Abstract: A system comprises an executable graph-based model. The executable graph-based model comprises a first overlay node. The first overlay node comprises processing logic that is operable to interact with one or more associated nodes of the executable graph-based model. Further, the executable graph-based model comprises a first node that has the first overlay node associated therewith. The system further comprises a processing unit configured to receive a first stimulus associated with the first overlay node and, in response to the first stimulus being received, cause execution of said processing logic of the first overlay node. Execution of said processing logic of the first overlay node is based on the first node.

Abstract: A system and a method are disclosed for automatic content upload and process. The system retrieves a set of files from a source location based on instructions received from a client device of a user. The system then classifies the set of files into a plurality of categories corresponding to a sequence of one or more services configured to process or store files. The system then generates a data structure storing key values, where the key values are derived based on respective processing of subsets of files. Responsive to receiving an input to execute logic relating to the set of files, the system determines that the input is associated with one or more of the key values, retrieves the one or more of the key values, and executing the logic using the one or more retrieved key values.

Abstract: Methods and systems for managing a queue are disclosed. In one aspect, an example method can comprise accessing at least a portion of a queue comprising a plurality of tasks. Each task of the plurality of tasks can be associated with a property, and the property associated with each task can comprise a respective value. An exclusion value can be determined based on a distribution of the respective values. A group of tasks that comprises respective values of the property that do not match the exclusion value can be selected from the queue, and the selected group of tasks can be processed.

Abstract: When a cooperation source user name corresponding to identification information of a user included in a received request is included in a cooperation destination system user information, a cooperation destination system converts the identification information of the user included in the request into a user ID corresponding to the cooperation source user name in a cooperation destination system user information. The cooperation destination system processes the request based on the user ID and determines whether or not a cooperation source user name corresponding to the user ID is included in the cooperation destination system user information. In a case where the cooperation source user name corresponding to the user ID is included in the cooperation destination system user information, the cooperation destination system converts the user ID into a cooperation source user name corresponding to the user ID in the cooperation destination system user information.

Abstract: A process operating system includes a process control platform, a process operation platform and an endpoint task robot. The process control platform is configured to receive operation information, extract a task from the operation information using a semantic analysis method, and publish the task. The process operation platform is configured to receive the task and store the task in a task queue. After receiving the task, the process operating platform defines a processing flow based on the task, and sorts the order of the task in the task queue. The endpoint task robot is configured to automatically obtain the task from the process operation platform, executes the task according to the processing flow. It then writes the execution result into the log queue and transmits the execution result to the process control platform.

Abstract: A system comprises an executable graph-based model. The executable graph-based model comprises a first overlay node. The first overlay node comprises processing logic that is operable to interact with one or more associated nodes of the executable graph-based model. Further, the executable graph-based model comprises a first node that has the first overlay node associated therewith. The system further comprises a processing unit configured to receive a first stimulus associated with the first overlay node and, in response to the first stimulus being received, cause execution of said processing logic of the first overlay node. Execution of said processing logic of the first overlay node is based on the first node.

Abstract: A system architecture for linking one or more derived objects to existing data objects in a data structure can be organized in a variety of forms. A method for establishing the architecture may include linking first and second data objects, linking first and second data objects. The first and second data objects may be associated with corresponding first and second pluralities of properties. The method may include receiving a first user limitation associated with at least one of the first or second pluralities of properties and receiving a second user limitation associated with at least one of the first or second pluralities of properties. The method can include deriving a conclusion object comprising a third plurality of properties comprising a subset of one or more of the first or second pluralities of properties and linking the conclusion object to the first and second data objects.

Abstract: The present disclosure describes methods, computer-readable media, and apparatuses for operating neural networks. For example, a first apparatus may receive a set of sparse weight vectors. The first apparatus may compress the set of sparse weight vectors to produce a compressed set of sparse weight vectors. The first apparatus may operate a neural network based on the compressed set of sparse weight vectors. In another example, a second apparatus may receive a set of sparse weight vectors. The second apparatus may perform a sparse computation based on the set of sparse weight vectors, and the performance of the sparse computation may produce one or more partial sums. The second apparatus may operate a neural network based at least in part on the one or more partial sums.

