Abstract: A computing device may comprise a volatile memory and a non-volatile storage device. Upon system shutdown, contents of the volatile memory may be preserved by memory transfer operations from the volatile memory to the non-volatile storage device. During memory preservation, the computing device may enter a low-power state. The low-power state may comprise suspension of power to a core of a processor while maintaining power to the processor's uncore, and disablement of interrupt signals not related to memory transfer operations. Power delivery to the core of the processor may be periodically resumed to initiate additional memory transfer operations.

Abstract: A computing device may comprise a volatile memory and a non-volatile storage device. Upon system shutdown, contents of the volatile memory may be preserved by memory transfer operations from the volatile memory to the non-volatile storage device. During memory preservation, the computing device may enter a low-power state. The low-power state may comprise suspension of power to a core of a processor while maintaining power to the processor's uncore, and disablement of interrupt signals not related to memory transfer operations. Power delivery to the core of the processor may be periodically resumed to initiate additional memory transfer operations.

Abstract: A computing device may comprise a processor, a volatile memory and a non-volatile storage device. An operating system or firmware of the device may cause one or more pages of the volatile memory to be treated, by applications executing on the computing device, as non-volatile memory pages. A maximum number of pages that may be treated as non-volatile may be determined based on an amount of energy available in a battery and an amount of energy needed to transfer a page of memory to the non-volatile storage device.

Abstract: A computing device may comprise a volatile memory and a non-volatile storage device. Upon system shutdown, contents of the volatile memory may be preserved by memory transfer operations from the volatile memory to the non-volatile storage device. During memory preservation, the computing device may enter a low-power state. The low-power state may comprise suspension of power to a core of a processor while maintaining power to the processor's uncore, and disablement of interrupt signals not related to memory transfer operations. Power delivery to the core of the processor may be periodically resumed to initiate additional memory transfer operations.

Abstract: Various techniques for managing power backup for computing devices are disclosed herein. In one embodiment, a method includes receiving data representing a backup capacity of one or more backup power units and data representing a backup power profile of one or more processing units sharing the one or more backup power units. A portion of the backup capacity may then be assigned to each of the one or more processing units based at least in part on both the received data representing the backup capacity of the one or more backup power units and the received data representing the profile of the one or more processing units.

Abstract: A direct network is described in which each resource is connected to a switching fabric via a set of two or more routing nodes. The routing nodes are distributed so as to satisfy at least one inter-node separation criterion. In one case, the separation criterion specifies that, for each resource, a number of routing nodes that share a same coordinate value with another routing node in the set (in a same coordinate dimension) is to be minimized. In some network topologies, such as a torus network, this means a number of unique loops of the direct network to which each resource is connected is to be maximized. The routing provisions described herein offer various performance benefits, such as improved latency-related performance.

Abstract: A computing device may comprise a volatile memory and a non-volatile storage device. Upon system shutdown, contents of the volatile memory may be preserved by memory transfer operations from the volatile memory to the non-volatile storage device. During memory preservation, the computing device may enter a low-power state. The low-power state may comprise suspension of power to a core of a processor while maintaining power to the processor's uncore, and disablement of interrupt signals not related to memory transfer operations. Power delivery to the core of the processor may be periodically resumed to initiate additional memory transfer operations.

Abstract: A computing device may comprise a processor, a volatile memory and a non-volatile storage device. An operating system or firmware of the device may cause one or more pages of the volatile memory to be treated, by applications executing on the computing device, as non-volatile memory pages. A maximum number of pages that may be treated as non-volatile may be determined based on an amount of energy available in a battery and an amount of energy needed to transfer a page of memory to the non-volatile storage device.

Abstract: Various techniques of managing storage devices in a computing system are described in this application. In one embodiment, a method includes receiving an input containing consumption data representing consumption of a storage device in one of the processing units and determining if the storage device in one of the processing units is consumed excessively. In response to determining that the storage device is consumed excessively, an indicator may be generated to indicate a potential program migration from the one of the processing units to another one of the processing units in the computing system.

Abstract: A computing device may comprise a volatile memory and a non-volatile storage device. Upon system shutdown, contents of the volatile memory may be preserved by memory transfer operations from the volatile memory to the non-volatile storage device. During memory preservation, the computing device may enter a low-power state. The low-power state may comprise suspension of power to a core of a processor while maintaining power to the processor's uncore, and disablement of interrupt signals not related to memory transfer operations. Power delivery to the core of the processor may be periodically resumed to initiate additional memory transfer operations.

