Micron Technology Incorporated

Minjian Wu

Hui Wang

Christopher Joseph Bueb

Kai Wang

Hongyan Li

Exemplary methods, apparatuses, and systems include a performance throttling manager for controlling performance levels of a memory subsystem using input voltages. The performance throttling manager monitors input voltages to a memory device to determine a current input voltage value, while the memory device is operating at a first performance level. The performance throttling manager detects that the current input voltage value satisfies a threshold voltage value. The performance throttling manager selects a second performance level that is lower than the first performance level and throttles the performance of the memory device in response to detecting that the current input voltage value satisfies a threshold voltage value.

Voltage-triggered performance throttling

Methods, systems, and devices for triggering a refresh for non-volatile memory are described. A host system may communicate with a memory system, where the host system and memory system may be included within a vehicle (e.g., an automotive system). The host system may receive an indication that the vehicle is powering down and may enter a power off state in response to the indication. The host system may detect a trigger (e.g., using a time or temperature input) to switch back to a power on state while the vehicle is powered down, the trigger associated with performing a refresh operation at the memory system. The host system may enter the power on state and may transmit a power on command to the memory system. The memory system may perform the refresh operation on one or more memory cells while the vehicle remains in the powered down state.

Triggering a refresh for non-volatile memory

Partitions within buffer memory, such as may be used as a cyclic buffer for caching or storage of time based telemetric sensor data, can be operated with different programming characteristics to provide a balance between total memory sub-system capacity and endurance. Received first time based telemetric sensor data can be written to one or more partitions or sub-partitions with different programming characteristics. One programming characteristic can provide a lesser data density and greater data endurance than another programming characteristic. The programming characteristic can be data density per memory cell. The different partitions or sub-partitions can be written to randomly, with equal probability, or another metric.

Partitions within buffer memory

A system to progressively generate responses configured to mitigate risk associated with row hammer attacks. Between two successive refreshing of memory cells in a memory device, increasing thresholds are used to detect row hammer attacks. For example, after a first alert of row hammer attacks is generated using a first lower threshold, a first operation associated with the first lower threshold is initiated to mitigate risk associated with row hammer attack; and a second higher threshold is used to detect row hammer attacks. After a second alert of row hammer attacks is generated using the second lower threshold, a second operation associated with the second lower threshold is initiated to mitigate risk associated with row hammer attack.

Progressive detection and response to row hammer attack

Data flush at power loss can include a solid state drive (SSD) receiving a power loss notification from a host. In response to the power loss notification, the SSD can flush dirty pages from a logical-to-physical (L2P) mapping table cached in volatile memory of the SSD to an L2P mapping table stored in non-volatile memory of the SSD. In response to the power loss notification, the SSD can flush dirty pages from an L2P mapping table cached in volatile memory of the host to the L2P mapping table stored in non-volatile memory of the SSD. In response to the power loss notification, the SSD can flush time based telemetric sensor data from volatile memory to non-volatile memory of the SSD.

Data flush at power loss

Apparatuses, systems, and methods for using defrag levels to reduce data loss are provided herein. In a number of embodiments of the present disclosure, a method can include setting a first defrag level for a memory device, determining if a buffer is full while performing defrag operations on the memory device according to the first defrag level, setting a second defrag level for the memory device in response to determining the buffer is full while performing defrag operations according to the first defrag level.

Defrag levels to reduce data loss

Methods, systems, and devices supporting an interface for refreshing non-volatile memory are described. In some examples, a host system may communicate with a memory system, where both the host system and the memory system may be included within a vehicle (e.g., an automotive system). The host system may receive an indication that the vehicle is powering down (e.g., shutting off an engine or lowering power output from a battery). The host system may switch from a first mode corresponding to a first power usage to a second mode corresponding to a second, lower power usage in response to the vehicle powering down, the second mode supporting initiation of a refresh operation at the memory device. The host system may transmit a refresh command to the memory system to refresh non-volatile memory while the vehicle is powered down if the host system is operating in the second mode of operation.