Abstract: A method for partitioning executable operations for a reconfigurable computing system includes receiving a set of expressions comprising a plurality of operations and dependencies for those operations, partitioning the plurality of operations into selected executable partitions wherein each selected executable partition conforms to resource constraints for a reconfigurable unit of the reconfigurable computing system. Partitioning the plurality of operations into selected executable partitions may include seeding a candidate partition with an operation, recursively generating an additional candidate partition for each operation adjacent to the candidate partition whose dependent operations are already within the candidate partition or a previously selected partition, and selecting a best candidate partition based on resource cost. A corresponding system and computer-readable medium are also disclosed herein.

Abstract: A graphics processing unit (GPU) or other apparatus includes a plurality of shader engines. The apparatus also includes a first front end (FE) circuit and one or more second FE circuits. The first FE circuit is configured to schedule geometry workloads for the plurality of shader engines in a first mode. The first FE circuit is configured to schedule geometry workloads for a first subset of the plurality of shader engines and the one or more second FE circuits are configured to schedule geometry workloads for a second subset of the plurality of shader engines in a second mode. In some cases, a partition switch is configured to selectively connect the first FE circuit or the one or more second FE circuits to the second subset of the plurality of shader engines depending on whether the apparatus is in the first mode or the second mode.

Abstract: In various examples disclosed herein, provided is a method and system for cross-validating files to facilitate code generation. The system can perform the cross-validation in phases of disparate files to enable the interoperability of the files. The files can include representations of tasks that are organized differently and differ in format and semantics. The method and system described herein can perform syntax and semantic validation on the files to check for errors, and then identify terms and names identified in the tasks associated with the files along with signatures (e.g., schemas for data models or function signatures for function invocations) associated with the names. After error checking is complete, and after the names and signatures are identified, the names and signatures in the respective files can be cross-validated to find matching tasks, and code is generated that utilizes the cross-validated and matching tasks.

Abstract: A fork support is provided for duplicating an application running inside an enclave entity. In this regard, a request to duplicate an application running inside a first enclave may be received by one or more processors of a host computing device of the first enclave. A snapshot of the first enclave including the application may be generated. The snapshot may be encrypted with a snapshot key and copied to untrusted memory of the host. A second enclave may be generated. The snapshot key may be sent from the first enclave to the second enclave through a secure communication channel. The encrypted snapshot may be copied from the untrusted memory of the host into the second enclave. The encrypted snapshot may be decrypted inside the second enclave with the snapshot key.

Abstract: Systems and methods which facilitate management of functionality provided with respect to data objects in a multiuser environment are described. Embodiments are adapted to provide concurrency with respect to the data of the data objects in the multiuser environments, such as through the use of channel grouping techniques for providing various data communications. Additionally or alternatively, embodiments are adapted to implement data object assignments for facilitating management functionality available with respect to the multiuser environment, such as through the use of data object assignment containers created for storage of instances of assigned data objects. Through pre-creation of data object assignments to assignees, multiple users of a multiuser environment are enabled to utilize the functionality provided with respect to data objects without duplication of tasks and/or undesired conflicts between the users.

Abstract: Techniques for subtask assignment for an artificial intelligence (AI) task are described, and may be implemented to leverage a local set of devices to distribute portions of an AI task between the devices. Generally, the described techniques enable AI task allocation based on a variety of factors, such as device capabilities, device availability, task complexity, and so forth.

Abstract: It is determined that a workflow comprising inter-resource requests of a distributed application is to be initiated. In response to detecting, at a first resource during execution of the workflow, that a triggering condition for retrying a request is met, and that a workflow-level retry budget of the workflow indicates that a retry is permitted, the retry is attempted and the budget is modified to indicate that the retry has been attempted. In response to detecting, at another resource, that a triggering condition for retrying another request is met, and that the workflow-level retry budget of the workflow indicates that a retry is not permitted, an indication that the workflow has failed is generated.

