摘要:
A method, system and article of manufacture are disclosed for registering and deregistering memory pages in a computer system. The method comprises the steps of hoisting register and deregister calls in a given routine where temporal locality is present to overlap computation and communication; using software pipelined registration and deregistration where spatial locality is observed; and using intra-procedural and inter-procedural analysis by a compiler of the computer system to deregister dynamically allocated buffers. The preferred embodiment of the invention is based on an optimizing compiler. The compiler is used to extract information such as addresses of buffers which are being reused repeatedly (temporal locality), preferably in a loop. The compiler may also find information about spatial locality, such as arrays whose indexes are used in a well-defined manner in a series of messages, for example, array pages being accessed in a pre-defined pattern in a loop.
摘要:
A method of measuring energy expenditure in a living subject comprises: (a) administering a specified dose of doubly-labeled water (2H218O) to a living subject; (b) obtaining samples at three or more times of body water from the living subject; (c) measuring 2H/1H, 17O/16O and 18O/16O ratios in each of the obtained samples using optical spectroscopy; and (d) determining (1) a combined value of flux of body water and exhaled carbon dioxide from a change in measured 18O/16O over time, (2) a value of flux of body water alone from a change in measured 2H/1H over time, and (3) a reference value of isotopic background fluctuation from a change in measured 17O/16O over time. Using 17O measurements to estimate background fluctuations of the 2H and 18O decreases the required isotope dosing of subjects or decreases uncertainty at current dosing levels.
摘要:
The present disclosure is directed to a method for managing tasks in a computer system having a plurality of CPUs. Each task in the computer system may be configured to indicate a migration ready indicator of the task. The migration ready indicator for a task may be given when the set of live data for that task reduces or its working set of memory changes. The method may comprise associating a migration readiness queue with each of the plurality of CPUs, the migration readiness queue having a front-end and a back-end; analyzing a task currently executing on a particular CPU, wherein the particular CPU is one of the plurality of CPUs; placing the task in the migration readiness queue of the particular CPU based on status of the task and/or the migration ready indicator of the task; and selecting at least one queued task from the front-end of the migration readiness queue of the particular CPU for migration when the particular CPU receives a task migration command.
摘要:
Systems and methods provide for dynamic generation of media content at the point that content is served. The system and method work within the bounds of simple, existing, open protocols and the media files served are playable by standard media playback clients. The method is driven by a request for media content from a client specifying an edit list. The server opens one or more source files and selects portions of one or more files to transmit based on edit list instructions, and sequentially writes those portions to an output for serving to the client. The method allows for various modifications of the content before serving.
摘要:
According to one aspect of the present disclosure, a method and technique for preprovisioning virtual machines is disclosed. The method includes analyzing requests for a plurality of different virtual machine (VM) configurations where each VM configuration has a plurality of configuration attributes. The method also includes determining a request frequency corresponding to each requested VM configuration, determining a configuration of each provisioned VM, and responsive to determining the configuration of each provisioned VM, determining a configuration for a preprovisioned VM based on a request frequency of the different VM configurations and the configurations of the provisioned VMs. The method further includes creating the preprovisioned VM.
摘要:
A request for video advertisements is received and video advertisements that can be provided in response to the request are identified. Performance features associated with the video advertisements are identified and are provided as input to a video advertisement performance model trained to estimate a play through rate for each of the video advertisements. The video advertisements are ranked based on the estimated play through rates for the video advertisements provided by the video advertisement performance model, and one or more video advertisements are provided in response to the request according to the ranking of the video advertisements.
摘要:
A query including one or more keywords is received. An advertisement associated with the one or more keywords is identified, and display data for displaying the advertisement is generated and a determination is made as to whether the advertisement is associated with a video. Video selection data is generated for displaying the video proximate to the advertisement if the advertisement is determined to be associated with the video, and the display data and the video selection data are provided.
摘要:
A computer implemented method is provided for optimizing at the time of compiling a program that employs a message-passing interface (MPI). The method includes: detecting an MPI application source file; identifying a non-blocking communication within the MPI application source file; and overlapping independent computation concurrently with the non-blocking communication. A system is also provided.
摘要:
Techniques are described for providing point-to-multipoint (P2MP) Ethernet service in a L2 network. Routers providing the Ethernet service allow an administrator to classify local attachment circuits as either “leaf” attachment circuits or “root” attachment circuits to define a tree-like architecture for forwarding Ethernet frames within a VPLS domain. Based on the classifications, each of router constructs flood domains, referred to herein as mesh groups, that control switching behavior between attachment circuits and pseudowires that transport the L2 communications through the VPLS domain. The routers utilize the mesh groups when switching L2 communications to enforce the requirements of E-TREE service or other L2 services in which L2 traffic is constrained within the L2 VPN to tree-like connectivity.
摘要:
At a client, a video is received. The video includes one or more advertisement slots. The video is played back to a user. During the playback of the video, an impending advertisement slot is detected. One or more advertisements are requested for placement in the advertisement slot. The one or more advertisements are received and placed in the advertisement slot.