Abstract: Technology relating to tuning for operating memory devices is disclosed. The technology includes a computing device that selectively configures operating parameters for at least one operating memory device based at least in part of performance characteristics for an application or other workload that the computing device has been requested to execute. This technology may be implemented, at least in part, in the firmware via a Basic Input/Output System (BIOS) or Unified Extensible Firmware Interface (UEFI) of the computing device. Further, this technology may be employed by a computing device that is executing workloads on behalf of a distributed computing system, e.g., in a data center. Such data centers may include, for example, thousands of computing devices and even more operating memory devices.

Abstract: A computing device may comprise a processor, a volatile memory and a non-volatile storage device. An operating system or firmware of the device may cause one or more pages of the volatile memory to be treated, by applications executing on the computing device, as non-volatile memory pages. A maximum number of pages that may be treated as non-volatile may be determined based on an amount of energy available in a battery and an amount of energy needed to transfer a page of memory to the non-volatile storage device.

Abstract: A computing device may comprise a volatile memory and a non-volatile storage device. Upon system shutdown, contents of the volatile memory may be preserved by memory transfer operations from the volatile memory to the non-volatile storage device. During memory preservation, the computing device may enter a low-power state. The low-power state may comprise suspension of power to a core of a processor while maintaining power to the processor's uncore, and disablement of interrupt signals not related to memory transfer operations. Power delivery to the core of the processor may be periodically resumed to initiate additional memory transfer operations.

Abstract: A computing device may comprise a processor, a volatile memory and a non-volatile storage device. An operating system or firmware of the device may cause one or more pages of the volatile memory to be treated, by applications executing on the computing device, as non-volatile memory pages. A maximum number of pages that may be treated as non-volatile may be determined based on an amount of energy available in a battery and an amount of energy needed to transfer a page of memory to the non-volatile storage device.

Abstract: Various techniques for managing power backup for computing devices are disclosed herein. In one embodiment, a method includes receiving data representing a backup capacity of one or more backup power units and data representing a backup power profile of one or more processing units sharing the one or more backup power units. A portion of the backup capacity may then be assigned to each of the one or more processing units based at least in part on both the received data representing the backup capacity of the one or more backup power units and the received data representing the profile of the one or more processing units.

Abstract: Various techniques for managing power backup for computing devices are disclosed herein. In one embodiment, a method includes receiving data representing a backup capacity of one or more backup power units and data representing a backup power profile of one or more processing units sharing the one or more backup power units. A portion of the backup capacity may then be assigned to each of the one or more processing units based at least in part on both the received data representing the backup capacity of the one or more backup power units and the received data representing the profile of the one or more processing units.

Abstract: A direct network is described in which each resource is connected to a switching fabric via a set of two or more routing nodes. The routing nodes are distributed so as to satisfy at least one inter-node separation criterion. In one case, the separation criterion specifies that, for each resource, a number of routing nodes that share a same coordinate value with another routing node in the set (in a same coordinate dimension) is to be minimized. In some network topologies, such as a torus network, this means a number of unique loops of the direct network to which each resource is connected is to be maximized. The routing provisions described herein offer various performance benefits, such as improved latency-related performance.

Abstract: Various techniques of managing storage devices in a computing system are described in this application. In one embodiment, a method includes receiving an input containing consumption data representing consumption of a storage device in one of the processing units and determining if the storage device in one of the processing units is consumed excessively. In response to determining that the storage device is consumed excessively, an indicator may be generated to indicate a potential program migration from the one of the processing units to another one of the processing units in the computing system.

Abstract: A direct network is described in which each resource is connected to a switching fabric via a set of two or more routing nodes. The routing nodes are distributed so as to satisfy at least one inter-node separation criterion. In one case, the separation criterion specifies that, for each resource, a number of routing nodes that share a same coordinate value with another routing node in the set (in a same coordinate dimension) is to be minimized. In some network topologies, such as a torus network, this means a number of unique loops of the direct network to which each resource is connected is to be maximized. The routing provisions described herein offer various performance benefits, such as improved latency-related performance.

Abstract: Various techniques for managing power backup for computing devices are disclosed herein. In one embodiment, a method includes receiving data representing a backup capacity of one or more backup power units and data representing a backup power profile of one or more processing units sharing the one or more backup power units. A portion of the backup capacity may then be assigned to each of the one or more processing units based at least in part on both the received data representing the backup capacity of the one or more backup power units and the received data representing the profile of the one or more processing units.