Abstract: Techniques for calculating and using a request cost index for throttling application programming interface (API) requests to execute operations in a provider network. The techniques encompass the step receiving a request to execute an operation at an API service in the provider network. Further steps include determining to execute the operation based on a request rate limiting algorithm, executing the operation to yield an operation result, and sending the operation result. Additional steps include calculating a request cost index that reflects an amount of computing resources utilized by executing the operation, determining an adjustment amount for a state variable of the request rate limiting algorithm based on the calculated request cost index, and adjusting (e.g., lowering) the state variable by the adjustment amount. Other (e.g., subsequent) requests to execute queries received at the API service that are metered by the state variable can be throttled by the API service.

Abstract: Embodiments of the present application provide a phased streaming system and process using a dynamic video game client. The dynamic video game client can utilize a state stream game engine in combination with a game application streaming service to provide users with the ability to begin playing games quickly on a huge range of devices.

Abstract: Systems and methods for controlling access to services. Methods may comprise receiving, from a first client of a plurality of clients, a first request to access a first service of a plurality of services. The first service may be associated with a first bulkhead. A first count of concurrent active requests to the first service via the first bulkhead may be determined. If the first count is equal to a first bulkhead maximum value, access to the first service via the first bulkhead may consequently be refused. A second count of concurrent active requests via a shared burst bulkhead may be determined. The second count may correspond to concurrent active requests to any of the plurality of services via the shared burst bulkhead. If the second count is less than a shared burst maximum value, the first request to the first service may be routed via the shared burst bulkhead.

Abstract: An electronic device includes a memory configured to store one or more applications, and at least one processor configured to control the electronic device. The processor may identify a target model file based on a neural network and associated with a target application among the one or more applications, determine a target initialization time of the target model file among a plurality of preset initialization times, and initialize the target model file at the target initialization time. In addition, various example embodiments may be implemented.

Abstract: Provided are a computer program product, system, and method to determine whether a holding of a lock to the resource by a first task satisfies a condition and whether the lock is swappable. The lock is released from the first task and granted to a second task waiting in a queue for the lock in response to determining that the holding of the lock satisfies the condition and that the lock is swappable. The first task is indicated in the queue waiting for the lock in response to granting the lock to the second task.

Abstract: Techniques are disclosed for storing and retrieving large amounts of data in a non-relational database using sort keys. A server computer system may receive a request for raw data specifying a start timestamp and an end timestamp. The server determines a start key and an end key for performing a query on a distributed non-relational database storing key-value pairs, where the determining is based on the start timestamp and the end timestamp. The server may compare the start key and the end key to a sort key included in row keys of key-value pairs stored in the non-relational database. Based on the comparing, the server retrieves one or more rows of raw data from the non-relational database. The server generates a graphical representation of the one or more rows of raw data retrieved from the non-relational database. The disclosed techniques may advantageously improve the efficiency of a database management system.

Abstract: A method for implementing collective communication, a computer device, and a communication system are provided, to resolve problems of high communication latency and resource consumption in the conventional technology. The method includes: identifying a work request having no communication dependency, and when forwarding work requests, directly forwarding the work request identified as having no communication dependency, and forwarding, after queue management, a work request that is not identified as having no communication dependency. According to the application, communication latency caused when the work request having no communication dependency is managed by using a queue can be avoided, and resource consumption caused by performing related management can be reduced, to improve overall communication performance of collective communication.

Abstract: Methods and systems for providing improved and interactive product documentations are provided. In one embodiment, a method is provided that includes providing a documentation page to a computing device. The documentation page may include an identifier of a data environment associated with the documentation page. A request may be received to display the data environment and data associated with the data environment may be retrieved from a database associated with the computing device. The data environment may be created based on the retrieved data and provided to the computing device for display.

Abstract: The embodiments of the present invention provide a dynamic production scheduling method, apparatus and electronic device based on deep reinforcement learning, which relate to the technical field of Industrial Internet of Things, and can reduce the overall processing time of jobs on the basis of not exceeding the processing capacity of production device.

