108results about How to "Reduce performance loss" patented technology

The invention relates to the technical field of server heat dissipation, in particular to a cloud-based server heat dissipation method and system, and the method comprises the following steps: monitoring the key operation parameters of a server in real time through a cloud platform, updating the state of a heat dissipation unit according to the heat dissipation demands monitored in real time, optimizing the heat dissipation efficiency, and obtaining the optimal configuration of the heat dissipation parameters. Through real-time monitoring and data-driven heat dissipation control, heat dissipation efficiency and energy consumption management are remarkably optimized, dynamic heat dissipation unit state updating and optimization are introduced, so that heat dissipation configuration can respond to real-time load and temperature changes of the server in real time, and by continuously adjusting feedback control logic including the feedback period and strength, the heat dissipation efficiency of the server is improved. The influence of the state of the heat dissipation unit can be simulated and predicted, the optimal configuration of the heat dissipation parameters is realized, the energy utilization of the server is optimized, the optimal matching between the heat dissipation control and the server load is ensured, the energy consumption is reduced, and meanwhile, the heat dissipation control efficiency is improved.

The invention relates to an environment energy management optimization method and system for a tag chip, and the method comprises the steps: obtaining the temperature data and environment electromagnetic data of a chip environment, and carrying out the spatial distribution analysis, and obtaining the corresponding environment parameter characteristics; acquiring real-time power consumption data of the tag chip, and performing correlation analysis on the real-time power consumption data and the environmental parameter characteristics to obtain corresponding energy consumption information; acquiring chip circuit acquisition data of the tag chip, and performing resonance frequency analysis on the chip circuit acquisition data according to a preset resonance optimization algorithm; performing fault diagnosis on the chip circuit acquisition data to obtain corresponding system state information and a corresponding system operation prediction scheme; and performing strategy optimization on the initial energy scheduling scheme according to the system operation prediction scheme to obtain a corresponding global optimization strategy. According to the invention, the system can flexibly cope with energy demand changes of different application scenes.

The invention provides a message processing method and device based on a message queue, equipment and a medium, and relates to the technical field of computers. The method comprises the steps that two message queues are created in advance and recorded as a first message queue and a second message queue respectively, one of the first message queue and the second message queue serves as a producer queue, and the other one serves as a consumer queue; controlling a producer thread to operate a producer queue according to a producer rule; and controlling the consumer thread to operate the consumer queue according to the consumer rule. Compared with the prior art, two independent message queues are created in advance, and the two message queues can be provided with independent lock mechanisms, which means that lock operation of a producer and a consumer is completely decoupled, and performance loss caused by lock competition can be remarkably reduced, so that a high-performance thread security queue is realized, and the safety of a thread is improved. And the harsh requirements of a modern large-scale distributed system on the data processing speed and the concurrency capability are met.

An electric pump pressure type liquid oxygen methane rocket main and auxiliary power integrated system adopts an electric pump to pressurize a low-pressure propellant in a rocket storage tank, the pressurized propellant can be controlled to enter an attitude and orbit control engine or return to the rocket storage tank through reversing of a reversing valve, and if the reversing valve continuously supplies the propellant into the attitude and orbit control engine, the propellant can enter the attitude and orbit control engine; continuous and stable work of the attitude and orbit control engine can be realized; if the reversing valve supplies propellants into the attitude and orbit control engine and then immediately switches to the rocket storage tank, pulse work of the attitude and orbit control engine can be achieved through circulation. The reversing valve can keep the electric pump in a normal circulation state at the downstream moment, and the electric pump can work normally and stably no matter whether the attitude and orbit control engine works stably or in a pulse mode. The cavitation pipe is installed in front of the reversing valve, when the attitude and orbit control engine works in a pulse mode, the rotating speed and the power of the electric pump are kept unchanged, the pressure and the flow behind the pump can be constant, the flow entering the attitude and orbit control engine is constant, and it can be guaranteed that thrust generated by the attitude and orbit control engine in each pulse work is constant.

