11results about How to "Improve latency" patented technology

The invention relates to an optical module communication system. The optical module communication system comprises a first optical signal transmission device, a second optical signal transmission device and a third optical signal transmission device. The first optical signal transmission device is respectively connected with the second optical signal transmission device and the third optical signal transmission device through transmission optical fibers; the first optical signal transmission device comprises a first optical transceiver and a first digital signal processor, and a half of a transmitting link and a half of a receiving link in the first optical transceiver are subjected to signal processing through the first digital signal processor. The other half of transmitting links and the other half of receiving links in the first optical transceiver do not adopt the first digital signal processor to perform signal processing; the second optical signal transmission device comprises a second optical transceiver and a second digital signal processor, the third optical signal transmission device comprises a third optical transceiver and a third digital signal processor, and all links of the second optical transceiver and the third optical transceiver adopt corresponding digital signal processors for signal processing.

This application provides a method and system for integrated protection of wireless network optimization and data security. The method includes: collecting wireless network operation index data and network security threat index data; evaluating network quality based on the wireless network operation index data, calculating network channel quality values, and generating network optimization strategies; calculating security threat values ​​based on the network security threat index data, and constructing security protection strategies based on the security threat values ​​and network channel quality values; linking and matching the network optimization strategies and security protection strategies to form linkage control commands; and synchronously executing network parameter adjustments and security management actions according to the linkage control commands. This application achieves bidirectional linkage between wireless network quality perception and data security threat detection, optimizing network speed and latency while simultaneously identifying attacks and dynamically adjusting security strategies, achieving synergistic optimization and balance between performance and security.

The application discloses a large language model inference acceleration method for real-time voice interaction and electronic equipment, and relates to the technical field of artificial intelligence. The method comprises the following steps: obtaining optimal recognition text output by an automatic speech recognition engine as a draft sequence; inputting input data containing the draft sequence into a target large language model, and calculating posterior probability distribution of all token positions in the draft sequence through single forward propagation parallel calculation; determining a model output candidate sequence according to the posterior probability distribution of each position, comparing the model output candidate sequence with the draft sequence, and completing inference output based on the comparison result. The application does not need to additionally train, deploy and maintain an independent draft model, effectively reducing system engineering complexity and memory occupation. Meanwhile, the application scheme is executed based on the original architecture of the target large language model throughout the whole process, without the need to modify the model network structure, and without the need to carry out retraining or fine-tuning, so that the original generality and recognition accuracy of the pre-trained large language model are completely retained.

The application discloses a display panel and a display device. The display panel comprises a display area; a first middle line, the width of the display panel in a first direction is D, the first middle line extends along a second direction, the distance between the first middle line and the first edge of the display panel in the first direction is d1, d1=D / 4, at least one gate drive circuit is arranged on both sides of the first middle line in the first direction; the first direction and the second direction intersect; and the gate drive circuit is located in the display area. According to the embodiment of the application, frameless display can be realized, and display uniformity can be improved.

The application discloses an asynchronous arbitration circuit, an asynchronous arbitration method and an asynchronous arbitrator, and relates to the technical field of communication, aiming to solve the problem that the asynchronous arbitrator in the prior art may appear in a metastable state, which may cause errors in circuit logic and thus negatively affect the stability of a system. The asynchronous arbitration circuit comprises a request input unit, an arbitration unit, a filter unit and a handshake signal generation unit. The request input unit is configured to transmit a request signal to the arbitration unit and the handshake signal generation unit respectively in the case that a received response feedback signal is in a valid state; the arbitration unit is configured to perform arbitration processing on the request signal and determine a target output signal; the filter unit is configured to perform filtering processing on the target output signal to eliminate a corresponding metastable voltage in the target output signal, and output a corresponding response signal; and the handshake signal generation unit is configured to output a request front transmission signal based on the request signal.

