45 results about "Working set" patented technology

The techniques disclosed herein provide a synchronization engine that operates in conjunction with a service worker to dynamically store and update a working set of user data and single page application (SPA) resources from a network server to a user device. The working set can be hosted across several domains and identified by association with a user account. Accordingly, the synchronization engine retrieves the working set from the network server to enable offline execution of the single page applications. As such, subsequent requests for interacting with a single page application are then serviced by the synchronization engine using the working set retrieved from the network server. For instance, the service worker can bind user data to the application resources to enable progressive rendering through an application controller using locally available resources. In this way, the disclosed system provides a consistent user experience irrespective of network connectivity.

The invention provides a task scheduling and arranging system, method and device and a storage medium, and relates to the technical field of task system management, and the system comprises a gateway cluster module which is used for obtaining a task creation demand of a client; matching in the gateway node according to a task creation demand to obtain a process definition; creating a process instance according to the process definition, and generating a task instruction corresponding to the task creation demand according to the process instance; the work cluster module is used for executing the task instruction through a work node to obtain a real-time state event of task execution; the gateway cluster module is also used for mapping the real-time state event to a process instance and judging whether the execution task is completed or not according to the mapped process instance; and after the task is executed, processing a task result according to the task creation requirement. According to the method, the problems of large delay, weak management capability and the like of task scheduling and arrangement in a large-scale scene in the prior art are solved, and the high requirement of a service system on task scheduling and arrangement can be better met.

Systems and methods for using object storage as a primary storage tier for a file system of a virtual storage system are provided. According to one embodiment, ephemeral storage of a host on which a virtual storage system is running in the cloud is used as an external cache in which all or some subset of file system metadata may be stored, for example, to boost read IOPS and reduce latency of the read path. In various examples, external cache policies may be defined to trigger adding to, updating, and / or re-warming the external cache to accelerate access to file system metadata, thereby making a virtual storage solution that maintains all file system data and file system metadata within object storage a suitable and cost-effective platform for certain types of workloads. In some example, the working set of a workload may also be kept in the external cache to further accelerate read performance of the virtual storage system.

The invention relates to a pointer analysis-based incremental data flow path extraction method. The method comprises the following steps of: firstly, acquiring an initial pointer variable related to an incremental code as a task variable set; then, performing reverse query on the task variables based on the pointing relation constraint graph to obtain a basic variable set and a memory object pointed by the basic variable set; and after the working set is initialized, executing pointing relation propagation, propagating the memory object to a pointing set of the task variable, and classifying a new association pointer found in propagation into an indirect variable set. And executing explicit dependency query on the indirect variable set, adding a new dependency variable into the task set, repeating the process until the constraint graph is stable, and finally obtaining pointing information of all variables. And an inter-process data dependence sub-graph is constructed, and an incremental data flow path is extracted. According to the method, through a locality principle and a reverse query strategy, an analysis range is limited to incremental code related variables, and the overhead and memory problems of global analysis are avoided.

The present invention proposes a method, system, medium, and computer device for optimizing container resource usage overhead. The method includes: recording the files accessed during the execution of the function process corresponding to each container, generating a file working set for each container; retaining only the files in the file working set for a single container and deleting other files, and identifying duplicate files between different file working sets for multiple containers to perform deduplication processing; generating a snapshot file for each container and pausing the operation of the container, identifying duplicate memory pages in each snapshot file to perform deduplication processing; and restoring the container from the snapshot file when a new container needs to be started. This method improves the degree and efficiency of resource sharing between containers on the same server, achieving the effect of reducing container resource overhead and accelerating container startup speed.

The application belongs to the technical field of computer memory management, and discloses an active page replacement method and system based on regional reuse mode identification, which comprises the following steps: working set identification based on memory access sample clustering, obtaining memory access samples by sampling the memory access instructions of a process, and clustering the samples to obtain the working set in the current monitoring window in real time; adaptive memory region division based on working set integration, dividing the address space into regions according to virtual address continuity and adaptively splitting / merging; reuse mode identification based on state transition, determining the reuse mode of each region according to the switching between hot and cold states of each region and the state duration; active page replacement and prefetching method based on the reuse mode, prefeteching the regions that will reach the reuse time, and triggering the replacement of the regions that are in the cooling state for a long time. Compared with the existing passive page fault triggering method, the application can reduce the page fault and jitter, reduce the replacement traffic, and improve the application execution efficiency and system throughput.

