Abstract: A system is described for enabling content-based read caching on a virtual machine without downtime for creating the digest file by launching the virtual machine before the digest file is produced and creating the digest file dynamically during runtime. After the virtual machine is executed, a background process creates the digest file in the background. The background process is paused during input/output operations and resumes when there are no pending input/output requests.

Abstract: A memory system may include: a nonvolatile memory device; a volatile memory suitable for storing write data; and a controller suitable for: allocating a normal write buffer in the volatile memory when normal write data are inputted, allocating a first write buffer in the volatile memory when first write data, which are grouped into a first transaction and first total size information on a total size of the first transaction, are inputted, allocating a second write buffer in the volatile memory when second write data, which are grouped into a second transaction and second total size information on a total size of the second transaction, are inputted, managing sizes of the first and second write buffers to change them in response to the first and second total size information, respectively, and managing a size of the normal write buffer to fix it to a set size.

Abstract: To reduce the computational resources used to generate a response, certain inputs from a first request and corresponding outputs are stored in a data structure that is returned to a client device concurrent with a response. When a subsequent request is received, if an input of the subsequent request matches an input stored in the data structure, the output that corresponds to that input is used in a response rather that recomputing the output. For inputs that do not match an input in the data structure, an output is computed and included in a response. The data structure is then returned to the client device concurrent with the subsequent response. Outputs in the data structure may be periodically validated, or may be associated with an expiration time, to ensure that the data structure includes current and valid data.

Abstract: To provide a storage device by which distribution of data to an added memory device may be carried out in a short period of time. In a case where any new SSD is added to the storage device, a storage controller carries out first data distribution processing moving any piece of user data in a parity group from existing SSDs to the added SSD and second data distribution processing increasing the number of pieces of the user data configuring the parity group from a first number to a second number after the first data distribution processing and storing the user data in a plurality of memory devices SSDs.

Abstract: Embodiments of the present invention include receiving, by an operating system, a request from an application to reserve a subset of a memory allocated to the application for mirroring. The request specifies a size of the subset. A first portion of the specified size and a second portion of the specified size of the memory are reserved by the operating system for the mirroring. Data to write to the first portion of the memory is received from the application. The operating system writes the data to the first portion of the memory and initiates a background write-back process of the data to the second portion of the memory.

Abstract: When the hash of the divided data is not duplicate with the hash registered in an in-memory hash table, the hash of the divided data is registered in the in-memory non-duplication data list. When the hash registered in the in-memory non-duplication data list is duplicated with the hash registered in an on-disk hash table, a duplicate count of the duplicate hash registered in the on-disk hash table is increased by 1 and an update time of the hash is updated to the latest value. When the duplication count of the hash registered in the on-disk hash table exceeds a threshold, the hash which is registered in the on-disk hash table and in which the duplication count has exceeded the threshold is moved from the on-disk hash table to the in-memory hash table.

Abstract: Techniques for splitting a disk set involve obtaining stripe shared information in a first disk set, wherein the information indicates a stripe set that shares a plurality of disks. The techniques further involve determining the number of disks to be split from the first disk set. The techniques further involve splitting the first disk set into a first portion and a second portion based on the information and the number of disks to be split, wherein the second portion and one or more newly added disks form a second disk set, and both the first disk set and the second disk set store data through Redundant Array of Independent Disks (RAID). In the case that a current disk set needs to be split due to adding one or more new disks, disks are selected using the collected information, thereby reducing data movements caused by splitting of the disk set.

Abstract: A network interface device comprises a programmable interface configured to provide a device interface with at least one bus between the network interface device and a host device. The programmable interface is programmable to support a plurality of different types of a device interface.

Abstract: Exemplary methods and apparatus are provided for implementing a deep learning accelerator (DLA) or other neural network components within the die of a non-volatile memory (NVM) apparatus using, for example, under-the-array circuit components within the die. Some aspects disclosed herein relate to configuring the under-the-array components to implement feedforward DLA operations. Other aspects relate to backpropagation operations. Still other aspects relate to using an NAND-based on-chip copy with update function to facilitate updating synaptic weights of a neural network stored on a die. Other aspects disclosed herein relate to configuring a solid state device (SSD) controller for use with the NVM. In some aspects, the SSD controller includes flash translation layer (FTL) tables configured specifically for use with neural network data stored in the NVM.

Abstract: Methods and apparatuses to improve the performance of a video game engine using an Entity Component System (ECS) are described herein. In accordance with an embodiment, the ECS creates and uses entities, to represent game objects, which are constructed entirely using value data types. The ECS constructs the entities within a memory in a densely packed linear way, and whereby the ECS constantly monitors (e.g., during game play) objects within a game and adjusts the entity distribution within the memory so that a maximum density of memory usage is maintained in real time as the game is being played.

Abstract: A transaction manager for use with memory is described. The transaction manager can include a write data buffer to store outstanding write requests, a read data multiplexer to select between data read from the memory and the write data buffer, a command queue and a priority queue to store requests for the memory, and a transaction table to track outstanding write requests, each write request associated with a state that is Invalid, Modified, or Forwarded.

