Embodiments are generally directed to enhanced power management for support of priority system events. An embodiment of a system includes a processing element; a memory including a registry for information regarding one or more system events that are designated as priority events; a mechanism to track operation of events that requires Turbo mode operation for execution; and a power control unit to implement a power management algorithm. The system is to maintain an first energy budget and a second residual energy budget for operation in a Turbo power mode, and wherein the power management algorithm is to determine whether to authorize execution of a detected system event in the Turbo power mode based on the second residual energy budget upon determining that the first energy budget is not sufficient for execution of the detected system event and that the detected system event is designated as a priority event. Priority designations for the priority events may include a first High Priority designation and a second Critical designation.

A computer-implemented data processing method, including the steps of: providing a first program including a group of operations arranged to satisfy a first set of operation dependencies, the group of operations being adapted for computing data from at least one data source, generating a second program including the group of operations, arranged to satisfy a second set of operation dependencies, and processing the data from the at least one data source with the second program. The group of operations includes a first operation, a second operation, and a third operation. The first set of operation dependencies includes a first dependency between the first operation and the second operation, a second dependency between the first operation and the third operation, and a third dependency between the second operation and the third operation.

A distributed processing system to which a plurality of distributed nodes are connected, each of the distributed nodes including a plurality of arithmetic devices and an interconnect device, wherein, in the interconnect device and/or the arithmetic devices of one of the distributed nodes, memory areas are assigned to each job to be processed by the distributed processing system, and direct memory access between memories for processing the job is executed at least between interconnect devices, between arithmetic devices or between an interconnect device and an arithmetic device.

The handling of protocol exceptions for deterministic code that communicates with external component(s). A protocol exception host updates an execution state object associated with the deterministic code as the execution of the deterministic code proceeds. The component also detects whether a protocol exception has occurred that was caused by the deterministic code communicating using the protocol with an external component. If the component detects that such a protocol exception has occurred, the component handles the protocol exception. The component also determines whether the handled protocol exception has been successfully handled. If the exception is not successfully handled, the component stops the execution of the deterministic code such that the execution state object includes execution state of the deterministic code up to the stop. Accordingly, the execution state of the deterministic code up to the stop may be later used to resume execution of the deterministic code.

Contextual paste target prediction is used to predict one or more target applications for a paste action, and do so based upon a context associated with the content that has previously been selected and copied. The results of the prediction may be used to present to a user one or more user controls to enable the user to activate one or more predicted application, and in some instances, additionally configure a state of a predicted application to use the selected and copied content once activated. As such, upon completing a copy action, a user may, in some instances, be provided with an ability to quickly switch to an application into which the user was intending to paste the content. This can provide a simpler user interface in a device such as phones and tablet computers with limited display size and limited input device facilities. It can result in a paste operation into a different application with fewer steps than is possible conventionally.

A highly parallelized parallel tempering technique for simulating dynamic systems, such as quantum processors, is provided. Replica exchange is facilitated by synchronizing grid-level memory. Particular implementations for simulating quantum processors by representing cells of qubits and couplers in grid-, block-, and thread-level memory are discussed. Parallel tempering of such dynamic systems can be assisted by modifying replicas based on isoenergetic cluster moves (ICMs). ICMs are generated via secondary replicas which are maintained alongside primary replicas and exchanged between blocks and/or generated dynamically by blocks without necessarily being exchanged. Certain refinements, such as exchanging energies and temperatures through grid-level memory, are also discussed.

Method and system for training a neural network. The neural network is split into first and second portions. A k-layer first portion is sent to a client training/inference engine and the second portion is retained by a server training/inference engine. At the splitting point, the kth layer is a one-way function in output computation has a number of nodes that are less than any other layer of the first portion. The client training/inference engine trains the first portion with input data in a set of training data. The server training/inference engine receives a batch of outputs from the client training and applies them to the second portion to train the entire neural network.

A computer may receive a request to generate a snapshot view of the enterprise network infrastructure. The computer may implement a multithread process to contemporaneously query a plurality of blade servers and server enclosures within the entire network infrastructure. The computer may contemporaneously receive a plurality of information files from the queried network resources (e.g. the blade servers, server enclosures). In active state modes, the computer may push firmware update binaries to the network resources. In a server processing and an active state mode, the computer may implement a multithreaded process to push the firmware update binaries to standalone servers or blade servers that can be accessed directly. In a blade enclosure processing and an active state mode, the computer may implemented a nested multi-threader, using child threads nested within a parent thread to a blade server enclosure to push firmware update binaries to blade servers in the enclosure.

Embodiments described herein provide techniques enable a graphics processor to continue processing operations during the reset of a compute unit that has experienced a hardware fault. Threads and associated context state for a faulted compute unit can be migrated to another compute unit of the graphics processor and the faulting compute unit can be reset while processing operations continue.

Disclosed are various embodiments for providing automated non-billing cycle remittances in order to load-balance batch processing jobs. An eligible user is identified for registration. Then, a registration request is sent to the user. Next, user registration data is received in response to the registration request, the user registration data specifying an event trigger. Then, it is determined that the processing event trigger has occurred. Subsequently, the processing event is executed in response to the processing event trigger occurring.