The present invention discloses a method for processing timer tasks, a software timer and a storage medium. A first task is provided with a threaded first timer queue. When the first task sets a timer, the timer is set to the first timer queue, and the earliest timer in the first task is set to the timer scanning task; if the timer scanning task detects that the earliest timer in the first task has expired, a timer expiration notification is sent to the first task; the first task scans the first timer queue according to the timer expiration notification. The main timer scanning work is allocated to different tasks to avoid the bottleneck caused by the scanning work being concentrated on one task when the timer is used on a large scale. The threaded idea is adopted to avoid the problem of software timer lock conflict, thereby improving the system throughput. The system is suitable for scenarios where a large number of software timers are used in a multi-core multi-thread concurrent environment.

An information processing device 100 according to the present disclosure comprises: a storage unit 121 that stores a business schedule in which scheduled execution work and scheduled execution time of workers are set; a specification unit 122 that acquires operation data about a specific worker and specifies the execution work and the execution time executed by the specific worker on the basis of the operation data; a detection unit 123 that detects a difference between the specified execution work and execution time of the specific worker and the scheduled execution work and the scheduled execution time set in the business schedule of the specific worker; and a setting unit 124 that, on the basis of the difference, specifies scheduled execution work to be changed among the scheduled execution work set in the business schedule of the specific worker, and sets the scheduled execution work to be changed, to a work schedule of another worker as the schedule execution work of the other worker. As a result, job scheduling is optimized.

Multiple co-resident flush working sets are created in a data storage system. Each flush working set indicates host data that was received by the data storage system from at least one host computing device during a period of time corresponding to that flush working set. Host data indicated by the flush working sets is to be flushed from a storage processor of the data storage system to non-volatile data storage devices of the data storage system, and indications of host data within the flush working sets are organized according to logical block addresses of the indicated host data. Multiple independently executing flusher threads flush the host data indicated by at least one of the flush working sets from the storage processor to the non-volatile data storage devices of the data storage system.

A working set of data blocks to be flushed from a cache to non-volatile data storage is identified, and a physical block of non-volatile data storage is allocated. At least some of the data blocks in the working set are compressed to generate compressed data blocks. The compressed data blocks are stored into the physical block of non-volatile data storage. For each one of the compressed data blocks stored into the physical block, an independent virtual layer pointer indicating a location of that compressed data block within the physical block of non-volatile data storage is stored into a single, flexibly sized virtual layer block that is allocated for storing the working set. The flexibly sized virtual layer block is part of a logical address mapping tree.

[Problem] To provide a work setting system for setting work of a mounted function unit that is mounted on a moving body capable of autonomous movement at a movement destination of the moving body. [Solution] A work setting system according to the present invention comprises: a movable moving body; the mounted function unit that executes a function at a movement destination by being mounted on the moving body; a map display unit that displays, on a user interface, a map that indicates a region in which the moving body can move; a point setting unit that sets a point that constitutes the movement destination of the moving body on the map, said point being the location where the function of the mounted function unit is executed; a point function display unit that displays, for the point, functions that can be executed by the mounted function unit, said functions being displayed in a manner that makes it possible to select a function; and a link function unit that associates the execution of the function that is selected by the point function display unit with the point that is set by the point setting unit.

To provide a cluster management method, an apparatus, an electronic device, and storage medium for improving the applicability of a cluster system through a meta cluster according to actual needs.SOLUTION: A cluster management method comprises: a step for initiating, via a meta cluster, a work cluster adding / deleting task that instructs adding / deleting a work cluster designated in a cluster system, wherein the cluster system includes one or more clusters that include one meta cluster and one or more work clusters, wherein the meta cluster is responsible for operation maintenance and management control for the cluster system, and the operation maintenance and management control include work cluster adding / deleting processing, wherein the work cluster is responsible for a business container and executes a business task; and a step for controlling the designated work cluster to realize the placement or removal of a corresponding cluster in the cluster system based on the instruction information of the adding / deleting task.SELECTED DRAWING: Figure 1

