66 results about "Physical computing" patented technology

The invention relates to the technical field of distributed computing, and discloses a distributed computing modeling system design method for high-dimensional discrete data, which comprises the following steps: acquiring a feature identifier access request in a high-dimensional discrete data stream, counting an access frequency, and marking a feature identifier of high-frequency access as a high-frequency access state; constructing a local parameter copy on a non-master node physical computing unit of the distributed cluster; routing gradient update data aiming at the feature identifier to a local parameter copy closest to the network and generating a local residual tensor; calculating the direction cosine similarity of the local residual tensor and the global gradient update vector; calculating a real-time access entropy based on the access source distribution and converting the real-time access entropy into a dynamic synchronous threshold value; according to the method, through a dynamic synchronization judgment mechanism based on semantic information gain, high-frequency but invalid redundant communication is effectively inhibited, and network congestion under long-tail distribution is reduced.

Techniques for executing multiple large language models (LLMs) on a single physical computing device, including receiving a task to be performed by at least a first LLM and a second LLM, creating a plurality of virtual compute nodes on the single physical computing device for processing at least some of first set of sub-tasks to be performed by the first LLM and second set of sub-tasks to be performed by the second LLM, wherein each one of the plurality of virtual compute nodes includes one or more graphics processing unit (GPU) cores allocated for performing a sub-task of the first set of sub-tasks or the second set of sub-tasks, and executing at least some of the first set of sub-tasks and the second set of sub-tasks substantially in parallel across the plurality of virtual compute nodes using shared memory. Inferences between the first and second LLMs may be shared.

The invention relates to the technical field of electric digital data processing, and discloses a data processing system for message management, which comprises a look-ahead buffer unit for maintaining a sliding window with variable window size parameters and pre-reading a message sequence; the mapping logic circuit is used for converting the resource identifier into a resource request mask; the state register is used for storing a global occupancy bitmap for representing resource occupancy; the scheduling controller is used for distributing messages according to bit operation results, counting bitmap setting quantity based on the population counting instruction to generate a resource saturation index, and then adjusting window size parameters in a negative feedback mode. The resource contention entropy value is directly sensed by utilizing a hardware instruction, nanosecond-level self-adaptive breathing control of a scheduling window is realized, and the service life of the scheduling window is prolonged. The problem of physical computing power collapse caused by bitmap scanning jitter in a high-concurrency scene is solved.

The disclosure relates to systems, methods and computer readable media for enabling distributed secure remote access from a device via a wide area network (WAN), to a designated physical site covered by local area network (LAN). Specifically, the disclosure relates to a computerized systems, methods and computer-readable media using hardware-based, virtual private network pairs having preshared encryption / decryption keys, operable to transmit data over WAN from a physical computing device to an exclusively designated site in a physical area covered by a LAN.

The invention relates to the technical field of computers, and discloses a computer memory data processing system comprising a calculation task compiling module used for determining physical setting parameters according to a calculation task and generating a preset target evolution path; the physical calculation execution module is used for performing state evolution according to the physical setting parameters; the dynamic feedback control module is used for monitoring the actual evolution state and generating a calibration signal according to the deviation between the actual evolution state and a preset target evolution path so as to dynamically adjust driving; and the result decoding module is used for converting the steady-state physical quantity into a calculation result after the system reaches a stable equilibrium state. By constructing closed-loop feedback control, evolution deviation generated by non-ideal factors of a physical system is corrected in real time, it can be effectively guaranteed that actual state evolution of physical calculation accurately tends to a preset ideal path, and therefore the accuracy and reliability of data processing through the physical process are remarkably improved.

A system and method for efficient tracking and rendering of dynamic entities in video games. In order to avoid costly computational loads associated with simulating physics for each dynamic entity independently, then rendering each entity, the system and method stores the pose of dynamic entities after physics calculations and amalgamates pose data into an optimized data structure to allow a graphics processing unit (GPU) to re-create the pose of and display a plurality of dynamic entities as part of a single rendering pass, referring to only a single set of reference data. The method achieves this by recording the position and rotation of a set of pre-defined bones of importance of a skeleton onto a single location, a single image file, which contains all the information necessary for the GPU to pose and display each of the dynamic entities simply by referencing some pre-defined definitions.

