Abstract: Systems and methods for processing computing jobs of a managed network are disclosed. Each of one or more worker nodes may implement a scheduler thread and a pool of worker threads. Upon waking up from a sleep state, the scheduler thread may determine a current number of jobs in an in-memory job queue that are waiting for processing by a worker thread, and may compute a job-completion rate of jobs processed by threads of the pool. Based on the job-completion rate, the scheduler thread may perform one or more of retrieving more jobs from a centralized database job queue and adding them to the in-memory job queue; removing one or more jobs from the in-memory job queue and returning them to the database job queue; leaving the in-memory job queue unchanged; or adjusting the duration of the sleep-interval timer. The scheduler thread may then return to a sleep state.

Abstract: Dynamic scheduling is disclosed. A plurality of physical nodes is included in a computer system. Each node includes a plurality of processors. Each processor includes a plurality of hyperthreads. In response to receiving an indication of an event occurring, a search is performed for a queue in a set of queues on which to place a virtual processor that had been waiting on the event. Queues in the set of queues correspond to hyperthreads in a physical node in the plurality of physical nodes. The queues in the set of queues are visited according to a predetermined traversal order.

Abstract: Examples may include techniques to decrease a live migration time for a virtual machine (VM). Examples include selecting data to copy or not copy during a live migration of the VM from a source host server to a destination host server.

Abstract: Hierarchical stalling strategies are disclosed. An indication is received of a stalling event caused by a requested resource being inaccessible. In response to receiving the indication of the stalling event, a set of cost functions usable to determine how to handle the stalling event is selected based at least in part on a type of the stalling event. The stalling event is handled based at least in part on an evaluation of the set of cost functions selected based at least in part on the type of the stalling event.

Abstract: A system for managing the end to end lifecycle of a cloud-hosted desktop virtualization environment within a public cloud service such as Microsoft Azure, Amazon AWS, includes a set of installation tools and an administration portal. A reference architecture for a cloud-hosted virtualization environment defines technology layers. A cloud orchestrator deploys public cloud tenants. A setup wizard tool installs and configures management software components, infrastructure software components, and application images. An environment management module completes the installation and configuration of the management software components of a virtualization environment and provides ongoing management of the infrastructure software and management software components. An image management module stores the configuration of the application images and provides ongoing management of the application images. A catalog management module maintains a library of the application images.

Abstract: A system includes a production cluster with a first plurality of nodes. The production cluster executes a workload. Jobs associated with the workload are allocated, according to a first configuration, across the first plurality of nodes. A workload simulator is coupled to the production cluster and a test cluster. The workload simulator extracts production cluster data, which includes production capability information, workload data, and production cluster usage information, as well as test capability information. The workload simulator determines a first job type to include in a simulated workload to be executed on the test cluster and a number of jobs of the first job type to include in the simulated workload. The system also includes a test cluster which includes a second plurality of nodes. The second plurality of nodes includes fewer nodes than does the first plurality of nodes. The test cluster executes the simulated workload.

Abstract: A workload/processor resource scheduling system is coupled to a processing system. The workload/processor resource scheduling system monitors a performance of first workload(s) by the processing system according to a workload/processor resource schedule, and identifies a correlation between the performance of the first workload(s) according to the workload/processor resource schedule, and an operating level of a processing system operating parameter for the processing system when performing the first workload(s) according to the workload/processor resource schedule. Based on the correlation, the workload/processor resource schedule and the processing system operating parameter are linked.

Abstract: Approaches for managing how the passage of time is observed by a software execution environment, such as a virtual machine or a sandbox environment. A computer system maintains a set of physical time sources. A set of virtual time sources are computed based on the set of physical time sources. The virtual time sources operate independently of the set of physical time sources. For example, the virtual time sources may observe time passing faster or slower than the set of physical time sources. The set of virtual time sources are presented to the software execution environment as the set of time sources. Many benefits may be obtained such as higher utilization of allocated resources and avoidance of timeouts.

Abstract: Disclosed are aspects of memory-aware placement in systems that include graphics processing units (GPUs) that are virtual GPU (vGPU) enabled. Virtual graphics processing unit (vGPU) data is identified for graphics processing units (GPUs). A configured GPU list and an unconfigured GPU list are generated using the GPU data. The configured GPU list specifies configured vGPU profiles for configured GPUs. The unconfigured GPU list specifies a total GPU memory for unconfigured GPUs. A vGPU request is assigned to a vGPU of a GPU. The GPU is a first fit, from the configured GPU list or the unconfigured GPU list that satisfies a GPU memory requirement of the vGPU request.

Abstract: Embodiments of the present disclosure disclose an artificial intelligence chip and an instruction execution method for an artificial intelligence chip. A specific embodiment of the artificial intelligence chip includes: an instruction memory, a data memory, at least one general execution unit, and at least one dedicated execution unit. The instruction memory is configured to: receive a kernel code including at least one code block. The general execution unit is configured to: receive the code block, lock the dedicated execution unit associated with the received code block, and send an instruction in the received code block to the locked dedicated execution unit. The dedicated execution unit is configured to: execute the received instruction, and store an execution result in the data memory. The data memory is configured to: store the execution result sent by the dedicated execution unit.

