Abstract: Techniques and mechanisms to provide write access to a memory device. In an embodiment, a memory controller sends commands to a memory device which comprises multiple memory banks. The memory controller further sends a signal specifying that the commands include back-to-back write commands each to access the same memory bank. In response to the signal, the memory device buffers first data of a first write command, wherein the first data is buffered at least until the memory device receives second data of a second write command. Error correction information is calculated for a combination of the first data and second data, and the combination is written to the memory bank. In another embodiment, buffering of the first data and combining of the first data with the second data is performed, based on the signal from the memory controller, in lieu of read-modify-write processing of the first data.

Abstract: Embodiments are generally directed to validation of memory on-die error correction code. An embodiment of a memory device includes one or more memory arrays for the storage of data; control logic to control operation of the memory device; and ECC (error correction code) logic, including ECC correction logic to correct data and ECC generation logic to generate ECC code bits and store the ECC bits in the one or more memory arrays. In a validation mode to validate operation of the ECC logic, the control logic is to allow generation of ECC code bits for a first test value and disable generation of ECC code bits for a second test value.

Abstract: Memory subsystem refresh management enables commands to access one or more identified banks across different bank groups with a single command. Instead of sending commands identifying a bank or banks in separate bank groups by each bank group individually, the command can cause the memory device to access banks in different bank groups. The command can be a refresh command. The command can be a precharge command.

Abstract: A memory device includes at least two independent interface paths, an interface path including multiple memory banks. The memory device can selectively operate in a bank mode or a bank group mode. In bank mode, banks are operated as logical banks, where separate physical banks from different interface paths operate in parallel. When a logic bank is accessed, all physical banks belonging to the logical bank are accessed in parallel across the interface paths. In bank group mode, banks are operated independently, but accessed in bank groups. A separate interface path is operated as an independent bank group, and a bank is individually accessed in its bank group. In bank group mode, back to back access to separate bank groups is possible without resulting in access delay.

Abstract: Detection logic of a memory subsystem obtains a threshold for a memory device that indicates a number of accesses within a time window that causes risk of data corruption on a physically adjacent row. The detection logic obtains the threshold from a register that stores configuration information for the memory device, and can be a register on the memory device itself and/or can be an entry of a configuration storage device of a memory module to which the memory device belongs. The detection logic determines whether a number of accesses to a row of the memory device exceeds the threshold. In response to detecting the number of accesses exceeds the threshold, the detection logic can generate a trigger to cause the memory device to perform a refresh targeted to a physically adjacent victim row.

Abstract: A memory subsystem has multiple memory devices coupled to a command/address line and an error alert line, the error alert line delay-compensated to provide deterministic alert signal timing. The command/address line and the error alert line are connected between the memory devices and a memory controller that manages the memory devices. The command/address line is driven by the memory controller, and the error alert line is driven by the memory devices.

Abstract: A memory subsystem includes a data bus to couple a memory controller to one or more memory devices. The memory controller and one or more memory devices transfer data for memory access operations. The data transfer includes the transfer of data bits and associated check bits over a transfer cycle burst. The memory devices include internal error checking and correction (ECC) separate from the system ECC managed by the memory controller. With a 2N transfer cycle for 2?N data bits for a memory device, the memory devices can provide up to 2N memory locations for N+1 internal check bits, which can leave up to (2N minus (N+1)) extra bits to be used by the system for more robust ECC.

Abstract: Examples include techniques for a write zero operation. Example techniques include forwarding a write 0 command to a memory device to cause internal activations of column select lines of one or more blocks of memory to cause bit values or contents of the one or more blocks to have or store a value of 0.

Abstract: Described herein are a method and an apparatus for dynamically switching between one or more finite termination impedance value settings to a memory input-output (I/O) interface of a memory in response to a termination signal level. The method comprises: setting a first termination impedance value setting for a termination unit of an input-output (I/O) interface of a memory; assigning the first termination impedance value setting to the termination unit when the memory is not being accessed; and switching from the first termination impedance value setting to a second termination impedance value setting in response to a termination signal level.

Abstract: A memory subsystem manages memory I/O impedance compensation by the memory device monitoring a need for impedance compensation. Instead of a memory controller regularly sending a signal to have the memory device update the impedance compensation when a change is not needed, the memory device can indicate when it is ready to perform an impedance compensation change. The memory controller can send an impedance compensation signal to the memory device in response to a compensation flag set by the memory or in response to determining that a sensor value has changed in excess of a threshold.