Abstract: Proposed is a system for executing a task of an operating system for a vehicle, the system including: a task information extracting unit which, when an event signal is received from an application, extracts task information related to the event signal from a pre-prepared event chain table; an execution cycle setting unit which sets a task execution cycle by using the extracted task information; a task executing unit which executes any one task among a plurality of tasks according to the task execution cycle; and a waiting time providing unit which provides a task waiting time to the task executing unit so as to wait for execution of a next task when the execution of the one task is completed.

Abstract: Methods, apparatus, systems, and articles of manufacture are disclosed to steal work in heterogeneous computing systems. An apparatus includes load balancing circuitry to obtain tasks from a workload by encoding minimum and maximum index ranges of a data parallel operation, allocate a first task from the workload to a first work queue based on a first capability of first computation circuitry, the first computation circuitry to process the first task in the first work queue, and allocate a second task from the workload to a second work queue, second computation circuitry to process the second task in the second work queue. The apparatus further includes first work stealer logic to steal the second task from the second work queue using an atomic operation to access the second work queue.

Abstract: Techniques are described for ensuring end-to-end fair-share resource utilization during cross-region replication. In certain embodiments, a fair-share architecture is used for communication among pipeline stages performing a cross-region replication between different cloud infrastructure regions. Cross-region replication-related jobs are distributed evenly from a pipeline stage into a temporary buffer in the fair-share architecture, and then further distributed evenly form the fair-share architecture to parallel running threads of next pipeline stage for execute. Techniques for static and dynamic resource allocations are also disclosed.

Abstract: A systolic array includes a plurality of basic computation units arranged in a matrix. A basic computation includes a feature input register configured to store first feature data, a result buffer configured to store first temporary data, a comparator connected to the feature input register and the result buffer, and a control register connected to the feature input register, the result buffer, and the comparator. The comparator is configured to compare the first feature data input with the first temporary data successively. The control register is configured to control the first feature data of the feature input register and the first temporary data to be input to the comparator, output a comparison result to the result buffer and a feature input register of a next basic computation unit, and after sorting, output the first temporary data last stored in the result buffer as a first data result.

Abstract: A system, method and computer program product to synchronize processing across multiple lanes. In a system, a synchronizing interface network controller (SINC) communicates with a plurality of processors. Each processor executes an application having thread(s) of operation. Each processor notifies the SINC when a specific thread is ready to perform a respective operation. The SINC releases the processors to perform the respective operation upon being notified by all processors that the specific thread is ready to perform the respective operation. Each processor is configured to monitor for the release of the processors and to also determine whether sufficient time remains within a time window to perform the respective operation. If insufficient time remains, a processor notifies the SINC that the specific thread is no longer ready to perform the respective operation. If the processors are released by the SINC while sufficient time remains, each processor performs the respective operation.

Abstract: A system-on-chip includes a processor, a memory and a memory interface coupled to the processor and to the memory. The processor, in operation, generates memory access requests. The memory includes one or more physical banks divided into a succession of sectors, each sector having a size equal to a smallest erasable size of the memory. The memory interface, in operation, responds to receiving memory configuration information by storing logical memory bank configuration information in the one or more configuration registers, the logical memory bank configuration information assigning each sector of the one or more physical banks of the memory to a respective logical memory bank of one or more logical memory banks. The memory interface, in operation, controls access to the memory by the processor based on the logical memory bank configuration information stored in the one or more configuration registers.

Abstract: An apparatus for use by a mobile edge control element or function, the apparatus comprising processing circuitry, and memory for storing instructions to be executed by the processing circuitry, wherein the memory and the instructions are configured to, with the processing circuitry, cause the apparatus to: receive and process an event report indicating that location information of a communication element communicating in a communication network are provided, check location related application information for determining whether or not a relationship of at least one application to a location corresponding to the location information indicated in the event report is present, prepare, in case the check is affirmative, an application identification indication for indicating at least one location area where location information of the communication element is of relevance for a processing conducted by at least one application, and ensure that the communication element is informed about the application identificati

Abstract: Techniques are provided for performing server device updates using update baselines that are tagged across multiple management consoles. One method comprises obtaining, by a first management console of multiple management consoles associated with a server device, a first update baseline comprising a first tag identifier, of multiple tag identifiers, that identifies a first update of a component of the server device across the multiple management consoles; and applying the first update baseline to the server device in a maintenance mode to update the component of the server device, wherein the server device is maintained in the maintenance mode when two or more of the tag identifiers are active in the server device following the application of the first update baseline. The server device can be restored from the maintenance mode when only one tag identifier is active in the server device following the application of the first update baseline.

