A system and method for providing dynamic I/O virtualization is herein disclosed. According to one embodiment, a device capable of performing hypervisor-agnostic and device-agnostic I/O virtualization includes a host computer interface, memory, I/O devices (GPU, disk, NIC), and efficient communication mechanisms for virtual machines to communicate their intention to perform I/O operations on the device. According to one embodiment, the communication mechanism may use shared memory. According to some embodiments, the device may be implemented purely in hardware, in software, or using a combination of hardware and software. According to some embodiments, the device may share its memory with guest processes to perform optimizations including but not limited to a shared page cache and a shared heap.

Disclosed herein includes a system, a method, and a device for reading and writing sparse data in a neural network accelerator. A plurality of slices can be established to access a memory having an access size of a data word. A first slice can be configured to access a first side of the data word in memory. Circuitry can access a mask identifying byte positions within the data word having non-zero values. The circuitry can modify the data word to have non-zero byte values stored starting at an end of the first side, and any zero byte values stored in a remainder of the data word. A determination can be made whether a number of non-zero byte values is less than or equal to a first access size of the first slice. The circuitry can write the modified data word to the memory via at least the first slice.

A queuing requester for access to a memory system is provided. Transaction requests are received from two or more requestors for access to the memory system. Each transaction request includes an associated priority value. A request queue of the received transaction requests is formed in the queuing requester. Each transaction request includes an associated priority value. A highest priority value of all pending transaction requests within the request queue is determined. An elevated priority value is selected when the highest priority value is higher than the priority value of an oldest transaction request in the request queue; otherwise the priority value of the oldest transaction request is selected. The oldest transaction request in the request queue with the selected priority value is then provided to the memory system. An arbitration contest with other requesters for access to the memory system is performed using the selected priority value.

The present disclosure includes methods and apparatuses for read cache memory. One apparatus includes a read cache memory apparatus comprising a first DRAM array, a first and a second NAND array, and a controller configured to manage movement of data between the DRAM array and the first NAND array, and between the first NAND array and the second NAND array.

Methods of operating a memory include performing a memory access operation, obtaining an address corresponding to a subsequent memory access operation prior to stopping the memory access operation, stopping the memory access operation, sharing charge between access lines used for the memory access operation and access lines to be used for the subsequent memory access operation, and performing the subsequent memory access operation.

Described is storage system and method for reducing power consumption. The storage system has first and second physical disks configured to provide mirroring. The first physical disk is placed into a power-saving mode of operation, while the second physical disk is in an active mode of operation responding to read and write requests. The first physical disk transitions from the power-saving mode of operation to an active mode of operation for destaging writes pending from cache to the first physical disk, while the second physical disk responds to read and write requests. The second physical disk transitions from the active mode of operation to the power-saving mode of operation, while the first physical disk responds to read and write requests.

Provided are a method and apparatus for performing error handling operations using error signals A first error signal is asserted on an error pin on a bus to signal to a host memory controller that error handling operations are being performed by a memory module controller in response to detecting an error. Error handling operations are performed to return the bus to an initial state in response to detecting the error. A second error signal is asserted on the error pin on the bus to signal that error handling operations have completed and the bus is returned to the initial state.

Provided are a method and apparatus for setting high address bits in a memory module. A memory module controller in the memory module, having pins to communicate on a bus, determines whether high address bits are available for the memory module, uses a predetermined value for at least one high address bit with addresses communicated from a host memory controller in response to determine that the high address bits are not available to address a first address space in the memory module, and uses values communicated from the host memory controller on at least one of the pins used for the at least one high address bit in response to determine that the high address bits are available to address a second address space, wherein the second address space is larger than the first address space.

The present disclosure includes methods and apparatuses for read cache memory. One apparatus includes a read cache memory apparatus comprising a first DRAM array, a first and a second NAND array, and a controller configured to manage movement of data between the DRAM array and the first NAND array, and between the first NAND array and the second NAND array.

Apparatus and methods are disclosed, including memory devices and systems. Example memory devices, systems and methods include a buffer interface to translate high speed data interactions on a host interface side into slower, wider data interactions on a DRAM interface side. The slower, and wider DRAM interface may be configured to substantially match the capacity of the narrower, higher speed host interface. In some examples, the buffer interface may be configured to provide multiple sub-channel interfaces each coupled to one or more regions within the memory structure and configured to facilitate data recovery in the event of a failure of some portion of the memory structure. Selected example memory devices, systems and methods include an individual DRAM die, or one or more stacks of DRAM dies coupled to a buffer die.