A programmable switch includes a plurality of ports for communication with devices on a network. Circuitry of the programmable switch is configured to receive a cache line request from a client on the network to obtain a cache line for performing an operation by the client. A port is identified for communicating with a memory device storing the cache line. The memory device is one of a plurality of memory devices used for a distributed cache. The circuitry is further configured to update a cache directory for the distributed cache based on the cache line request, and send the cache line request to the memory device using the identified port. In one aspect, it is determined whether the cache line request is for modifying the cache line.

The devices within an inter-device processing system maintain data coherency in the last level caches of the system as a cache line of data is shared between the devices by utilizing a directory in one of the devices that tracks the coherency protocol states of the memory addresses in the last level caches of the system.

A system includes a plurality of host processors and a plurality of hybrid memory cube (HMC) devices configured as a distributed shared memory for the host processors. An HMC device includes a plurality of integrated circuit memory die including at least a first memory die arranged on top of a second memory die, and at least a portion of the memory of the memory die is mapped to include at least a portion of a memory coherence directory; and a logic base die including at least one memory controller configured to manage three-dimensional (3D) access to memory of the plurality of memory die by at least one second device, and logic circuitry configured to implement a memory coherence protocol for data stored in the memory of the plurality of memory die.

A method includes receiving, by a L2 controller, a request to perform a global operation on a L2 cache and preventing new blocking transactions from entering a pipeline coupled to the L2 cache while permitting new non-blocking transactions to enter the pipeline. Blocking transactions include read transactions and non-victim write transactions. Non-blocking transactions include response transactions, snoop transactions, and victim transactions. The method further includes, in response to an indication that the pipeline does not contain any pending blocking transactions, preventing new snoop transactions from entering the pipeline while permitting new response transactions and victim transactions to enter the pipeline; in response to an indication that the pipeline does not contain any pending snoop transactions, preventing, all new transactions from entering the pipeline; and, in response to an indication that the pipeline does not contain any pending transactions, performing the global operation on the L2 cache.

A computer-implemented method according to one aspect includes receiving a request to perform a transaction in persistent memory; determining a correlation between volatile memory address locations in a volatile transaction cache and persistent memory locations in the persistent memory; performing the transaction within the volatile memory address locations of the volatile transaction cache; identifying modified volatile memory address locations in the volatile transaction cache that have been written during the transaction; logging, within the persistent memory, data within the modified volatile memory address locations; copying the data within the modified volatile memory address locations to corresponding persistent memory locations in the persistent memory, utilizing the determined correlation; and removing the logged data from the persistent memory, in response to determining that the copying has completed.

A method for controlling the data flow in the storage device is applied to a host, and includes obtaining a cache input and output parameter, determining whether the cache input and output parameter meets an overload condition, when the cache input and output parameter meets the overload condition, obtaining a first bandwidth value, where the first bandwidth value is less than a current flushing bandwidth value of the cache, determining a quantity of tokens based on the first bandwidth value, and controlling the data flow in the storage device.

A semiconductor device includes a device memory, and a device coherency engine (DCOH) that shares a coherency state of the device memory based on data in a host device and a host memory. A power supply of device memory is dynamically adjusted based on the coherency state.

A host server in a server cluster has a memory allocator that creates a dedicated host application data cache in storage class memory. A background routine destages host application data from the dedicated cache in accordance with a destaging plan. For example, a newly written extent may be destaged based on aging. All extents may be flushed from the dedicated cache following host server reboot. All extents associated with a particular production volume may be flushed from the dedicated cache in response to a sync message from a storage array.

A memory device includes a page cache comprising a cache register, a memory array configured with a plurality of memory planes, and control logic, operatively coupled with the memory array. The control logic receives, from a requestor, a cache release command indicating that data associated with a first subset of the plurality of memory planes and pertaining to a previous read command was received by the requestor. Responsive to the cache release command, the control logic returns to the requestor, data from the cache register and associated with a second subset of the plurality of memory planes and pertaining to the previous read command, while concurrently copying data associated with the first subset of the plurality of memory planes and pertaining to a subsequent read command into the cache register.

Embodiments for mitigating security vulnerabilities in a heterogeneous computing system are provided. Anomalous cache coherence behavior may be dynamically detected between a host and one or more accelerators using a cache controller at a shared last level cache based upon a pair-based coherence messages functioning as a proxy for indicating one or more security attack protocols.