Abstract: Distributed timer task execution management is disclosed. A cluster member generates a first timer task that can be executed on any cluster member of a plurality of cluster members including the first cluster member that composes a cluster. A first timer task schedule that identifies at least one future point in time at which the first timer task is to be executed is generated. A second cluster member of the plurality of cluster members is selected as a cluster member owner for the first timer task that is to schedule the first timer task and to execute the first timer task at the at least one future point in time. The first timer task and the first timer task schedule are transferred to the second cluster member.

Abstract: Approaches for managing how the passage of time is observed by a software execution environment, such as a virtual machine or a sandbox environment. A computer system maintains a set of physical time sources. A set of virtual time sources are computed based on the set of physical time sources. The virtual time sources operate independently of the set of physical time sources. For example, the virtual time sources may observe time passing faster or slower than the set of physical time sources. The set of virtual time sources are presented to the software execution environment as the set of time sources. Many benefits may be obtained such as higher utilization of allocated resources and avoidance of timeouts.

Abstract: Computing resources are provided to a user by identifying applications used by the user, and provisioning virtual computing resources that are adapted to the resource requirements of the identified applications. The resource requirements of the identified applications can be combined into a single set of resource requirements and used to acquire a virtual machine that is able to host the identified applications. In other examples, virtual machines may be acquired for each identified application. Each virtual machine generates a display stream via a streaming agent. The display stream is received by an application streaming client on the user's client computer system, and is displayed to the user on a client display. Multiple virtual machines may generate multiple display streams which can be combined by the application streaming client and presented to the user on the single client display.

Abstract: System and method for terminating instances and autoscaling instance groups of computing platforms. For example, a method includes determining whether an instance of an instance group is identified as eligible for termination. The method further includes, in response to determining that the instance of the instance group is identified as eligible for termination, terminating the eligible instance. The terminating the eligible instance includes, in response to a runtime of the eligible instance being equal to or larger than a predetermined maximum lifetime, terminating the eligible instance.

Abstract: A technique is described herein that intelligently deploys resources in a data center for a new program. The new program has, at least in part, unknown runtime characteristics. The technique involves collecting plural input factors that provide evidence of an expected runtime behavior of the new program. It does so by identifying at least one related program that differs from the new program, but has a specified degree of relatedness to the new program. The collecting operation then obtains information that describes an amount of resources that the related program(s) consume when run. Based on at least some of the plural input factors, the technique generates an estimated amount of resources that the new program is expected to consume when it is run. The technique then determines and deploys a configuration of resources in the data center that will provide the estimated amount of resources.

Abstract: Techniques are provided for adaptive resource allocation for workloads with initial condition setting. One method comprises obtaining a dataset comprising data from previous executions of a workload, wherein the data comprises a plurality of different resource allocations and parameterizations of the workload; determining an initial allocation of an amount of a resource for the workload based on a regression model characterizing a behavior of the workload, the data, a predefined service metric and a characterization of a target infrastructure; and initiating an application of the determined initial allocation of the amount of the resource for the workload. A performance of one or more of the plurality of workloads can be evaluated based on a percentage of time within a predefined error range. The regression model can be updated and/or replaced over time with new data for additional executions of the at least one workload.

Abstract: A technique for determining when it is safe to use scheduler lock-acquiring wakeups to defer quiescent states in real-time preemptible read-copy update (RCU). A determination may be made whether a deferred quiescent-state reporting request that defers the reporting of an RCU quiescent state on behalf of a target computer task is warranted. If so, it may be determined whether a previous deferred quiescent-state reporting request on behalf of the target computer task remains pending. A request may be issued for deferred quiescent-state report processing that reports a deferred quiescent state. The request for deferred quiescent-state report processing may be issued in a manner selected according to a result of the determining whether a previous deferred quiescent-state reporting request remains pending.

Abstract: Disclosed methods, systems, and apparatus, include computer programs encoded on computer storage media, for performing allocation of M resources among N users into K pools by solving a knapsack problem (KP) using a distributed computing system. The method includes: receiving data representing K global constraints and L local constraints of the KP; transforming the KP into a dual problem using K dual multipliers; decomposing the dual problem into N sub-problems; performing two or more iterations in solving the dual problem, wherein in one iteration, for each dual multiplier corresponding to a global constraint corresponding to a pool: determining an updated dual multiplier for the global constraint corresponding to the pool to be a non-negative threshold; and computing M decision variables of each of the N users corresponding to the updated dual multiplier in solving each of the N sub-problems corresponding to the each of the N users.

Abstract: Executing a big data analytics pipeline in a storage system that includes compute resources and shared storage resources, including: receiving, from a data producer, a dataset; storing, within the storage system, the dataset; allocating processing resources to an analytics application; and executing the analytics application on the processing resources, including ingesting the dataset from the storage system.

Abstract: Roughly described, a task control system for managing multi-scale simulations receives a case/task list which identifies cases to be evaluated, at least one task for each of the cases, and dependencies among the tasks. A module allocates available processor cores to at least some of the tasks, constrained by the dependencies, and initiates execution of the tasks on allocated cores. A module, in response to completion of a particular one of the tasks, determines whether or not the result of the task warrants stopping or pruning tasks, and if so, then terminates or prunes one or more of the uncompleted tasks in the case/task list. A module also re-allocates available processor cores to pending not-yet-executing tasks in accordance with time required to complete the tasks and constrained by the dependencies, and initiates execution of the tasks on allocated cores.