The application relates to the technical field of cloud computing, in particular to a computing resource scheduling method and device, equipment and a storage medium, the method comprises the following steps: constructing a self-defined type of resource, and constructing a corresponding controller according to the self-defined type of resource; when the category of the controller is a virtual machine, a container containing the virtual machine is constructed; according to the type corresponding to the self-defined type of resource, corresponding resource configuration is extracted, and the creation instruction of the container containing the virtual machine is rewritten, the virtual machine is scheduled, and thus a single physical computer can support mixed scheduling of virtual machines and containers simultaneously.

In a computer system with multiple physical computers at different physical locations, a transformer module receives an original query from a client-side computer, analyzes the query statements and annotates the query. The transformer module forwards the annotated query to server-computers. This approach allows a data-analyst 190-ALPHA to use a query that is relatively simple, wherein a data-analyst 190-BETA who does not benefit from the transformer module would have to write a more complex query.

Techniques are described for managing communications between multiple computing nodes, such as for computing nodes that are part of managed virtual computer networks provided on behalf of users or other entities. In some situations, one or more of the computing nodes of a managed virtual computer network is configured to perform actions to extend capabilities of the managed virtual computer network to other computing nodes that are not part of the managed virtual computer network, such as by forwarding communications between computing nodes of the managed virtual computer network and the other external computing nodes so as to enable the other external computing nodes to participate in the managed virtual computer network. In some situations, the computing nodes may include virtual machine nodes hosted on one or more physical computing machines or systems, such as by or on behalf of one or more users.

A device may receive, via executable instructions of an electronic document running on a physical computing device, a request from an automated auditing system to audit data of the electronic document. A device may process, via the executable instructions of the electronic document running on the physical computing device, the request to audit the data of the electronic document. A device may deliver, to the automated auditing system and from the electronic document via the executable instructions running on the physical computing device, data in a format that is native to the electronic document and auditable by the automated auditing system in a state in which it is stored in the electronic document.

The application provides a kind of industrial cloud platform computing resource dynamic scheduling method and system combined with deep learning, it is related to industrial cloud platform technical field, first, the real-time request queue parameter set and real-time container resource parameter set of containerized application instance set in industrial cloud platform are acquired, then correlation coupling analysis is carried out to generate service quality sensitivity descriptor set, then call deep learning request flow prediction network to identify request flow evolution mode, generate predicted request timing distribution set in future prediction time window, according to the node available resource parameter set of the two sets above Resource reservation dynamic programming is carried out to physical computing node cluster, generate dynamic binding relationship mapping set, finally, based on the dynamic binding relationship mapping set Migration scheduling instruction sequence is generated and issued to container orchestration scheduler, to realize the dynamic accurate scheduling of computing resource.

The embodiment of the invention discloses an interface interaction control method, a computer system, equipment and a computer storage medium. According to an interface type of a target virtual interface indicated by the creation request, creating a target virtual computing power device corresponding to the interface type in a target virtual environment, responding to a deep learning request of the target virtual environment through the target virtual computing power device, and forwarding the request to a host machine, and scheduling a physical computing power device corresponding to the interface type to respond to the request, and returning a processing result of the request to the target virtual environment through a target virtual computing power device. According to the method, virtualization is carried out from a software interaction level instead of virtualization of computing power hardware, the scheme is completely realized based on software, and a physical GPU is not required to have any special virtualization hardware function, so that almost all mainstream and even old models of GPUs on the market and accelerators such as NPUs with different architectures can be managed. And the user does not need to pay for expensive virtualization dedicated hardware and software license, so that the virtualization deployment cost is reduced.

Components of a distributed database may operate faster by maintaining connections to multiple other components. These components may be hosted on various physical computing devices. An application layer communication system comprising a network proxy on each computing device involved in the distributed database may decrease the overall number of connections that are to be established and maintained during operation of the distributed database. The network proxies at sending devices may reduce traffic across the connections by combining packets directed from a particular device to another.