An apparatus comprises at least one processing device that includes a processor coupled to a memory. The processing device is configured to receive in a host device, for a plurality of targets of a storage system, information characterizing particular sets of one or more logical storage volumes accessible to the host device via respective ones of the targets, to compare the received information to a working set maintained by the host device, the working set identifying one or more logical storage volumes utilized by at least one application deployed on the host device, and to control a number of input-output (IO) queues configured in the storage system for each of one or more of the targets based at least in part on a result of the comparison of the received information to the working set. The targets illustratively comprise NVMe targets, and the logical storage volumes comprise respective NVMe namespaces.

Examples of performing tensor operations by a neural network-based computing system, are described. In an example, a first output working set generated by a first operation, wherein the first output working set is a set of processed partitioned tensors, is obtained. The first output working set is then copied to the output working set, for retrieving by the second operation.

A working-set structure is used to organize cached data for storing to persistent storage, which includes leaf structures and page descriptors (PDs) for data pages to be persisted. Upon adding a new PD located in an address range of an existing leaf structure, a PD population count of the existing leaf structure is compared to a predetermined PD population threshold. When the count is below the threshold, the new PD is incorporated into an existing set of PDs for the existing leaf structure, and otherwise (a) a new leaf structure is created, and (b) the new leaf structure is used for the new PD and later-added PDs in the address range. Flush parallelism is enhanced by avoiding large differences in PD population across a set of leaf structures.

The invention discloses an interaction method and device, electronic equipment and a computer readable storage medium, and the method comprises the steps: displaying a time-limited group carrying a group theme through a target interface, and enabling members in the group to share a creation work in a preset group effective time period through the time-limited group; in response to a group participation event for a target time-limited group, displaying a group interaction page of the target time-limited group, the target time-limited group providing an intra-group work set for interaction of intra-group members; and in response to a work sharing event for the target created work, updating the work set in the group and / or a work sharing message sent in the group, and after the current moment exceeds the end moment of the effective period of the time-limited group, sending the work sharing message to the target created work. At least part of the created works in the intra-group works set of the time-limited group are stored and displayed in association with the group theme and / or the effective time period and / or the current account, and the creation passion of the user can be stimulated by newly adding an interaction mode for the created works of the user.

The invention discloses a cross-cluster application scheduling method and device based on K3s and computer equipment, and belongs to the technical field of computers. The method is applied to a centralized control cluster based on K3s, the centralized control cluster comprises a distributed working cluster of a plurality of working clusters, and the method comprises the following steps: when a scheduling request for a first target application is received, in the plurality of working clusters currently in a survival state in the distributed working cluster, scheduling the first target application according to the scheduling request, determining a first target cluster according to the scheduling preference of the first target application; and scheduling the first target application to the first target cluster according to access configuration information pre-reported to the centralized control cluster by the first target cluster. According to the method provided by the invention, the optimal cluster can be automatically screened out according to the scheduling preference in a large-scale distributed environment, the application requirements of large-scale and multi-region deployment can be supported, and the application deployment efficiency is improved.

1. The name of the design product: the working setting graphical user interface of electronic equipment. 2. The use of the design product: for an electronic equipment. 3. The design points of the design product: the graphical user interface. 4. The picture or photo that best shows the design points: design 1 front view. 5. Design 1 is designated as the basic design. 6. The use of the graphical user interface: the present graphical user interface is used for working setting. Design 1 front view to design 8 front view are working setting interfaces, and the user can make relevant settings through touch operation according to the content prompt of the interface. Design 9 front view is a working setting interface, and the user can set more parameters by sliding up in the pop-up window, as shown in the interface change state diagram of design 9.

The application relates to a GPU progressive optimization method for neural operator backbone network reasoning and belongs to the electronic information technical field.The method comprises the following steps: step S1, initializing constraints and tolerances;step S2, obtaining baseline execution time and reference results through baseline measurement;step S3, obtaining an estimated split value based on a working set model and expanding;step S4, performing KSweep evaluation and consistency checking, and preferably obtaining; a process of scanning and measuring each candidate set is called KSweep; step S5, fixing, adding fusion reasoning optimization, and performing consistency verification, if the consistency verification is passed, outputting and saving the optimal configuration, and directly loading and executing in the running stage; otherwise, backtracking or adjusting and recalibrating, and returning to step S3. The application shortens the reuse distance between adjacent layers, reduces the round-trip overhead of intermediate activation between the video memory and the cache, and thus improves the reasoning efficiency.