Abstract: A software agent executing on a computing device receives a request from a client to provide data associated with neighboring devices to the computing device. The client includes a scan engine to perform a network scan of a network that includes the computing device. The software agent accesses device data in a cache of an operating system command, determines, based on the device data, an identifier associated with each device that is neighboring the computing device, converts the device data into a standardized format to create neighboring device data, and sends the neighboring device data to the client.

Abstract: Methods and systems that facilitate automatic generation of Application Programming Interface (API) specification from web traffic. Methods include obtaining a plurality of API requests and responses to the plurality of API requests. Methods include processing these API requests and responses to API requests to identify one or more attributes, such as, for example, variables, query parameters, response status codes, and response schemas. Methods include identifying variables using a tree data structure to represent resource paths. Methods include identifying query parameters based on resource items in resource paths. Methods include determining that the API call does not conform to the API specification by comparing one or more attributes of the API call with the attributes of the API specification.

Abstract: A request to configure a connect cluster including one or more connectors for a cloud computing system is received. The request include a desired connector state. A connector specification file is automatically generated based on the desired connector state for the connect cluster via a declarative application programming interface (API). Application resources associated with the connect cluster are automatically configured based on the specification file.

Abstract: In some examples, a system includes an operating system (OS) kernel, one or more hardware processors, and a kernel scheduler. The kernel scheduler receives a first indication that a first thread is to wait for an event, receives a second indication that the first thread has transitioned to a blocked state in which the first thread is blocked on the event, and receives a third indication from a second thread delivering the event, the third indication to unblock the first thread. In response to the third indication, the kernel scheduler causes a transition of the first thread from the wait state to an active state if the first thread is in the wait state and has not yet transitioned to the blocked state, and causes a transition of the first thread from the blocked state to a ready state if the first thread is in the blocked state.

Abstract: There are provided systems and methods for increasing availability of a micro-computation decision service by utilizing execution flow configurations. A service provider, such as an electronic transaction processor for digital transactions, may utilize different decision services that implement rules and/or artificial intelligence models for decision-making of data including data in production computing environment. A decision service may normally be used for data processing and decision-making through an execution flow configuration and/or graph identifying a flow of task executions and other computing operations.

Abstract: The present invention discloses a fair task offloading and migration method for edge service networks, taking the Pareto optimality of the utility function of all user tasks executed by the edge system as the optimization objective, this approach not only takes into account the constraints of edge network resources, but also ensures the maximization of the utility function of all user tasks in the system, it proposes a new quantitative measurement index for improving the task utility quality under multi-user competition.

Abstract: The present disclosure generally relates to ensuring a data storage device consumes as little power as possible. Different HW modules in the data storage device can operate at different frequencies to ensure any bottleneck HW modules operate at as fast a frequency as possible, while non-bottleneck HW modules operate at slower frequencies and hence, consume less power. The frequency for each HW modules is dynamic and is adjusted based upon detected bottlenecks so that the data storage device can operate as efficiently as possible and consume as little power as possible.

Abstract: In some examples, a system includes one or more hardware processors, an interrupt service thread to handle an interrupt, and a kernel scheduler to schedule the interrupt service thread as part of a collection of threads to be scheduled.

Abstract: Embodiments of the present disclosure may provide dynamic and fair assignment techniques for allocating resources on a demand basis. Assignment control may be separated into at least two components: a local component and a global component. Each component may have an active dialog with each other; the dialog may include two aspects: 1) a demand for computing resources, and 2) a total allowed number of computing resources. The global component may allocate resources from a pool of resources to different local components, and the local components in turn may assign their allocated resources to local competing requests. The allocation may also be throttled or limited at various levels.