A nonlinear physical computing architecture comprising a continuous nonlinear domain in which excitation produces persistent reshaping of at least one spatially continuous parameter field. Said reshaping alters non-local coupling coefficients of governing nonlinear equations and modifies attractor basin topology. Subsequent excitation follows a measurably reduced state-space trajectory length relative to baseline measured following reset. Persistence magnitude may be continuously modulated, and multiple domains may be orchestrated in staggered reset configurations. The architecture excludes discrete memory cells and programmable weight matrices.

The invention provides a computing power aggregation method and device of a common computing power intelligent computing cloud platform, and relates to the technical field of intelligent computing centers, intelligent computing centers and computing power infrastructure, the intelligent computing cloud platform comprises a logic computing power cluster and a plurality of physical computing power clusters, each physical computing power cluster comprises a plurality of computing power nodes, and each physical computing power cluster comprises a plurality of computing power nodes. The method comprises the steps that a logic computing power cluster synchronizes information of each computing power node in a physical computing power cluster in real time, the computing power nodes are mapped into virtual computing power resources, and the virtual computing power resources are put into a computing power resource pool; receiving a container creation request or a calculation task creation request of a user, and creating a container based on the request; based on the information of each virtual computing power resource in the computing power resource pool, screening out a target virtual computing power resource matched with the container from the computing power resource pool; and determining a target computing power node corresponding to the target virtual computing power resource, and scheduling the target computing power node operation container through a physical computing power cluster where the target computing power node is located.

The invention relates to a vGPU performance interference perception-based large model reasoning load resource allocation method, which comprises the following steps of: giving a group of large model reasoning loads as well as allowable maximum first Token response time TTFT, maximum average single Token generation time TPOT and request arrival rate of the large model reasoning loads; obtaining a physical GPU required by the reasoning load, and calculating the minimum proportion of vGPU resources required by running the reasoning load in the physical GPU; taking as an initial allocation benchmark; carrying out descending sorting on the large model reasoning load according to vGPU resource requirements; for each to-be-allocated load, traversing the existing candidate GPU equipment, and dynamically allocating resources through an iterative adjustment mechanism on the premise of ensuring that the allocated resources of the current physical GPU do not exceed the maximum capacity of the current physical GPU; and if the existing device cannot accommodate the new load, expanding a new GPU device. According to the method, the vGPU resource scale required by the reasoning task can be accurately evaluated.

A system including an unmanned aerial vehicle. The unmanned aerial vehicle includes image sensors and one or more physical computer processors. The image sensors are configured to capture images and / or video segments. The one or more physical computer processors create a user ID and / or other identifying information for a user and / or a consumer when a user and / or consumer registers with an authentication component. The processors associate the user ID and / or the other identifying information with one or more client computing platforms. The processors store the user ID and / or the other identifying information associated with the user and / or the consumer on the authentication component. The processors associate the user and / or the consumer with one or more accounts. The processors authenticate the user and / or the consumer with the authentication component so that the user and / or the consumer is permitted to access a repository of the images and / or the video segments.

The invention relates to a cooperative control-oriented end side network trust evaluation method and system. The method comprises the following steps of: configuring an information interaction topology for an end side node and constructing a corresponding dynamic model; establishing a feature description mechanism for neighbor end-side nodes of the end-side network, and depicting the gain of neighbor end-side node information on collaboration, the stability of the behavior of the neighbor end-side nodes and the influence of the neighbor end-side node information on constraints; setting a response mechanism, realizing global trust iteration of the end side network, and generating a global trust evaluation result; the system comprises one or more nodes arranged in an end-side network, each node is configured with an end-side agent module deployed on an independent physical computing unit, and the end-side agent module comprises a processing module for executing the method, a communication module and a cooperative control execution module. According to the method, the behavior credibility of the neighbor end-side network in the cooperative task is reflected more accurately, the capability of resisting abnormal data propagation and burst interference is remarkably improved, and the method can be efficiently operated in a resource-limited end-side network system.