Abstract: A memory controller includes a central processing unit (CPU) configured to translate a logical address corresponding to an operation that is to be performed by a memory device into a physical address, and an addressing component configured to acquire information about an addressing rule supported by the memory device, among a plurality of addressing rules, and to configure an addressing table corresponding to the operation using the acquired addressing rule information and the physical address.

Abstract: A method for use in a storage node, the method comprising: instantiating a first poller for detecting whether pending storage device operations have been completed; executing the first poller to identify a first storage device operation that has been completed, wherein executing the first poller includes: (a) executing a first function to detect whether a completion queue corresponding to a storage device driver is empty, the first function being arranged to read a content of a memory location that is associated with the completion queue, (b) terminating the execution of the first poller when the completion queue is empty, and (c) executing a system call function to the storage device driver when the completion queue is non-empty; identifying a first thread that is waiting for the first storage device operation to be completed; and transitioning the first thread from a waiting state to a ready state.

Abstract: Systems and methods for metadata tag inheritance during backup are disclosed. The method includes maintaining a plurality of data storage systems in communication with an external metadata management system, operating the metadata management system to store metadata corresponding to data residing on the data storage systems as a plurality of entries that include or more custom tags, detecting an execution of a backup data operation command on data residing in a data storage system that causes creation of a backup copy of a data set in that data storage system in a destination data storage system, and operating the metadata management system to create a new metadata entry corresponding to the execution of the backup data operation command in response to detecting the execution of the backup data operation command. The new metadata entry includes at least one custom tag created before the execution of the backup data operation command.

Abstract: Systems and methods for metadata tag inheritance are disclosed. The methods include maintaining a plurality of data storage systems in communication with an external metadata management system, operating the metadata management system to store metadata corresponding to data residing on the data storage systems as a plurality of entries, detecting an execution of a data operation command on data residing in a data storage system that causes movement of a data set from that data storage system to a destination data storage system that is different from that data storage system, and operating the metadata management system to create a new metadata entry corresponding to the execution of the data operation command. Each of the plurality of entries may include one or more custom tags, and the new metadata entry may include at least one custom tag created before the execution of the data operation command.

Abstract: A memory controller comprises a command interface to transmit a memory command to a plurality of memory devices associated with the memory controller. The memory controller also comprises an acknowledgement interface to receive an acknowledgment status packet from the plurality of memory devices over a shared acknowledgement link coupled between the memory controller and the plurality of memory devices, the acknowledgement status packet indicating whether the command was received by the plurality of memory devices. In addition, the memory controller comprises a memory controller core to decode the acknowledgment status packet to identify a portion of the acknowledgement status packet corresponding to each of the plurality of memory devices.

Abstract: Utilizing physical cache address comparison for maintaining coherency. Operations are performed on data in lines of a cache of the computing system and virtual addresses are loaded into a cache controller. The virtual addresses correspond with lines associated with performing the operations. A physical address of a line is determined in response to having performed a first cache directory lookup of the line. The physical address from the first operation is compared with other physical addresses associated with other operations to determine whether the other operations utilize the same physical address as the first operation. In response to matching physical locations, determinations are made as to whether a conflict exists in the data at the physical addresses that match. Thus, the coherency maintenance is free from looking up virtual addresses to determine whether the line of the cache includes incoherent data.

Abstract: Systems and methods for metadata tag inheritance are disclosed. The methods include maintaining a plurality of data storage systems in communication with an external metadata management system, operating the metadata management system to store metadata corresponding to data residing on the plurality of data storage systems as a plurality of entries, detecting an execution of a data operation command in a data storage system that causes movement of a data set from source storage location to a destination storage location in that data storage system, and operating the metadata management system to create a new metadata entry corresponding to the execution of the data operation command in response to detecting the execution of the data operation command. Each of the plurality of entries includes one or more custom tags, and the new metadata entry includes at least one custom tag created before the execution of the data operation command.

Abstract: Techniques involve determining updated disk sets. The techniques further involve determining, from the updated disk sets, a target disk set for a storage unit associated with a plurality of storage disks. The techniques further involve determining position information of disk slices corresponding to the storage unit in the plurality of storage disks, in which the position information indicates respective updated disk sets in which the disk slices are located. The techniques further involve in accordance with a determination that an updated disk set indicated by position information of at least one of the disk slices is different from the target disk set, moving data in the at least one disk slice to the target disk set. Accordingly, an arrangement of a storage units across different disk sets can be avoided, and utilization rates of a plurality of storage disks can be made more uniform.

Abstract: Systems and methods for metadata tag inheritance are disclosed. The methods include maintaining a plurality of data storage systems including at least one tiered data storage system in communication with an external metadata management system, operating the metadata management system to store metadata corresponding to data residing on the plurality of data storage systems as a plurality of entries, detecting an execution of a data operation command in the tiered data storage system that causes movement of a data set from a source storage tier to a destination storage tier, and in response to detecting the execution of the data operation command, operating the metadata management system to create a new metadata entry corresponding to the execution of the data operation command. The new metadata entry includes at least one custom tag created before the execution of the data operation command.