The invention discloses an automatic coupling method and device for a photoelectric detector wafer, and the method comprises the steps: employing a visual recognition technology to achieve the coarse positioning of the coupling of a to-be-detected wafer chip, taking an optical power count value collected in real time as a search signal, carrying out the real-time analysis of the signal through combining with a breadth-first search algorithm, and carrying out the detection of the signal. Path planning is carried out on a three-dimensional space to determine a search path, precise control of a motion path is carried out on the three-axis precision platform through a software control system, finally, optical power value signals under each path position are traversed to find out a maximum value signal, the maximum value signal is moved to the position, and the maximum value signal is the optimal coupling position. Therefore, automatic coupling of wafers in a limited space is realized. The wafer coupling device has the beneficial effects that the automation degree of wafer coupling can be improved, the manual operation cost and the time cost are reduced, meanwhile, the precision and the stability of wafer coupling are improved, and adverse effects caused by manual operation errors are reduced.

The invention discloses an anti-attack lightweight defense method for a malicious client in federated learning, and the method comprises the steps: calculating a gradient parameter # imgabs0 # of a local model after a client i completes the local training of a current round, and uploading the gradient parameter # imgabs0 # to a server; the server extracts a gradient parameter # imgabs2 # of an output layer from the gradient parameter # imgabs1 #, calculates a gradient parameter change amplitude # imgabs3 #, updates a gradient parameter change matrix M, finds out an outlier client with inconsistent gradient parameter update, and determines the outlier client as a malicious client; and aggregating the gradient parameter # imgabs4 # uploaded by the benign client, updating global model parameters based on the aggregated gradient parameter Gt, issuing the global model parameters to all clients, updating local model parameters by the clients, adding 1 to the number of iterations, and starting t + 1 turns of iterations. By introducing a reputation point mechanism, different weights are distributed during client gradient aggregation according to historical manifestation of each client, and the method reduces model performance loss caused by misjudgment of benign clients while punishing malicious clients.

The invention belongs to the field of light-emitting diodes, and relates to a light-emitting diode based on a core-shell composite structure hole blocking layer and a preparation method thereof.The light-emitting diode comprises a cathode substrate, the hole blocking layer, a porous insulating layer and a porous anode, wherein the pores of at least one of the hole blocking layer, the porous insulating layer and the porous anode are also filled with a light-emitting active material; the hole blocking layer is composed of a composite material with a core-shell structure, the composite material with the core-shell structure adopts a conductive material as a core layer and a charge transport material as a shell layer, the porous core-shell hole blocking layer is used for replacing a common charge transport material, efficient composite luminescence of current carriers can be achieved, and the hole blocking layer is a porous hole blocking layer. And high-efficiency transmission of charges in a bulk phase can be ensured, and a high-performance light-emitting diode device is realized.

The invention relates to a resource request processing method and device, computer equipment, a computer readable storage medium and a computer program product. The method is applied to a first cache server in a cluster, and comprises the following steps: in response to a received resource request forwarded by a gateway, extracting first identification information of a resource carried by the resource request, and updating request load data corresponding to the first identification information; under the condition that the request load data is greater than a preset threshold value, determining the resource as a hotspot resource, and adding second identification information to the resource request; and sending the hotspot resource to a second cache server except the first cache server in the cluster, and returning the resource request carrying the second identification information to the gateway, so that the gateway forwards the resource request to the second cache server for processing based on the second identification information. By adopting the method, the access quality of the CDN service can be improved.