The application provides a resource data query method and device, electronic equipment and storage medium. The method comprises the following steps: acquiring a plurality of binlog log data from a database of the OpenStack, and performing format processing on each binlog log data to obtain a plurality of to-be-screened log data; determining a resource identifier to be monitored based on a current deployment task; determining target log data to be monitored from the plurality of to-be-screened log data based on the resource identifier, and determining the target log data as resource data of the current deployment task, so as to determine the OpenStack resource state by acquiring the binlog log data in real time, and optimize the delay and state loss caused by periodically calling the API to query the resource state.

An undercarriage oil gas buffer device vertical force calculation and constant coefficient identification method adopts a method of introducing an acceleration term and a fractional power term of acceleration, combines a pseudo-linear least square method to carry out parameter estimation, and utilizes two-parameter selection and coefficient fitting to identify the vertical force of an undercarriage oil gas buffer device through an arranged buffer displacement sensor. All physical quantity calculation models and expressions are established, and the model fitting precision is remarkably improved through optimization of a multivariate nonlinear mathematical model structure of the undercarriage buffer.

This invention relates to the field of epoxy resins, and discloses an epoxy resin composition, epoxy resin material, its preparation method, and its application. The composition comprises 40-60 parts by weight of glycidyl amine epoxy resin, 10-20 parts by weight of polyenol acetal, 15-45 parts by weight of phenolic epoxy resin, 10-40 parts by weight of phosphorus-containing flame retardant, 1-5 parts by weight of carbon nanotubes, 1-10 parts by weight of curing agent, and 1-10 parts by weight of curing accelerator; the curing agent is a latent curing agent. The epoxy resin composition and epoxy resin material provided by this invention have excellent flame retardant properties, as well as good mechanical and thermal properties.

Provided are a bandwidth sharing method, device and equipment for multiple access points, the method comprising: sending a transmission opportunity allocation frame to a second access point, the transmission opportunity allocation frame being used for indicating a transmission sub-channel allocated by a first access point to the second access point in a shared transmission opportunity TXOP, the second access point performs data transmission in a C-FDMA frequency division mode based on the transmission opportunity allocation frame; wherein the second access point is an access point capable of carrying out bandwidth sharing with the first access point. A transmission control frame is sent to a first access point to indicate the first access point to allocate a transmission sub-channel in a shared TXOP for a second access point, so that the second access point independently performs data transmission in a C-FDMA frequency division mode based on the transmission sub-channel, and concurrent transmission with other access points in the shared TXOP is realized. The utilization rate of frequency domain resources is improved, and the time delay and the throughput rate are improved.

A method for operating a first network device, a first network device and a method for operating a communication network are disclosed. A method for operating a first network device of a communication network, wherein the method comprises: determining or receiving (102), by means of an ingress interface, an ingress data stream comprising payload data to be transmitted towards a second network device; determining or receiving (104), by means of an association observer, at least one association value for a plurality of communication paths between the first network device and the second network device, wherein each of the plurality of communication paths comprises a different physical channel of a plurality of physical channels; determining (106), by means of a multi-connectivity controller, a plurality of egress data streams depending on the at least one association value and depending on the ingress data stream; and transmitting (108) via a respective egress queue of a plurality of egress queues, wherein each egress data stream is associated with a different path of a plurality of paths.

Static random access memory (SRAM) global bitline circuits for reducing glitches during read accesses, and related methods and systems are disclosed. A global bitline scheme in SRAM can reduce output load, reducing power consumption. In certain embodiments, SRAM includes an SRAM array. The SRAM includes a global bitline circuit for each SRAM array column. Each global bitline circuit includes memory access circuit that pre-charges local bitlines corresponding to bitcells in SRAM array. The data read from selected bitcell is read from its local bitline onto aggregated read bitline, an aggregation of local bitlines. The SRAM includes bitline evaluation circuit that sends data from aggregated read bitline onto global bitline. Instead of sending data based on rising transition of clock trigger, data is sent onto the global bitline based on falling transition of clock trigger. A global bitline scheme can be employed that reduces glitches and resulting increases in power consumption.