Innovations are described for training machine learning models using an accelerator pipeline with deferred updates to model parameters. The accelerator identifies one or more first class micro-batches ("MBs") and second class MBs of the working set. The first class of MBs contain, as input, frequent access embedding stored in graphics processing unit ("GPU") memory. The accelerator schedules the first class MB (s) for training using one or more GPUs. During training, the accelerator obtains a second type of MB, which contains as input the infrequent access embedding stored in the main memory. At least some updates to the model parameters from training with the first class MB (s) are delayed until after training with the second class MB. The accelerator schedules a second type of MBs for training. Finally, after training with the second class of MBs, the accelerator updates non-frequent access values for the second class of MBs. At this time, the GPU (s) also update the model parameters, applying delayed update.

Automatically generating definitions of materialized views. Generally, automated MV generation identifies a set of candidate MVs by examining the working set of query blocks. Once candidates are formed, the candidate MVs are further evaluated to compute the benefit to the candidate MVs. An improved method for generating a candidate set of MVs is described herein. The improved method is referred to as the extended covered sub-expression technique (ECSE). Under ECSE, various relationships between join sets other than strict equivalence are used to generate new result join sets. Such relationships include subsets, intersections, supersets, and unions, which are described in further detail below. In some cases, relationships between the result join sets and the initial join sets are considered to generate new result join sets. The final result join sets are then used to form a candidate set of MVs.

The present disclosure provides a cluster management method, device, electronic equipment and storage medium. The cluster management method comprises: initiating a work cluster addition and deletion task by a meta cluster, the work cluster addition and deletion task being used to instruct to add a specified work cluster in a cluster system or to instruct to delete the specified work cluster in the cluster system; wherein the cluster system comprises at least one cluster, the at least one cluster comprising one meta cluster and at least one work cluster, the meta cluster carrying operation and maintenance control of the cluster system, the operation and maintenance control comprising work cluster addition and deletion processing, and the work cluster carrying a business container to execute a business task; and according to indication information of the addition and deletion task, controlling the specified work cluster to implement corresponding cluster deployment or removal in the cluster system. According to the present application, the work cluster can be expanded or deleted according to actual needs through the meta cluster, thereby helping to improve the applicability of the cluster system.

A working set of data blocks to be flushed from a cache to non-volatile data storage is identified, and a physical block of non-volatile data storage is allocated. At least some of the data blocks in the working set are compressed to generate compressed data blocks. The compressed data blocks are stored into the physical block of non-volatile data storage. For each one of the compressed data blocks stored into the physical block, an independent virtual layer pointer indicating a location of that compressed data block within the physical block of non-volatile data storage is stored into a single, flexibly sized virtual layer block that is allocated for storing the working set. The flexibly sized virtual layer block is part of a logical address mapping tree.

According to the method for improving the crosspiece chart, the super crosspiece chart obtained through the method, the terminal product of the super crosspiece chart and the storage medium of the super crosspiece chart, a string of work completed in a natural sequence in the longitudinal axis direction of the crosspiece chart is combined into one work to be expressed, and the work is defined as a work set; a different code name is correspondingly set for each subitem work under the work set to serve as an assignment of each subitem work, and the same type of work of different work sets has the same code name; the code names of the sequential sub-item work of each work sink are recorded on the corresponding position, representing the work process through the time continuation, of the row where the work sink is located in the crosspiece chart, so that the time continuation expressed by the transverse line of the position is replaced. According to the method, the number of longitudinal tables of the crosspiece chart is greatly reduced, adverse effects on vision and psychology caused by huge increase of the number of lines and the number of the tables are eliminated, the application range of the crosspiece chart is greatly expanded, the method is visual, vivid, simple and clear, and key paths of work are easy to find.

