Abstract: A system includes a processing device and a memory device communicatively coupled to the processing device. The memory device can include a cyclic buffer partition portion and a snapshot partition portion coupled to the cyclic buffer partition portion via hold-up capacitors. The snapshot partition portion can further include a first sub-partition portion having a first programming characteristic and a second sub-partition portion having a second programming characteristic. The processing device can write received data sequentially to the cycle buffer partition portion and write, based at least in part on a determination that a trigger event has occurred, data from the cyclic buffer partition portion to the first sub-partition portion or the second sub-partition portion, or both.

Abstract: A computing system has a processing device (e.g., CPU, FPGA, or GPU) and memory regions (e.g., in a DRAM device) used by the processing device during normal operation. The computing system is configured to: monitor use of the memory regions in volatile memory; based on monitoring the use of the memory regions, identify at least one of the memory regions of the volatile memory; initiate a hibernation process; and during the hibernation process, copy data stored in the identified memory regions to non-volatile memory.

Abstract: Disclosed are devices and methods for improved operations of a memory device. In one embodiment, a method is disclosed comprising periodically recording memory statistics of a dynamic random-access memory in a device, while the device is in a power on state; detecting a command to enter a standby state; analyzing the memory statistics to determine whether a health check should be performed; powering down the device when determining that the health check should be performed; and placing the device in standby mode when determining that the health check should not be performed.

Abstract: Methods, systems, and apparatuses related to multiple instructions sets or redundant instructions for memory access and management. In one approach, a controller of the memory device provides a first set of functions for management of a storage media of the memory device. The controller uses the first set of functions to service one or more requests received from the host device. The controller provides a second set of functions for management of the storage media. The second set includes one or more functions that are redundant to the first set. When a request is received from the host device, in response to determining that the first set of functions is unavailable to service the request, the controller services the request using one or more of the redundant functions of the second set.

Abstract: Devices and methods for preventing unauthorized access to memory devices are disclosed. A one-time programmable (OTP) memory is included in both a memory device and a processing device. The OTP memories store encryption keys and the encryption and decryption of messages between the two devices are used as a heartbeat to determine that the memory device has not been separated from the processing device and, in some instances, connected to a malicious processing device.

Abstract: A memory system includes a processing device (e.g., a controller implemented using a CPU, FPGA, and/or logic circuitry) and memory regions (e.g., in a flash memory or other non-volatile memory) storing data. The processing device receives an access request from a host system that is requesting to read the stored data. In one approach, the memory system is configured to: receive, from the host system over a bus, a read command to access data associated with an address in a non-volatile memory; in response to receiving the read command, access, by the processing device, multiple copies of data stored in at least one memory region of the non-volatile memory; match, by the processing device, data from the copies with each other; select, based on matching data from the copies with each other, first data from a first copy of the copies; and provide, to the host system over the bus, the first data as output data.

Abstract: A vehicle having a control element for the speed, acceleration or direction of the vehicle, multiple identical or redundant computing devices (e.g., each implemented as a system on chip (SoC)) to separately generate driving commands in parallel during autonomous driving of the vehicle, and a command controller coupled between the control element and the computing devices. The commands may have one or more matching groups, where commands within each respective group agree with each other and thus vote for a candidate command representing the group. The computing device outputs a candidate command that represents the largest group for execution by the control element.

Abstract: A system includes a memory device having a plurality of groups of memory cells and a processing device communicatively coupled to the memory device. The processing device is be configured to read a first group of memory cells of the plurality to determine a calibrated read voltage associated with the group of memory cells. The processing device is further configured to determine, using the calibrated read voltage associated with the first group of memory cells, a bit error rate (BER) of a second group of memory cells of the plurality. Prior to causing the memory device to perform a copyback operation on the plurality of groups of memory cells, the processing device is further configured to determine whether to perform a subsequent read voltage calibration on at least the second group of the plurality based, at least partially, on a comparison between the determined BER and a threshold BER.

Abstract: A system for ANN training through processing power of parked vehicles. The system can include a master computing device having a controller configured to control training of an ANN. The training can be performed at least partially in separate parts by computing devices of parked vehicles. The controller can be configured to separate computing tasks of training the ANN into separated tasks. Also, the controller can be configured to assign at least some of the separated tasks to selected computing devices of parked vehicles. The controller can also be configured to receive and assemble results of the separated tasks to train the ANN. The controller can also be configured to train the ANN according to the results. The master computing device can be configured to send the assigned tasks to the selected devices of the vehicles as well as receive, from the selected devices, the results of the assigned tasks.

Abstract: The disclosed embodiments are related to storing critical data in a memory device such as Flash memory device. In one embodiment, a method performed by a controller of a memory device comprises receiving a critical operation from a host processor, the critical operation accessing a memory array; retrieving a temperature value of the memory array from a temperature sensor; and conditionally processing the critical operation based on the temperature value.