Technology for a mobile prefabricated factory for constructing a regional data center is disclosed. This mobile prefabricated factory may comprise a storage enclosure configured to house the physical computing resources of the data center, a networking device, a power supply electrically connected to the networking device, and multiple computing devices of physical computing resources that are communicated to the networking device and electrically connected to the power supply. A management service can configure the computing devices for delivery to the target site by implementing a seed server device for the multiple computing devices and implementing a software resource repository in the seed server device. During transport of the storage enclosure, the seed server device can deploy software resources to the multiple computing devices.

The invention provides a load balancing processing method and device and related equipment, and the method comprises the steps: monitoring physical computing nodes for data processing in a system, and determining the physical computing nodes with data hotspots; adding a new virtual node in the target idle physical computing node; generating an adjusted physical execution graph according to the new virtual node; and according to the physical execution graph, transferring a data processing task on the physical computing node with the data hot spot to the corresponding virtual node for processing. According to the load balancing method provided by the invention, the data processing task on the physical computing node where the hotspot data appears is transferred to the newly added virtual node on the idle physical computing node for processing, so that a streaming data processing scene taking an event as a unit, such as invoice data processing, can be dealt with in time; the problem of data skew in the Flink engine is effectively solved, the complexity of a program is low, and the implementation cost is low.

The invention relates to the technical field of intelligent fire fighting and building information model application, and discloses a BIM-based basement fire fighting integrated system and method and electronic equipment. The system comprises a BIM operation and maintenance database module, a multi-source signal receiving and mapping module, a fire simulation and active environment control module, a dynamic path planning and navigation module and an intelligent resource scheduling module. Real-time fire data are constructed through multi-source signals and BIM space mapping, and the necessary air supply amount needed for maintaining the positive pressure of an escape channel is calculated on the basis of the fluid mechanics principle in combination with air leakage parameters of a building envelope so as to regulate and control heating and ventilation equipment; meanwhile, a dynamic weight updating mechanism based on positive pressure establishment feedback is constructed, and a person is guided to avoid a positive pressure failure area; and when the capacity of a local fan is insufficient, a cross-regional resource coordination or personnel density preemption strategy is executed based on air volume gap analysis, and active smoke prevention and closed-loop feedback evacuation navigation driven by physical calculation are realized.

Provided herein are various enhancements for deployment of workloads or jobs on a computing cluster. In one example implementation, a method includes identifying a container pod deployment request having a pod specification, and responsive to the container pod reaching a pending state for insufficient resources to support deployment of the container pod on a computing cluster, identifying resources indicated in the pod specification and determining one or more physical computing components to attach to a target node. The method also includes attaching the one or more physical computing components to the target node, where a change in resources available in the target node is detected by a workload manager that deploys the container pod.

The application relates to a method and system for trust evaluation of an end-side network facing cooperative control, wherein an information interaction topology is configured for an end-side node, and a corresponding dynamic model is constructed; a feature description mechanism for neighbor end-side nodes of the end-side network is established, the gain of neighbor end-side node information on cooperation, the stability of the behavior of the neighbor end-side node itself and the influence of the neighbor end-side node information on constraints are described; a response mechanism is set, global trust iteration of the end-side network is realized, and a global trust evaluation result is generated; the system comprises one or more nodes arranged in the end-side network, each node is configured with an end-side intelligent agent module arranged on an independent physical computing unit, and the end-side intelligent agent module comprises a processing module for executing the method, a communication module and a cooperative control execution module. The application more accurately reflects the behavior credibility of neighbor end-side networks in a cooperative task, significantly improves the ability to resist abnormal data propagation and sudden interference, and can efficiently operate in a resource-limited end-side network system.

A method including: receiving a single usage value for a plurality of physical computing resources over a period of time; receiving a cost for the plurality of physical computing resources; and determining, using one or more processors, a normalized usage rate cost of the plurality of physical computing resources over the period of time by normalizing the cost for the plurality of physical computing resources for a unit of the single usage value

The invention discloses a physical signal sensing method and system approaching a data processing inequality limit, and belongs to the technical field of intelligent sensing and information processing. According to the method, a data processing inequality in an information theory is taken as a theoretical basis, time, space and frequency multi-dimensional physical waveforms are designed to improve an information upper limit, and a physical calculation model is utilized to carry out end-to-end direct feature extraction and classification identification on the received original physical waveforms. The preprocessing steps of generating images, time-frequency spectrums and other intermediate representations are completely omitted, so that information loss caused by information conversion in a traditional perception normal form is avoided, and the theoretical performance limit specified by a data processing inequality is approached. The physical calculation model can be realized through software or special hardware, and has the advantages of high processing precision, low delay and low power consumption. According to the method, a traditional recognition normal form based on traditional image processing is broken, and an innovative solution is provided for a next-generation high-precision intelligent sensing system.