The invention provides a cross-medium coaxial double-rotor unmanned aerial vehicle with foldable propellers and a control method of the cross-medium coaxial double-rotor unmanned aerial vehicle. The unmanned aerial vehicle comprises a main fuselage module, a vector motor base module, a coaxial propeller power module, a foldable propeller module, a waterproof sealing module and electronic equipment. According to the cross-medium power system, by folding the propellers, the effective area of the propellers is reduced, the influence of different working media of the cross-medium unmanned aerial vehicle on the efficiency of the propellers is fundamentally reduced, the capacity loss is reduced, and the endurance time is prolonged; the number of actuators is reduced, a complex gear structure is omitted, the weight of the unmanned aerial vehicle is reduced, the reliability of the unmanned aerial vehicle is improved, finally, control codes suitable for the cross-medium coaxial double-rotor unmanned aerial vehicle are designed, and automatic control over the rotating speed of motors and folding of propellers of the unmanned aerial vehicle under different working media are achieved. And the power distribution of the unmanned aerial vehicle in the cross-medium task is more reasonable.

This invention relates to the field of computer technology, specifically to a method and apparatus for integrating third-party algorithm components based on dual-core collaboration. This application directly integrates third-party algorithm components into the BCM dual-core MCU, eliminating reliance on ZCU and CAN bus communication. This fundamentally eliminates the latency and security overhead caused by cross-ECU transmission and E2E protection, improving signal real-time performance and communication security. By clearly defining the functions and allocating resources of the BCM dual-core, a dedicated storage area is planned for the third-party algorithm components, and a standardized integration environment is built, achieving efficient and stable component integration. Simultaneously, peripheral access strategies are optimized, unnecessary protections and redundant polling are disabled, releasing idle computing power and hardware resources of the BCM and improving utilization. Without increasing the ZCU hardware cost, the chassis braking algorithm is deployed and executed locally, reducing performance losses caused by cross-domain communication and effectively improving the overall system operating efficiency and resource utilization efficiency.

Provided are a channel estimation method and apparatus, a program product, an electronic device, and a program product, which are applied to the technical field of communications, the method comprising: acquiring channel-related parameters corresponding to a current channel, the channel-related parameters being used for representing the performance of the current channel; determining an MMSE filtering order according to the channel related parameters; performing channel estimation according to the reference signal corresponding to the current channel to obtain a corresponding first channel estimation value; and determining an MMSE filtering matrix according to the MMSE filtering order and the channel related parameters, and performing filtering and interpolation on the first channel estimation value by using the MMSE filtering matrix to obtain a second channel estimation value. According to the channel estimation method provided by the embodiment of the invention, the MMSE filtering order is selected by identifying the performance of the current channel, so that the problem that the currently determined MMSE filtering order is mismatched with the current channel due to quick change of the state of the current channel can be avoided, and the performance loss is reduced.

The invention relates to a directional communication method of an instant messaging system during clustering based on an MQTT protocol, which comprises the following steps of: when a plurality of back-end Broker service nodes communicate with each other, only sending communication data to a specified node, and storing a mapping relationship between each user and the back-end Broker service node connected with the user by using a uniform storage component, according to the method, a plurality of Broker service nodes in a cluster environment can conveniently obtain the serial numbers of the Broker nodes to be connected with all users, and meanwhile, the messages are directionally forwarded to the specified Broker service nodes, so that the network flow can be greatly reduced, the performance loss of non-target Broker service nodes is reduced, the possible avalanche effect is avoided, the processing efficiency of a system and the network throughput are improved, and the user experience is improved. The system stability is enhanced.

The invention discloses a method for identifying kitchen waste occluded objects and judging occlusion relations, constructs an occlusion data set of kitchen waste images, and marks the visible mask, complete mask, and category of each target object therein; first, uses the occlusion data set to train an instance segmentation model and a mask restoration model; the input of the instance segmentation model is the kitchen waste image, and the output is the category and visible mask of each target object; the input of the mask restoration model is the visible mask and the kitchen waste image output by the instance segmentation model, and the output is the complete mask; secondly, the kitchen waste image to be identified and judged is used to obtain the category, visible mask, and complete mask of each target object using the trained instance segmentation model and mask restoration model; finally, the visible mask and the complete mask are combined to judge whether each target object is occluded and the occlusion relationship. The invention performs instance segmentation and mask restoration on recyclable garbage, and can obtain object category, position, occlusion relationship and complete shape information.

