of physical address locations corresponding to the data, and taking an action involving the physical address locations based, at least in part, on the information corresponding to the health of the plurality of physical address locations corresponding to the data.

Abstract: A valid data merging method, a memory control circuit unit, and a memory storage device are provided. The method includes: obtaining a first system parameter corresponding to a first region and a second system parameter corresponding to a second region; determining whether the first system parameter is greater than the second system parameter; selecting a third physical erasing unit from the second region preferentially and performing a valid data merging operation by using the third physical erasing unit when the first system parameter is greater than the second system parameter; and selecting a fourth physical erasing unit from the first region preferentially and performing the valid data merging operation by using the fourth physical erasing unit when the first system parameter is not greater than the second system parameter.

Abstract: An apparatus includes a cache controller circuit and a multi-ported cache memory including a plurality of cache ways. The cache controller circuit is configured to maintain rank values and a threshold value usable to classify the rank values. A given rank value corresponds to a least recently used one of the plurality of cache ways. The cache controller circuit is further configured to receive, in a common access cycle, first and second memory access requests for the cache memory, and, in response to a determination that the first and second memory access requests correspond to respective first and a second cache ways, compare the corresponding rank values for the first and second cache ways to the threshold value. The cache controller circuit is further configured to, based on the comparison, modify the rank value of a selected one of the first and second cache ways.

Abstract: An operation method of a memory system includes a memory device including plural level memory cells. The operation method includes allocating a physical address according to a physical address allocation scheme which is determined based on an attribute of a write command; and performing a write operation on the allocated physical address.

Abstract: An operation method of a memory system includes: searching for a valid physical address in memory map segments stored in the memory system, based on a read request from a host, a logical address corresponding to the read requests, and a physical address corresponding to the logical address and performing a read operation corresponding to the read request; caching some of the memory map segments in the host as host map segments based on a read count threshold indicating the number of receptions of the read request for the logical address; and adjusting the read count threshold based on a miss count indicating the number of receptions of the read request with no physical address, and a provision count indicating the number of times the memory map segment is cached in the host.

Abstract: One embodiment facilitates data storage. During operation, the system receives data to be stored in a non-volatile memory of a storage device. The system determines, by a flash translation layer module of a control unit which is distinct from the storage device, a physical page address at which the data is to be stored in the non-volatile memory, wherein the flash translation layer module of the control unit determines physical page addresses for data to be stored in a plurality of storage devices. The system stores, by the flash translation layer module of the control unit, a mapping between a logical page address for the data and the physical page address. The system writes the data to the non-volatile memory at the physical page address.