The invention relates to the field of distributed computing, and particularly provides a distributed computing power grid-connected platform and a self-adaptive load balancing method thereof, and the method comprises the steps that a physical resource layer is a platform hardware foundation and carries out physical computing, storage and network resource use; the resource abstraction layer abstracts and standardizes the hardware of the physical resource layer, shields the bottom layer hardware difference, and converts the physical resources into scheduling and management logic resources; the core scheduling layer carries out intelligent scheduling and dynamic optimization of computing power resources based on the logic resources provided by the resource abstraction layer in combination with a self-adaptive load balancing mechanism; the application interface layer provides standardized and diversified access modes; the application scene layer is an actual application direction of platform computing power resources, and the platform performs scene application. Compared with the prior art, the application access threshold can be reduced, multi-field scene landing is supported, and cross-industry computing power collaboration and innovation are promoted.

An example computing device accesses a resource allocation network service utilizing a wireless network connection in accordance with embedded SIM (eSIM) functionality to identify and reserve computing resources. The computing device transmits a request to the resource allocation server that includes computing device profile information stored in the eSIM. The resource allocation network service processes the requests and receives additional profile information from the identified physical computing resource. The resource allocation network service then establishes a reservation and generates allocation information that includes credential information and is sent to the requesting computing device and the physical computing resource. When the computing device attempts to access the computing resource (e.g., the workstation), the physical computing resource authenticates the computing device based allocation information transmitted by the computing device.

This will allow users to transparently specify the computing resources on which to run deep learning applications. [Solution] When a call is made to the deep learning framework 30 during the execution of a deep learning application 20, a first identifier that identifies the computing resources described in the application 20 is obtained, and based on the first identifier, a second identifier corresponding to the first identifier is obtained by referring to the correspondence relationship 105 between the first identifier and a second identifier that identifies the physical computing resources managed by the framework 30, the first identifier included in the call is replaced with the obtained second identifier to generate a new call, and the new call is sent to the framework 30.

The application relates to the technical field of reactor physics calculation, and discloses a particle simulation transport system and method for simplifying an input process, the system comprising: a front-end modeling system configured to introduce a WebGL-based Three.js library as a 3D engine, perform three-dimensional modeling, and generate a particle simulation model; wherein the particle simulation model at least comprises a geometric body, a radiation source and a detector; a front-end parameter setting system configured to set MCNP parameters of the particle simulation model; and a back-end calculation system configured to introduce a Django framework to constitute a network server, convert the particle simulation model and the MCNP parameters into an MCNP input file by using the Django framework, and perform simulation calculation according to the MCNP input file. The application models the particle simulation model in the Three.js library, sets parameters for the model, converts the particle simulation model and the corresponding parameters into the MCNP input file, simply generates the MCNP input file, and thus can perform MCNP simulation calculation.

The application discloses an intelligent network anomaly monitoring method and device and a storage medium, relates to the field of communication, and is used for improving the efficiency of intelligent network anomaly monitoring, and comprises the following steps: acquiring time series data corresponding to key performance indicators of physical computing nodes and virtual computing instances; performing window slicing on the time series data in the time dimension, extracting multi-dimensional time series features from each data window to obtain a feature vector; determining a target prediction model from a plurality of prediction models according to the time characteristics of the abnormal mode corresponding to the time series data; analyzing the feature vector based on the target prediction model, determining a performance indicator prediction sequence of a target time window, and determining an abnormal probability score according to the deviation degree of the performance indicator prediction sequence and a preset normal baseline. The application is applied to the process of intelligent network anomaly monitoring.