The invention discloses a forming method of a magnesium oxide-based carbon dioxide adsorbent suitable for large-scale preparation. The forming method comprises the following steps: calcining a magnesium-containing precursor to prepare porous magnesium oxide raw powder; the preparation method comprises the following steps: carrying out impregnation modification on magnesium oxide raw powder and an alkali metal salt solution to obtain wet mixed thick slurry, drying the slurry, and carrying out secondary calcination to obtain modified adsorbent raw powder; and mixing the raw powder with deionized water, a forming aid and the like to prepare a plastic wet material, and performing extrusion-spheronization balling, drying and calcining to obtain regular spherical adsorbent particles. The formed adsorbent prepared through the method has the advantages of being controllable in particle size, high in mechanical strength, good in adsorptive property after being formed, excellent in cycling stability and the like, can be used for CO2 chemical adsorption trapping of CO2-containing gas under the medium-temperature condition, and is good in CO2 adsorptive property, economical efficiency of the technological method and environmental protection performance, the process technology is easy to amplify, and the method is suitable for industrial production. The method is suitable for industrial large-scale production.

This invention discloses a file decryption method, system, electronic device, and readable storage medium, particularly relating to the field of network security technology. The method includes: obtaining the address of a target encryption function in a process and instrumenting the target encryption function; monitoring file open operations and initializing the state record of the corresponding file in thread-local storage; intercepting the call to the target encryption function and reading the state record of the corresponding file in thread-local storage; if the state record meets the initial recording condition, extracting the encryption parameters and ciphertext fragments involved in the call to the target encryption function, storing the encryption event containing the encryption parameters and ciphertext fragments, and updating the state record; obtaining the file to be decrypted, extracting the data blocks of the file to be decrypted, and matching the data blocks with the ciphertext fragments in the stored encryption event; when a match is successful, performing a decryption operation on the file to be decrypted based on the encryption parameters in the matched encryption event.

The invention discloses a data updating method and device, equipment, a storage medium and a program product, and is applied to the technical field of data processing. The method comprises the steps of obtaining updating result information of a database; detecting whether service process data related to the update result information exists in a cache or not; under the condition that the service process data exists in the cache, whether the service process data in the cache is updated or not is determined according to the updating result information; and under the condition of determining that the service process data is not updated, updating the service process data in the cache according to the updating result information. The method can effectively solve the problem that when the database is updated and the data query request is processed at the same time, the data query request obtains old data, and service data in the cache and the database are inconsistent.

The present application discloses a compilation-free loading method, device and medium for front-end engineering configuration files, and relates to the field of software development technology. The method comprises: constructing a corresponding configuration file to be loaded based on a preset configuration file deployment requirement; when a vue project corresponding to the configuration file to be loaded is started, precompiling the configuration file to be loaded through the Node environment; judging whether there is a window variable in the precompiled configuration file corresponding to the configuration file to be loaded, and determining the loading method to be applied according to the judgment result, so as to load the precompiled configuration file. The present application realizes dynamic loading and real-time parsing of the front-end engineering configuration file through the above method without recompiling the entire application.

The present application discloses an encryption and decryption engine module, an encryption and decryption method, a storage medium and a chip, and belongs to the field of chip technology. The encryption and decryption engine module includes n PPCs, n vCEs, multiple encryption and decryption modules, a memory protection controller MPC and n memory blocks, n≥2; each PPC is connected to a vCE, and each vCE corresponds to a memory block; the PPC is used to send an external encryption and decryption request to the corresponding vCE; the vCE is used to obtain a key according to the encryption and decryption request and select a target encryption and decryption module from multiple encryption and decryption modules, create an encryption and decryption context according to the target encryption and decryption module and the key, schedule the target encryption and decryption module to process the encryption and decryption request based on the encryption and decryption context, and send the data in the processing process to the MPC; the MPC is used to store the data in the memory block corresponding to the vCE. The present application improves the security of the system and reduces the performance loss of the system through hardware isolation.

