Amazon Technologies, Inc.

Steven Scott Larson

Thomas A Volpe

Mark Banse

Asif Khan

Kiran Kalkunte Seshadri

Carlos Javier Cabral

Shao-Wen Yang

Hoshgeldy Tachmuradov

Zheng Wang

Douglas Lloyd Mainz

Hebaallah Aly Abdelhalim Aly Ismail

Ruth Ravichandran

Yang Pan

Andrey Li

David Hung Huynh

Various embodiments are disclosed for providing machine learning routines with peripheral device data to infer driver activity and location. Peripheral device data may be collected on a peripheral device having a machine learning routine executing thereon to infer driver activity and perform improved estimation of driver location. Using driver activity and location estimation, contextually relevant delivery workflow assistance may be automatically provided to a delivery driver or other individual without requiring manual input, thereby improving driver safety and operational efficiency.

Machine learning item delivery using sensor data

Disclosed herein are techniques for tracking usage of a storage-class memory. In one embodiment, a method includes receiving a first statistics update entry and a second statistics update entry by a memory controller for a memory, and assembling the statistics update entries into a statistics log entry. The first statistics update entry indicates a number of operations performed on a first memory block in the memory, and the second statistics update entry indicates a number of operations performed on a second memory block in the memory. The method also includes determining a persistent memory region in a persistent memory for storing the statistics log entry, and writing the statistics log entry into the persistent memory region, where the statistics log entry persists in the persistent memory region until the statistics log entry is read back through the memory controller.

Tracking persistent memory usage

An access device may be implemented to provide one or more access channels to non-volatile memory. Memory mapping implemented at the access device may direct a memory controller of the access device to perform access requests, replacing an initial storage location with a different storage location to access in the non-volatile memory device. Address scrambling, encryption, and other modifications to performing an access request may be implemented at the access device, in some embodiments, in addition to the memory mapping techniques.

Memory mapping in an access device for non-volatile memory

Disclosed herein are techniques associated with a Direct Memory Access (DMA) engine that can include a data generation module. The DMA engine can receive, from a processing entity, a particular type of write command with an indicator to write a data pattern to an address. Upon receipt of the particular type of write command, the DMA engine can generate, using a data generation module of the DMA engine, the data pattern to be written to the address. The processing entity can be Central Processing Unit (CPU) including a core configured to process serial commands. The DMA engine can be disposed on a same die as a processing entity or a network interface port.

Pattern generation using a direct memory access engine

Provided are systems and methods for an address translation circuit for a memory controller. In various implementations, the address translation circuit includes an address translation table operable to include a subset of address translations for a processor memory. An address translation memory can include all address translations for the processor memory. The address translation circuit can be operable to receive an input address for a transaction to processor memory. The address translation circuit can determine an index for the address translation table by inputting the input address into a hash function. The address translation circuit can read a row from the address translation table using the index. The address translation circuit can determine whether an entry in the row includes an address translation for the input address. The address translation can generate and output a translated address using the address translation.

Address translation and address translation memory for storage class memory

Provided are systems and methods for an integrated circuit comprising a search engine, which a memory controller can use to manage a page table. In various implementations, the search engine can generate a series of read transactions to read the page table, which is stored in a memory. Each page table entry includes an address translation for processor memory. The memory controller may periodically change the address translations. The search engine can further determine whether data read from an entry in the page table corresponds to a search parameter. The search engine can further output a response, where the response is affirmative when the data read from the entry corresponds to the search parameters, and where the response is negative when no data read from any entry corresponds to the search parameter.

Page table search engine

Provided are systems and methods for an address translation circuit for a memory controller. The address translation circuit can include an address translation table. A first set of rows in the address translation table can be associated with all virtual machine identifiers supported by the memory controller. A second set of rows can be associated with only a particular virtual machine identifier. The address translation circuit can receive an input address for a transaction to processor memory. The address translation circuit can determine an index by inputting the input address into a hash function. The address translation circuit can read a row from the address translation table using the index. The address translation circuit can determine whether an entry in the row includes the address translation for the input address. The address translation circuit can generate and output a translated address using the address translation.

Address translation for storage class memory in a system that includes virtual machines

A contiguous region in memory may be configured to store data so that a first portion of the data is addressable using a first indexing scheme and a second portion of the data is addressable using a second indexing scheme. The first portion of the data may include information which may be used by one entity and the second portion of the data may include different information which may be used by another entity.

Indexing a memory region

A microcontroller may be implanted to manage migration events performed by a separate memory controller to non-volatile memory. Migration events, such as failed writes at the memory controller or wear leveling migrations to move data from one storage location to another may be detected by the microcontroller. The destination location and the data to be migrated may be identified by the microcontroller. The microcontroller may then instruct the memory controller to copy the data to the destination location.

Managing migration events performed by a memory controller

Disclosed herein are techniques for monitoring the performance of a storage-class memory (SCM). In some embodiments, a performance monitoring circuit at an interface between the SCM and a memory controller of the SCM receives transaction commands from the memory controller to the SCM, measures statistics associated with the transaction commands, and determines a utilization rate of the SCM based on the statistics. Based on the determined utilization rate of the SCM, future transaction requests can be optimized to improve the utilization rate of the SCM.

Performance monitoring for storage-class memory

Failed write handling can be implemented at a memory controller for non-volatile memory. Failure of a write to a storage location in the non-volatile memory may be detected. An indication of the failure may be sent to a microcontroller for the non-volatile memory which may return an instruction to write to a different location in the non-volatile memory. Reads and writes to the storage location of the failed write may still be allowed, in some embodiments, by redirecting the reads and writes to a copy of data of the failed write stored in a copy buffer in the memory controller.

Write failure handling for a memory controller to non-volatile memory

Peripheral devices may implement multiple reporting modes for signal interrupts to a host system. Different reporting modes may be determined for interrupts generated at a host system. Reporting modes may be programmatically configured for various operations at the peripheral device. Reporting modes may indicate a reporting technique for transmitting an indication of the interrupt and may indicate a priority assigned to reporting the interrupt. An interrupt controller for the peripheral device may report generated interrupts according to the reporting mode determined for the interrupts.

Growing community of inventors

Georgetown, TX, United States of America

Steven Scott Larson