A computer-implemented method according to one embodiment includes causing a plurality of I/O queues to be created between an initiator and a storage target device. The created I/O queues are reserved for I/O requests for which adjustments of current priorities of extents of data associated with the I/O requests are to be performed. The method further includes determining identifying information of an I/O request sent from the initiator to the storage target device and determining whether the I/O request was sent from the initiator to the storage target device using one of the created I/O queues. In response to a determination that the I/O request was sent using a first of the created I/O queues having one of the adjustments associated therewith, a tiering manager of the storage target device is instructed to perform the adjustment on the current priority of the extent of data associated with the I/O request.

Methods, systems, and devices for adjusting a granularity associated with read disturb tracking are described. In some examples, a memory system may receive a set of read commands from a host system instructing the memory system to read data stored at a memory array. The memory system may track a quantity of executed read commands corresponding to a first portion of the memory array according to a first granularity and determine whether the quantity of read commands satisfies a threshold. If the quantity of read commands satisfies the threshold, the memory system may adjust the granularity for tracking executed read commands for the first portion from the first granularity to a second granularity. For example, the memory system may increase or decrease the granularity for tracking executed read commands for the first portion. The memory system may use the tracked quantities of executed read commands for read disturb remediation.

A method for computational storage may include storing, at a storage device, two or more portions of data, wherein a first one of the two or more portions of data comprises a first fragment of a record and a second one of the two or more portions of data comprises a second fragment of the record, and performing, by the storage device, an operation on the first and second fragments of the record. The method may further include performing, by the storage node, a second operation on first and second fragments of a second record. The operation may include a data selection operation, and the method may further include sending a result of the data selection operation to a server. The method may further include sending a result of a first data selection operation to a server.

Methods, systems, and devices for improved data management for memory are described. An apparatus may include a memory array including one or more blocks of memory cells. Data read from a block of memory cells may be written to a buffer, to support providing the data to a host system or modification of the data by the host system. If a quantity of read commands performed at the block of memory cells satisfies a threshold, the data may be written from the buffer to a different block of memory cells, rather than the block from which the data was previously read.

Methods, systems, and devices for voltage regulation distribution for stacked memory are described. A stacked memory device may support various techniques for coupling between voltage regulation circuitry of multiple memory dies, or for coupling of voltage regulation circuitry of some memory dies with circuitry associated with operating memory arrays of other memory dies. In some examples, such techniques may include cross-coupling of voltage regulation circuitry based on access activity or a degree of access activity for array circuitry. In some examples, such techniques may include isolating voltage regulation circuitry based on access activity or a degree of access activity for array circuitry. Dynamic coupling or isolation between voltage regulation circuitry may be supported by various signaling related to a stacked memory device, such as signaling between the stacked memory dies, signaling between a memory die and a central controller, or signaling between the stacked memory device and a host device.

Implementations disclosed describe methods and systems to perform the methods of optimizing a size of memory used for accumulation of neural node outputs and for supporting multiple computational paths in neural networks. In one example, a size of memory used to perform neural layer computations is reduced by performing nodal computations in multiple batches, followed by rescaling and accumulation of nodal outputs. In another example, execution of parallel branches of neural node computations include evaluating, prior to the actual execution, the amount of memory resources needed to execute a particular order of branches sequentially and select the order that minimizes this amount or keeps this amount below a target threshold.

The present invention discloses a method for external devices accessing computer memory, which includes: the external device applying to a computer for a memory space with a certain size, and receiving multiple memory blocks fed back by the computer; the external device establishing a memory mapping relation between the external device and the computer by means of a sequential storage structure or a chain storage structure; and when initiating a read-and-write operation, the external device finding the corresponding offset address in said computer according to the memory mapping relation between the external device and the computer, generating a read-and-write burst command, and actualizing read-and-write operations in the computer memory. The present invention can achieve the rapid and continuous access to multiple discontinuous memory areas of the computer memory, and improve the speed in the computer’s operating system and external devices accessing the memory.

A storage system. In some embodiments, the storage system includes a plurality of object stores, and a plurality of data managers, connected to the object stores. The plurality of data managers may include a plurality of processing circuits. A first processing circuit of the plurality of processing circuits may be configured to process primarily input-output operations, and a second processing circuit of the plurality of processing circuits may be configured to process primarily input-output completions.

Method and system for training a machine learning model based on a training dataset formed by data objects distributed across a virtual object storage service. The method comprises fetching from the virtual object storage service, the training dataset; copying the fetched training dataset on a first local storage device and maintaining a list of modifications executed on the training dataset that occurred on the virtual object storage service. The method comprises, upon receiving a request to initiate training of the machine learning model, generating a synchronized training dataset mirroring the training dataset stored in the virtual object storage service; storing the synchronized training dataset in a second local storage device; and fetching training data from the synchronized training dataset stored in the second local storage device as the training of the machine learning model is executed.

In a case of performing control of migrating data from a relatively-old-generation magnetic tape as a migration-source magnetic tape included in one storage pool of the plurality of storage pools to relatively-new-generation magnetic tapes as migration-destination magnetic tapes included in each of the plurality of storage pools, an information processing apparatus excludes, from migration targets, data recorded in a relatively-new-generation magnetic tape included in a storage pool to which the migration-destination magnetic tape belongs.