The present invention discloses a method for verifying the performance of a storage engine of a Cockroach database system, and relates to the technical field of distributed databases. The method uses a TPCC performance test to record information of a corresponding storage engine interface called process through a recording function, and serializes the information of the called process and writes it into a file according to a specified path. The file is called according to a storage engine interface corresponding to the file to obtain data reflecting the performance of a database storage engine to verify the storage performance of the storage engine.

The application provides a terminal and edge server cooperative model reasoning optimization method and system and a storage medium. The method comprises the following steps: acquiring the computing power parameters and communication parameters of local terminals and edge servers in an initial model deployment environment; combining the computing power parameters, the communication parameters and layer structure information of the initial model to calculate a total splitting cost; taking the total splitting cost and model reasoning accuracy as optimization objectives, processing the initial model based on an optimal pruning strategy to generate a lightweight model; selecting a target splitting point matched with a real-time communication rate from the matching relationship between the communication rate interval and the splitting point, splitting the lightweight model into a terminal-side sub-model and an edge-side sub-model based on the target splitting point, deploying the terminal-side sub-model on the local terminal and deploying the edge-side sub-model on the edge server to carry out cooperative reasoning. The embodiment of the application can ensure reasoning accuracy while realizing low delay.

The invention belongs to the technical field of data storage, and provides a cache data processing method and system, a storage medium and electronic device.The method comprises the steps that a search result cache is obtained according to a user request; verifying the validity of the search results in the search result cache to obtain invalid search results; and processing the search result cache based on the invalid search result. According to the method, the validity period of the cache is prolonged to the maximum extent, performance loss caused by frequent refreshing is reduced, it is ensured that data obtained by a user is relatively accurate, and the performance and stability of the whole system are improved.

The embodiment of the invention provides an Android cross-thread message tracking method, which belongs to the technical field of data processing, and specifically comprises the following steps: intercepting a native method related to message processing in an Android system; when a client thread delivers a Message object to the message queue, a recording process is triggered by an enqueueMessage method of Hook, and call stack information of the client thread is acquired and stored and is subjected to associated storage; when the server thread takes out the Message object from the message queue and processes the Message object, an application process is triggered through a DispatchMessage method of Hook, and the current server thread is associated with the Message object; in response to a request for obtaining the stack information, judging whether the current thread is associated with a message object which is being processed or not; through the scheme of the invention, efficient, stable and low-intrusion message tracking is realized.

The invention provides an image coding method and device, equipment and a storage medium, and relates to the technical field of computers. The method comprises the following steps that: a CPU (Central Processing Unit) analyzes an image to be coded in a wavelet transform coding process to obtain original pixel data, and transmits the original pixel data to a GPU (Graphics Processing Unit); the GPU performs preprocessing operation on the original pixel data in parallel by using a plurality of threads to obtain standard pixel data meeting the wavelet transform requirement; wherein one thread processes one pixel column in the original pixel data; the GPU performs forward wavelet transform on the standard pixel data to obtain a sub-band coefficient; and the GPU performs entropy coding processing on the sub-band coefficients to obtain coded data. According to the scheme, image analysis is completed in the CPU, and preprocessing, wavelet transform and entropy coding processing are executed in parallel in the GPU, so that the coding throughput rate and the resource utilization rate in the image coding process can be improved.