The invention discloses a large model reasoning method and system based on active load awareness, and belongs to the technical field of large model reasoning services. Aiming at the problem that reasoning service is difficult to consider resource efficiency and service quality due to burstiness and heterogeneity of mixed loads in the prior art, the invention provides a closed-loop method of cycle planning, online scheduling and self-adaptive adjustment and optimization. In the period planning stage, prospective instance pre-configuration is carried out by utilizing long-term load prediction; in the online scheduling stage, the request is routed and scheduled through response length sensing and a heterogeneous scheduling strategy; and in the self-adaptive tuning stage, fine-grained correction is carried out on the instance scale based on peak working set pre-judgment. According to the method, resource supply and request scheduling are coordinated, the SLO is guaranteed, the delay peak is effectively inhibited, the GPU utilization rate is improved, and the method is suitable for an enterprise-level LMaaS platform.

The present application relates to the technical field of artificial intelligence computing power scheduling, in particular to a load balancing and performance guarantee method for multi-modal AI computing power service, comprising: firstly, analyzing request characteristics, extracting a computing work set and a minimum safe capacity to generate a local key-value cache baseline, and fitting to generate a minimal algebraic arrival curve of network direct writing traffic; then, combining the baseline, the curve and the network topology routing hop count to build a key-value cache non-eviction model and solve the physical cache capacity limit; further, determining the number of isolated paths according to the limit and the single-path capacity, calling the hardware register to set the capacity bitmask, and building a space-time physical isolation architecture; finally, when the traffic breaks through the upper bound of the curve, establishing an orthogonal damping constraint relationship based on equivalent arithmetic strength, and synchronously reducing the computing end micro-batch and the network end transmission unit. The present application solves the problem of physical squeezing of local cache by network direct writing through bottom-layer physical isolation and double-end orthogonal feedback.

Innovations in training a machine learning model using an accelerator pipeline with deferred updates to model parameters are described. An accelerator identifies one or more first-class micro-batches (“MBs”) and a second-class MB of a working set. A first-class MB contains, as inputs, frequently accessed embeddings stored in graphics processing unit (“GPU”) memory. The accelerator schedules the first-class MB(s) for training using one or more GPUs. During the training, the accelerator obtains the second-class MB, which contains, as inputs, non-frequently accessed embeddings stored in main memory. At least some updates to the model parameters from the training with the first-class MB(s) are deferred until after training with the second-class MB. The accelerator schedules the second-class MB for training. Finally, after the training with the second-class MB, the accelerator updates the non-frequently accessed values for the second-class MB. At this point, the GPU(s) also update model parameters, applying deferred updates.

In an information processing apparatus, a work setting information acquisition unit configured to acquire a type of work performed by a movable apparatus, information with respect to a work region, and map information for self-position estimation created in advance, and a starting position and orientation calculation unit configured to calculate a starting position and orientation when the work is started based on the type of the work, the work region, and the map information are provided.

In one embodiment, a computer-implemented method includes receiving, at a wearable mask, first information pertaining to a work setting, wherein the first information comprises video, audio, haptic feedback, or some combination thereof, determining, using an edge processor communicatively coupled to the wearable mask and based on the first information, one or more first characteristics of the work setting, generating, using the edge processor and based on the one or more first characteristics of the work setting, one or more first control instructions configured to modify one or more first operating parameters of a tool, and transmitting, to the tool, the one or more first control instructions to modify the one or more first operating parameters of the tool.

The invention belongs to the technical field of computer memory management, and discloses an active page changing method and system based on regional reuse pattern recognition, and the method comprises the steps: based on the working set recognition of memory access sample clustering, sampling a memory access instruction of a process to obtain a memory access sample, the samples are clustered to obtain a working set in the current monitoring window in real time; carrying out adaptive memory region division based on working set integration, dividing an address space into regions according to virtual address continuity, and carrying out adaptive splitting / merging; identifying a reuse mode based on state transition, and determining the reuse mode of each region according to the switching of each region between cold and hot states and the state duration; according to the active page swap-out and prefetching method based on the reuse mode, a region about to reach the reuse moment is prefetched, and swap-out is triggered for a region in a cooling state for a long time. Compared with an existing passive missing page triggering method, the method has the advantages that missing pages and jitter can be reduced, swap-in and swap-out flow is reduced, and application execution efficiency and system throughput are improved.