Abstract: A memory system having a processing device (e.g., CPU) and memory regions (e.g., in a DRAM device) on the same chip or die. The memory regions store data used by the processing device during machine learning processing (e.g., using a neural network). One or more controllers are coupled to the memory regions and configured to: read data from a first memory region (e.g., a first bank), including reading first data from the first memory region, where the first data is for use by the processing device in processing associated with machine learning; and write data to a second memory region (e.g., a second bank), including writing second data to the second memory region. The reading of the first data and writing of the second data are performed in parallel.

Abstract: An improved black box data recorder for use with autonomous driving vehicles (AVD). In one embodiment, two cyclic buffers are provided to record vehicle sensors data. A first cyclic buffer records raw vehicle sensor data on a volatile memory, while a second cyclic buffer records the same vehicle sensor data, as compressed data, on a non-volatile memory. In a case of a collision or near collision, in one embodiment the buffers are flushed into a non-volatile (NV) storage for retrieval. As long as there is no power interruption, the raw vehicle sensor data will be accessible from the NV storage. If a power interruption occurs, the raw vehicle sensor data held in the volatile memory of the first cyclic buffer will be lost and only the compressed form of the vehicle sensor data from the second cyclic buffer will survive and be accessible.

Abstract: Systems, methods and apparatus to collect sensor data generated in an autonomous vehicle. Sensors of the vehicle generate a sensor data stream that is buffered, in parallel and in a cyclic way, in a first cyclic buffer and a larger second cyclic buffer respectively. An advanced driver assistance system of the vehicle generates an accident signal when detecting or predicting an accident and provides a training signal when detecting a fault in object detection, recognition, identification or classification. The accident signal causes a sensor data stream segment to be copied from the first cyclic buffer into a slot of a non-volatile memory, selected from a plurality of slots in a round robin way. The training signal causes a sensor data stream segment to be copied from the second cyclic buffer into an area of the non-volatile memory outside of the slots reserved for the first cyclic buffer.

Abstract: A method includes writing received data sequentially to a particular location of a cyclic buffer of a memory device according to a first set of threshold voltage distributions. The method further includes performing a touch up operation on the particular location by adjusting the first set of threshold voltage distributions of the data to a second set of threshold voltage distributions in response to a determination that a trigger event has occurred. The second set of threshold voltage distributions can have a larger read window between adjacent threshold voltage distributions of the second set than that of the first set of threshold voltage distributions.

Abstract: A dynamic memory system having multiple memory regions respectively storing multiple types of data. A controller coupled to the dynamic memory system via a communication channel and operatively to: monitor usage of a communication bandwidth of the communication channel; determine to reduce memory bandwidth penalty caused by refreshing the dynamic memory system; and in response, reduce a refresh rate of at least one of the memory regions based on a type of data stored in the respective memory region.

Abstract: A system for ANN training through processing power of parked vehicles. The system can include a master computing device having a controller configured to control training of an ANN. The training can be performed at least partially in separate parts by computing devices of parked vehicles. The controller can be configured to separate computing tasks of training the ANN into separated tasks. Also, the controller can be configured to assign at least some of the separated tasks to selected computing devices of parked vehicles. The controller can also be configured to receive and assemble results of the separated tasks to train the ANN. The controller can also be configured to train the ANN according to the results. The master computing device can be configured to send the assigned tasks to the selected devices of the vehicles as well as receive, from the selected devices, the results of the assigned tasks.

Abstract: A computing system has a first processing device (e.g., CPU, FPGA, or GPU) and memory regions (e.g., in a DRAM device) used by the processing device during normal operation. In one approach, the computing system is configured to: collect data associated with operation of an autonomous vehicle; monitor, by a first processing device, the collected data; and based on the monitoring, determine that an event on the autonomous vehicle has occurred. The computing system is further configured to, in response to determining that the event has occurred, initiate a transfer of data controlled by a second processing device, the transfer including copying data stored in volatile memory of the autonomous vehicle to non-volatile memory of the autonomous vehicle, wherein the second processing device controls copying of the data independently of the first processing device.

Abstract: Disclosed are techniques for improving an advanced driver-assistance system (ADAS) by pre-processing image data. In one embodiment, a method is disclosed comprising receiving one or more image frames captured by an image sensor installed on a vehicle; identifying a position of a skyline in the one or more image frames, the position comprising a horizontal position of the skyline; cropping one or more future image frames based on the position of the skyline, the cropping generating cropped images comprising a subset of the corresponding future image frames; and processing the cropped images at an advanced driver-assistance system (ADAS).

Abstract: The disclosed embodiments are directed to adjusting the operational characteristics of a black box installed in a vehicle. In one embodiment a method is disclosed comprising classifying a mental state of a driver of an automobile, loading at least one setting based on the mental state, the setting defining an operational characteristic of a black box installed in the automobile, configuring the black box using the at least one setting, and recording one or more events by the black box.

Abstract: A map in a cloud service stores physical objects previously detected by other vehicles that have previously traveled over the same road that a current vehicle is presently traveling on. New data received by the cloud service from the current vehicle regarding new objects that are being encountered by the current vehicle can be compared to the previous object data stored in the map. Based on this comparison, an operating status of the current vehicle is determined. In response to determining the status, an action such as terminating an autonomous navigation mode of the current vehicle is performed.