Interface for refreshing non-volatile memory

Programming video data to different portions of memory is described herein. An example system includes a host interface, a memory device having a first portion and a second portion, and a controller coupled to the host interface and the memory device. The controller can be configured to program video data received via the host interface to the first portion of the memory device, and program video data received via the host interface to the second portion of the memory device instead of to the first portion of the memory device in response to receiving a signal that a trigger event has occurred.

Programming video data to different portions of memory

A production host can learn the production state awareness (PSA) modes supported by a memory device and select a particular of one of the supported PSA modes. The memory device can receive host image data from the production host and write the host image data according to the selected PSA mode. The memory device can set a PSA state to load complete after writing the host image data. The memory device can thereby be better situated for being soldered to a memory sub-system.

Production state awareness selection

A redundant multiport memory for vehicle applications can have different ports coupled to different hosts that are configured to provide advanced driver assistance system (ADAS) for the vehicle. Different multiport memory devices can provide primary or secondary storage of data for the hosts. At least one of the hosts can perform a functionality check on at least one of the multiport memory devices and make use of a second multiport memory device to which it is coupled if a first multiport memory device to which it is coupled fails the check.

Redundant multiport memory for vehicle applications

A memory sub-system can allocate a first portion of blocks of a memory device for storage of file system metadata based on a file system and a capacity of the memory device, write video data received from a host within a second portion of the blocks at a first data density, and write file system metadata within the first portion of the blocks at a second data density lesser than the first data density.

Storage of video data and file system metadata

Methods, systems, and devices for fast mode for a memory device are described. In some examples, a memory device may be initialized during a system boot procedure. The memory device may support multiple modes of operation, including at least a first mode that includes a first set of capabilities, and a second made that includes the first set of capabilities, as well as one or more additional capabilities. The memory device may perform the initialization while operating the memory device according to the first mode, which may include delaying one or more actions associated with the one or more additional capabilities. After the system boot procedure is complete, the memory device may operate according to the second mode, which may include performing an action delayed during the system boot.

Fast mode for a memory device

A memory device and method of operation are described. The memory device may include memory cells of a first type that each store a single bit of information and memory cells of a second type that each store multiple bits of information. The memory cells of the first type may be more robust to extreme operating conditions than the second type but may have one or more drawbacks (e.g., lower density). The memory device may identify data to be written, and in response, may identify a temperature of the memory device. If the temperature is within a nominal operating range associated with a low risk of memory errors, the memory device may write the data to the memory cells of the second type. If the temperature is outside the nominal operating range, the memory device may write the data to the memory cells of the first type.

Temperature-based memory management

A memory device and method of operation are described. The memory device may include NAND memory. The memory device may configure a host device to maintain a host-side buffer for data backup. When the memory device determines an error associated with attempting to write data to a page of memory in a memory block, the memory device may indicate the error to the host device. The host device may, based on receiving the indication of the error, transmit to the memory device a backup copy of the data and other impacted data from the circular buffer. The memory device may configure the host-side buffer to have at least a particular size based one or more structural or operational aspects of the memory device.

Data recovery management for memory

Methods, systems, and devices for an improved power supply for a memory device are described. An apparatus may include a memory device, one or more voltage detectors, and one or more voltage converters. A voltage detector may generate an output indicating whether a voltage at a first pin of the apparatus satisfies a threshold. A voltage converter may be coupled with the voltage detector and may be configured to selectively output a second voltage depending on the output of the voltage detector. Circuitry within the memory device may be coupled with one or more voltage detectors and one or more voltage converters and configured to select a supply voltage for another component of the memory device from among the first voltage (e.g., received from the first pin) and the second voltage (e.g., selectively generated and output by the voltage converter) based on the output from the voltage detector.

Growing community of inventors

Shanghai, China

Minjian Wu