The application discloses a method for instruction branch prefetching in an AIGPU processor, which comprises the following steps: 1) defining a branch prefetching instruction in a GPU instruction set, the branch prefetching instruction is used to check whether the instruction at label exists in the cache, if the instruction at label does not exist in the cache, one or more instruction data at label are fetched from the lower storage to the cache; 2) according to the PC absolute value or relative value coded by the branch prefetching instruction, the data at the program branch instruction address is prefetched. The application provides a mechanism for branch prefetching through an instruction, which can reduce the performance loss during branch jumping and improve the system performance. The application does not use a complex branch prediction circuit, and the prediction delay and prediction error are avoided.

A beamforming design method for intelligent metasurface arrays based on orthogonal time-frequency-space technology includes: aggregating multipath channels under direct illumination into cascaded channels with near-field direction vectors; randomly initializing the phase shift of intelligent metasurface units; establishing an effective channel matrix based on the cascaded channels; then establishing a multi-user and rate objective function; maximizing the objective function through phase shift optimization using a hierarchical search strategy; repeatedly executing the hierarchical search strategy to search and optimize subsequent blocks until the number of intelligent metasurface units in each block reaches a preset value; then performing an exhaustive search in the final block, thus completing one iteration; repeating the iteration until the maximum number of iterations is reached; outputting the final optimal phase shift; and implementing beamforming design based on the final optimal phase shift. This invention can reduce the complexity of beamforming design and ensure reliable transmission of high capacity during high-speed movement.

This invention relates to the field of high-performance computing and provides a method and apparatus for optimizing the performance of a graphics processing unit (GPU) aggregate communication operator. The method includes: allocating device-side storage space on the host side; copying the structured data to be communicated to the device-side storage space to obtain a target data pointer; adjusting the kernel function parameter of the aggregate communication operator from a structure type to a pointer type and passing the target data pointer; executing the kernel function on the GPU side, directly accessing the device-side storage space based on the target data pointer to complete the aggregate communication operation; and releasing the device-side storage space after the aggregate communication operation is completed. This invention solves the problem of excessive GPU register pressure and decreased cache hit rate caused by inefficient memory access instructions generated by the compiler to pass structure parameters in existing technologies. It avoids the generation of inefficient instructions, reduces register pressure, and improves execution efficiency.

The invention discloses a retry threshold table generation method and device and a storage medium, and relates to the technical field of memorizers, and the method comprises the steps: obtaining a target threshold voltage range, and generating a plurality of threshold points on a preset coordinate system according to the target threshold voltage range; determining a threshold point corresponding to the target average threshold under the data retention time as a first threshold point, and determining at least one second threshold point in the plurality of threshold points, the second threshold point being a threshold point which is not determined as the first threshold point; obtaining the distance from each threshold point to the origin of the coordinates, and according to the sequence of the distances from large to small, sorting the first threshold points, and then sorting the second threshold points; and generating a retry threshold table according to a sorting result. According to the method and the device, the retry threshold table which meets the requirement of a read retry mechanism and has fewer retry times can be generated.

The invention provides a preparation method of a spherical padding for sintering a resistor disc of a lightning arrester, and belongs to the technical field of manufacturing of lightning arresters. The method comprises the following steps: firstly, preparing slurry by taking granulation waste as a raw material, and then preparing the spherical padding by virtue of a breakthrough titration forming process for the first time, so that the forming precision is high, the spherical quality is excellent, the use performance of the padding can be directly improved, the material structure is stable, the performance loss is low, and the original characteristics of the waste can be reserved to the greatest extent; the damage to the material performance in the preparation process is reduced. In addition, the preparation method provided by the invention is simple, the process is mature, and the use amount of additives is small, so that the production cost can be effectively reduced, and the economic value is relatively high.

The present invention discloses a hybrid dynamic pruning method for large medical model compression, which adopts a sparse predictor based on hybrid sparse prediction, and has the advantage of reducing performance loss compared with other dynamic pruning methods based solely on dynamic sparse prediction. It effectively solves the problem of serious performance degradation after compression of large medical models, improves the model reasoning speed, and provides a feasible solution for the promotion and application of high-sparse, high-performance large models in practical applications.