Abstract: A host system includes a memory and a processing device coupled to the memory to perform operations including obtaining log data from a memory sub-system of a plurality of memory sub-systems, wherein the log data reflects memory usage of a memory device of the memory sub-system, wherein the memory device is shared by the plurality of host systems, including the host system, connected to the plurality of memory sub-systems; and determining, based on the log data, a schedule of a plurality of processes running on the plurality of host systems, wherein the plurality of processes share the memory device.

Abstract: Methods, systems, and devices for access heatmap generation at a memory device are described. In some examples, a memory device may maintain a register for tracking access operation occurrence, for which access operations of an address of the memory device may be mapped to multiple fields of the register. In some cases, in response to a first access operation performed on a first address of the memory device, the memory device may increment a first field and a second field of the register and, in response to a second access operation performed on a second address of the memory device, the memory device may increment the first field and a third field of the register. In some examples, the memory device may maintain a second register having a set of fields that each indicate a respective address for which an access occurrence satisfies a threshold.

Abstract: Disclosed in some examples are systems, devices, machine-readable mediums, and methods for customizing an in-memory versioning mode for each memory location according to a predicted access behavior to optimize memory device performance. Usage data in a previous time period may be utilized along with policy rules to determine whether to configure a particular memory address as a zero copy or direct copy mode. For example, memory addresses that are read frequently may be configured as direct copy mode to reduce the read latency penalty. This improves the functioning of the memory system by reducing read latency for memory addresses that are frequently read but written to less frequently, and reduces write latency for memory locations that are frequently written to, but not read as frequently.

Abstract: Methods, systems, and devices for performance monitoring for a memory system are described. A memory system may use a set of counters to determine state information for the memory system. The memory system may also use a set of timers to determine latency information for the memory system. In response to a request for performance information, the memory system may transmit state information, latency information, or both to a host system.

Abstract: To implement a multi-format data object in memory, a device can receive an allocation request for a data object that includes a set of data elements. This allocation request includes respective details for a set of formats for the data object, such as details for a first format in the set of formats including. The details can include memory address information and a mapping between a first data element of the data object in the first format to a second data element in a second format in the set of formats. The details can also include identification of a conversion function configured to convert the first data element to the second data element. The device can provide access to the second format of the data object from the first format of the data object in the memory based on the mapping data structure or the conversion data structure.

Abstract: Methods, systems, and devices for techniques for data transfer between tiered memory devices are described. A memory system may include a data transfer engine to manage data transfers between different tiers of memory devices within the memory system. The data transfer engine may receive a command which includes a set of source addresses of each of a set of data sets and a set of destination addresses to which the data sets are to be transferred. The data transfer engine may schedule and perform a transfer operation to transfer each of the set of data sets from the respective source address to the respective destination address. The command may further include an indication of an interrupt policy of a set of interrupt policies supported by the data transfer engine. The set of interrupt policies may determine how the data transfer engine may handle interruptions to the data transfer operation.

Abstract: A system includes a memory device and a processing device coupled to the memory device, and the processing device is to perform operations including determining, by monitoring accesses to the memory device, a plurality of values of one or more memory usage statistics reflecting memory usage by a plurality of requestors connected to the memory sub-system; generating memory usage data by processing the plurality of values of the one or more memory usage statistics; and transmitting, to a requestor of the plurality of requestors, the memory usage data.

Abstract: Methods, systems, and devices for inter-tier metadata storage are described. A controller associated with a memory system may manage metadata storage across tiers of memory within the memory system or across memory systems. The controller may transfer metadata between tiers of memory based on whether an access count associated with the metadata satisfies a threshold. For example, the controller may transfer metadata from a first tier of memory to a second tier of memory if the access count satisfies a threshold count. The controller may transfer the metadata from the second tier of memory to the first tier of memory if the access count fails to satisfy the threshold count.

Abstract: Methods, systems, and devices for techniques for coupled host and memory dies are described. A controller of a memory system may facilitate data rearrangement within a block-addressable memory device based on metadata associated with prefetching data to a byte-addressable memory device or to a host system. For example, the controller may utilize the metadata and various access commands to rearrange associated data within the block-addressable memory device such that the data is written to a singular superblock of the block-addressable memory device. In some examples, one or more counters may be utilized by the controller to determine whether to rearrange the data within the block-addressable memory device.

Abstract: An access counter associated with a segment of a memory device is maintained. The segment comprises a plurality of lines. A first count of the plurality of lines is identified. A subset of the plurality of lines of the segment is monitored. A second count of the subset of the plurality of lines is identified. An access notification for a first line of the subset of the plurality of lines is received. A first value of the access counter is changed by a second value. The second value is weighted based on the first count and the second count. Based on the first value of the access counter, a memory management scheme is implemented.