Abstract: A method includes, by a scheduling controller, receiving from a user a request for an application to be executed by a computing system associated with a data center, wherein the application includes a plurality of tasks, and wherein the request includes an estimated execution time corresponding to an estimated amount of real-world time that the tasks will be actively running on the computing system to fully execute the application. The method includes receiving from the user a service level objective indicating a target percentage of a total amount of real-world time that the tasks will be actively running on the computing system and generating, in response to determining that the job can be completed according to the service level objective and the estimated execution time, a notification indicating acceptance of the job.

Abstract: Disclosed are a task processing method, equipment, storage medium and device. The method includes: acquiring associated conditions of target tasks, and matching the associated conditions with the target tasks to obtain a matching result; establishing a task association table among the target tasks according to the matching result; acquiring an initial execution sequence of the target tasks, and generating a task matrix according to the initial execution sequence and the task association table; in a determination that the task matrix is not in a preset format, adjusting the initial execution sequence until a task matrix obtained according to the task association table and an adjusted execution sequence meets the preset format, and taking the adjusted execution sequence as a target execution sequence; and taking the target execution sequence as a task planning scheme of the target tasks.

Abstract: Various embodiments are generally directed to an apparatus, system, and other techniques for shared, trusted token generation for a token-bucket flow control policy implemented in a distributed ledger. Tokens may be generated by blocks in the distributed ledger, which may be a blockchain system. Trusted execution hardware may be used as a proof algorithm to ensure that a token exists, and that client has legitimately acquired or accumulated the token prior to using it to submit a transaction.

Abstract: Migrating workloads among execution environments including storage systems includes: selecting a target execution environment for supporting a workload and migrating the workload to the target execution environment utilizing active disaster recovery. Migrating the workload can include: assigning storage resources of the workload to a first pod; linking for replication, the first pod to a second pod of the target execution environment; and replicating the storage resources of the workload to the second pod of the target execution environment.

Abstract: A storage cluster is provided. The storage cluster includes a plurality of storage nodes coupled together as the storage cluster. The plurality of storage nodes is configured to assign data to two or more logical arrays and the plurality of storage nodes is configured to establish data striping across the plurality of storage nodes for user data of each of the two or more logical arrays.

Abstract: A host system can be queried to determine whether new data has been received based on a first time interval. After completion of the first time interval, a determination can be made as to whether the new data has been received and whether a portion of the new data was not stored. In response to the portion of the new data not being stored, the host system can be queried to determine whether subsequent data has been received based on a second time interval where the second time interval is different from first time interval.

Abstract: A method may include receiving a version of software widget source code, storing the version of software widget source code, and determining that a version of application source code and a view of software widget source code are used in a deployment environment. The version of application source code may call the version of software widget source code. The method may further include determining that the version of software widget source code corresponds to the view of software widget source code, constructing a build including the version of application source code and the version of software widget source code, and deploying the build in the deployment environment.

Abstract: A system comprises an interface configured to receive an identifier, a processor configured to determine a grouping associated with the identifier, wherein the grouping is determined using a first clustering, wherein the first clustering is based at least in part on a language processing system, determine a sub-grouping of the grouping associated with the identifier, wherein the sub-grouping is determined using a second clustering, determine a final identifier based at least in part on the identifier and the sub-grouping, determine a resource based at least in part on the final identifier, and store the final identifier associated with the resource, and a memory coupled to the processor and configured to provide the processor with instructions.

Abstract: Example implementations relate to method and system for an action contextual grouping of servers for applying one or more actions to each group. The method includes analyzing data corresponding to a plurality of servers and a management system to generate a dataset including contextual data and server features. The method further includes obtaining a plurality of actions from the management system, and mapping each action to a plurality of criteria to generate an action criteria table. The plurality of criteria includes dynamic contextual criteria derived by the management system based on at least some of contextual data. The method further includes comparing the plurality of criteria for each action in the action criteria table to the contextual data and/or the server features for each server to tag a corresponding server to the action, and grouping the plurality of servers into a plurality of action contextual groups based on the tag.