9results about How to "Avoid communication delays" patented technology

This invention relates to the field of artificial intelligence technology, providing a method, electronic device, and storage medium for calculating attention mechanisms. The method includes: configuring multiple thread bundles that share the same set of thread-local registers; breaking down the attention mechanism calculation process into multiple computational sub-steps and assigning these sub-steps to the multiple thread bundles, wherein each thread bundle is configured to execute its assigned sub-steps; and controlling the multiple thread bundles to execute the sub-steps in parallel in a pipelined manner by reading and writing to the shared thread-local registers to obtain the calculation result of the attention mechanism. This invention, by utilizing shared thread-local registers to construct a pipelined calculation mode for multi-threaded bundle collaboration, achieves efficient data copying and deep parallelism between different computational units during the attention mechanism calculation process, offering advantages such as low communication overhead, high hardware resource utilization, and strong operator execution performance.

The invention discloses a main and standby machine switching control method and system for a CDU unit. The method comprises the steps that S1, performance parameters of a main CDU are monitored in real time; s2, evaluating the health state of the main CDU based on the performance parameters, and generating an evaluation result; s3, comparing the evaluation result with a preset switching threshold value, and when the evaluation result reaches the switching threshold value, starting a standby CDU and controlling the standby CDU to form an internal circulation loop on a secondary side; and S4, the primary side valve opening degrees of the main CDU and the standby CDU are cooperatively adjusted, and the cooling load is gradually migrated to the standby CDU from the main CDU. According to the invention, through predictive switching, fault processing is changed from post-remedy to beforehand prevention, the risk of overheating of the data center caused by slow attenuation of the performance of the main CDU is avoided, and the operation reliability of the data center is improved. The control logic of first voltage building and then grid connection is adopted, the water hammer effect and pressure fluctuation in a traditional switching mode are eliminated, a data center server cold plate and a pipeline system are protected, and the service life of equipment is prolonged.

The application discloses a quadruped robot substation autonomous and teleoperation dual-mode inspection system and method, and belongs to the technical field of power system intelligent operation and maintenance and robot control. In the autonomous operation mode, the device semantics, voltage level and power safety insulation distance are fused to construct a rule semantic safety map, and from geometric obstacle avoidance to deep understanding of electrical regulations is realized. A double-layer local path planning strategy is adopted, the absolute safety insulation distance is taken as a constraint, and a safe reference path conforming to the power safety regulations is searched out. Then, the lower layer is solved under the constraint of the upper layer reference path boundary constraint and other constraints. In the remote teleoperation mode, global rule safety evaluation is carried out on the continuous virtual escape trajectory, and then the continuous virtual escape trajectory is issued and executed. The problems that the quadruped robot fails to inspect in the strong electromagnetic interference and the structure height similar scene, lacks semantic understanding of electrical safety regulations, and lacks autonomous and teleoperation smooth switching when it is trapped in the dead zone and cannot make autonomous decisions are solved.

The invention relates to an operation parameter determination method and device of energy storage power generation equipment, computer equipment, a computer readable storage medium and a computer program product. The method comprises the steps of obtaining target power load data of a target area corresponding to energy storage power generation equipment, determining a target calculation strategy of operation parameters of the energy storage power generation equipment corresponding to the target power load data in a preset strategy library, and calculating the operation parameters of the energy storage power generation equipment according to the target calculation strategy and the target power load data. And target operation parameters of the energy storage power generation equipment are determined. By adopting the method, the power supply stability of the energy storage power generation equipment can be improved.