Abstract: Systems, devices, and methods related to a deep learning accelerator and memory are described. For example, the accelerator can have processing units to perform at least matrix computations of an artificial neural network via execution of instructions. The processing units have a local memory store operands of the instructions. The accelerator can access a random access memory via a system buffer, or without going through the system buffer. A fetch instruction can request an item, available at a memory address in the random access memory, to be loaded into the local memory at a local address. The fetch instruction can include a hint for the caching of the item in the system buffer. During execution of the instruction, the hint can be used to determine whether to load the item through the system buffer or to bypass the system buffer in loading the item.

Abstract: Methods, systems, and devices for access heatmap generation at a memory device are described. In some examples, a memory device may maintain a register for tracking access operation occurrence, for which access operations of an address of the memory device may be mapped to multiple fields of the register. In some cases, in response to a first access operation performed on a first address of the memory device, the memory device may increment a first field and a second field of the register and, in response to a second access operation performed on a second address of the memory device, the memory device may increment the first field and a third field of the register. In some examples, the memory device may maintain a second register having a set of fields that each indicate a respective address for which an access occurrence satisfies a threshold.

Abstract: Devices and methods are disclosed, including receiving, by a memory controller of a memory device, a memory request from a host device; collecting packet trace data from the memory request; including the packet trace data in a log stored in a memory array of the memory device; and returning the log to the host device.

Abstract: Disclosed in some examples are improved address prediction and memory preloading that leverages next-delta prediction and/or far-delta prediction for scheduling using a DNN. Previous memory access sequence data that identify one or more memory addresses previously accessed by one or more processors of a system may be processed and then converted into a sequence of delta values. The sequence of delta values are then mapped to one or more classes that are then input to a DNN. The DNN then outputs a predicted future class identifier sequence that represents addresses that the DNN predicts will be accessed by the processor in the future. The predicted future class identifier sequence is then converted back to a predicted delta value sequence and back into a set of one or more predicted addresses.

Abstract: Methods, systems, and devices for management and control for ephemeral data are described. A host system may configure a memory device of a memory system to store ephemeral data. The host system may configure the memory device for a performance mode, which may include suspending memory management operations of a portion of the memory device based on allocating the portion for storing the ephemeral data. After storing the ephemeral data to the portion, the memory system may perform an error control procedure to identify one or more errors in the ephemeral data. The memory system may transmit an indication of the one or more errors to the host system, where the host system may determine to ignore the errors, or transfer back-up ephemeral data to the portion of the memory device from another portion of the memory device or from another memory device of the memory system.

Abstract: Various examples are directed to systems and methods for generating a set of pseudorandom numbers in a computing system comprising a compute element and a memory device. A memory controller of the memory device may receive, from the compute element, an indication to generate a set of pseudorandom numbers. The memory controller may generate the set of pseudorandom numbers and write the set of pseudorandom numbers to a memory array of the memory device for access by the compute element.

Abstract: Systems, devices, and methods related to a Deep Learning Accelerator and memory are described. For example, the accelerator can have processing units to perform at least matrix computations of an artificial neural network via execution of instructions. The processing units have a local memory store operands of the instructions. The accelerator can access a random access memory via a system buffer, or without going through the system buffer. A fetch instruction can request an item, available at a memory address in the random access memory, to be loaded into the local memory at a local address. The fetch instruction can include a hint for the caching of the item in the system buffer. During execution of the instruction, the hint can be used to determine whether to load the item through the system buffer or to bypass the system buffer in loading the item.

Abstract: Disclosed in some examples, are methods, systems, and machine-readable mediums in which application state is saved using in-memory versioning in a shared memory pool of disaggregated memory. By utilizing a disaggregated memory pool, the processing resources may be on separate devices than the memory those resources are using. As a result of this architecture, a failure of hardware of processing resources or an application does not necessarily also cause the hardware resources of the memory devices to fail. This allows a standby application executing on standby processing resources to quickly resume execution when a primary application fails by utilizing the memory pool assigned to the primary application in the memory pool.

Abstract: An access counter associated with a segment of a memory device is maintained. An access notification for a first line of the segment is received. An access type associated with the access notification is identified. A first value of the access counter is changed by a second value based on the access type. Based on the first value of the access counter, a memory management scheme is implemented.

Abstract: Systems, apparatuses, and methods for memory management are described. For example, these may include a first memory level including memory pages in a memory array, a second memory level including a cache, a pre-fetch buffer, or both, and a memory controller that determines state information associated with a memory page in the memory array targeted by a memory access request. The state information may include a first parameter indicative of a current activation state of the memory page and a second parameter indicative of statistical likelihood (e.g., confidence) that a subsequent memory access request will target the memory page. The memory controller may disable storage of data associated with the memory page in the second memory level when the first parameter associated with the memory page indicates that the memory page is activated and the second parameter associated with the memory page is greater than or equal to a threshold.