Abstract: A programmable data pattern for repeated writes to memory can enable efficient writing of a data pattern to multiple memory locations without transmitting the data pattern for each write. In one embodiment, a memory device includes input/output (I/O) circuitry to receive a command, a register to store a value to indicate a source of a data pattern to write in response to receipt of the command, and access circuitry to, in response to receipt of the command, write the data pattern to memory based on the source indicated by the value in the register.

Abstract: A memory controller issues a targeted refresh command. A specific row of a memory device can be the target of repeated accesses. When the row is accessed repeatedly within a time threshold (also referred to as “hammered” or a “row hammer event”), physically adjacent row (a “victim” row) may experience data corruption. The memory controller receives an indication of a row hammer event, identifies the row associated with the row hammer event, and sends one or more commands to the memory device to cause the memory device to perform a targeted refresh that will refresh the victim row.

Abstract: Examples include techniques for a write zero operation. Example techniques include forwarding a write 0 command to a memory device to cause internal activations of column select lines of one or more blocks of memory to cause bit values or contents of the one or more blocks to have or store a value of 0.

Abstract: Embodiments of the invention describe a dynamic random access memory (DRAM) device that may abort a self-refresh mode to improve the exit time from a DRAM low power state of self-refresh. During execution of a self-refresh mode, the DRAM device may receive a signal (e.g., a device enable signal) from a memory controller operatively coupled to the DRAM device. The DRAM device may abort the self-refresh mode in response to receiving the signal from the memory controller.

Abstract: A memory subsystem enables a refresh abort command. A memory controller can issue an abort for an in-process refresh command sent to a memory device. The refresh abort enables the memory controller to more precisely control the timing of operations executed by memory devices in the case where a refresh command causes refresh of multiple rows of memory. The memory controller can issue a refresh command during active operation of the memory device, which is active operation refresh as opposed to self-refresh when the memory device controls refreshing. The memory controller can then issue a refresh abort during the refresh, and prior to completion of the refresh. The memory controller thus has deterministic control over both the start of refresh as well as when the memory device can be made available for access.

Abstract: A memory subsystem includes a multi-device package including multiple memory devices organized as multiple ranks of memory. A control unit for the memory subsystem sends a memory access command concurrently to some or all of the ranks of memory, and triggers some of all of the memory ranks that receive the memory access command to change on-die termination (ODT) settings. One of the ranks is selected to execute the memory access command, and executes the command while all ranks triggered to change the ODT setting have the changed ODT setting.

Abstract: Memory refresh includes timing offsets for different memory devices, to initiate refresh of different memory devices at different times. A memory controller sends a refresh command to cause refresh of multiple memory devices. In response to the refresh command, the multiple memory devices initiate refresh with timing offsets relative to another of the memory devices. The timing offsets reduce the instantaneous power surge associated with all memory devices starting refresh simultaneously.

Abstract: A memory subsystem includes a group of memory devices connected to an address bus. The memory subsystem includes logic to uniquely map a physical address of a memory access command to each memory device of the group. Thus, each physical address sent by an associated memory controller uniquely accesses a different row of each memory device, instead of being mapped to the same or corresponding row of each memory device.

Abstract: A memory subsystem includes a command address bus capable to be operated at double data rate. A memory circuit includes N command signal lines that operate at a data rate of 2R to receive command information from a memory controller. The memory circuit includes 2N command signal lines that operate at a data rate of R to transfer the commands to one or more memory devices. While ratios of 1:2 are specified, similar techniques can be used to send command signals at higher data rates over fewer signal lines from a host to a logic circuit, which then transfers the command signals at lower data rates over more signal lines.

Abstract: Techniques and mechanisms to provide selective access to data error information by a memory controller. In an embodiment, a memory device stores a first value representing a baseline number of data errors determined prior to operation of the memory device with the memory controller. Error detection logic of the memory device determines a current count of data errors, and calculates a second value representing a difference between the count of data errors and the baseline number of data errors. The memory device provides the second value to the memory controller, which is unable to identify that the second value is a relative error count. In another embodiment, the memory controller is restricted from retrieving the baseline number of data errors.