The invention relates to the technical field of intelligent manufacturing and edge calculation, in particular to a processing equipment operation parameter self-adaptive correction method based on edge calculation. The method comprises the steps of state acquisition, model calculation, residual error comparison and parameter correction. The method comprises the following steps: synchronously acquiring a real-time state and a working condition instruction of equipment at an edge node, and inputting the real-time state and the working condition instruction into an ideal dynamic model to obtain ideal reference data; the core of the method is to inject a degradation operator to generate theoretical damaged simulation data, calculate a residual error of real and theoretical data relative to a reference, and calculate a topological similarity of the real and theoretical data; when the similarity is greater than a preset threshold value, extracting a target degradation operator, converting the target degradation operator into a correction parameter, and issuing and executing; otherwise, judging that the noise is kept unchanged; according to the method, cloud communication delay is avoided, feedback and instructions are ensured to be strictly aligned, and the real-time efficiency of adaptive correction is improved.

The invention discloses a Beidou short message-based power distribution terminal communication optimization method and system, and the method is applied to a Beidou short message-based power distribution automation system, a Beidou communication module obtains a response message of a power distribution automation terminal, sets the highest priority for a remote control message, reports the remote control message in real time, shields a frequent displacement message for a remote signaling message, and carries out the real-time monitoring of the remote signaling message. Key information is extracted from the remaining remote signaling messages, protocol conversion is carried out according to a private protocol, then the messages are reported, and the telemetering messages are reported through total calling and important remote signaling deflection triggering; the data is compressed and numbered before being reported, so that channel blockage and message loss are avoided; and the Beidou short message communication manager obtains a remote control instruction of the power distribution automation master station, directly responds to a corresponding message to the power distribution automation master station if the remote control instruction is a remote control selection instruction or a remote control revocation instruction, directly issues the corresponding message if the remote control instruction is a remote control execution instruction, and starts a retransmission mechanism if the remote control execution confirmation message is overtime. According to the invention, the integrity and reliability of data transmission are systematically improved.

Embodiments of the present specification provide a device control method, device and computing device, wherein the device control method is applied to a first device in a rail transit system, and the method comprises: determining a second device associated with the first device; obtaining first space-time information of the first device, second space-time information of the second device, and a connection relationship between the first device and the second device; constructing first group information of the first device according to the first device, the second device, the first space-time information, the second space-time information and the connection relationship, wherein the first group information is used to store a structural relationship between the first device and the second device in a token ring network, and group information of each device in the rail transit system is used to store a token ring network of the constructed rail transit system; and communicating with the second device through token scheduling based on the first group information. The token ring network of the rail transit system is constructed, communication between different devices is realized, and efficient, conflict-free communication and collaborative control are realized.

The invention relates to the technical field of wireless communication networks, and discloses an adaptive power control method based on channel sensing, which comprises the following steps: S1, a sending node sends a request sending frame; s2, the target receiving node and the cooperative neighbor node synchronously receive the request sending frame; s3, calculating a preliminary optimal transmitting power set and an optimal data channel; s4, judging the current path of the target receiving node; a corresponding sending strategy is adopted; s5, performing fine granularity power distribution; and S6, performing data transmission on the specified data channel. According to the method, a power distribution algorithm with maximization of a signal-to-noise ratio harmonic average value as a target is introduced, and adjacent node interference information broadcasted by a sending node is fused in calculation, so that network-level collaborative optimization is realized; the method not only improves the transmission quality of the target link, but also actively suppresses the potential interference to the adjacent nodes, effectively improves the overall spectrum efficiency of the system and the user fairness, and reduces the overall interference level of the network.

The invention discloses a virtual machine emergency processing method and device during host machine downtime, a medium and equipment, and relates to the technical field of computers.The method comprises the steps that a virtual machine list and virtual machine key information corresponding to a downtime host machine are obtained, and the virtual machine list comprises multiple virtual machines; based on the virtual machine key information, whether each virtual machine drifts or not is judged one by one; aiming at the first virtual machine which does not drift, restarting the first virtual machine after checking the fault of the downtime host machine; and in the drifting second virtual machine and / or the restarted first virtual machine, screening a target virtual machine of which the application is not started, and executing an application recovery operation on the target virtual machine according to an application recovery processing method matched with the target virtual machine. According to the method and the device, automatic emergency processing of the virtual machine can be realized when the host machine crashes, and high requirements of the financial industry on business continuity and stability are met.