180results about How to "Improve throughput" patented technology

An integrated cleaning and liquid supply device and online cleaning method for polyurethane polishing pads are disclosed. The device includes a base, a rotating seat, a swing arm, an independent liquid supply pipeline system and cleaning pipeline, a high-pressure nozzle array, and a positioning sensor assembly. During the polishing stage, polishing liquid is dripped through the liquid supply outlet. During the cleaning stage, the swing arm moves centripetally to a preset position aligned with the center of the polishing pad. The positioning sensor triggers a high-pressure pump, causing pure water to be sprayed as an atomized jet along the diameter of the pad surface through the high-pressure nozzle array. Simultaneously, the polishing pad rotates at a low speed, utilizing the synergistic effect of centrifugal force and high-pressure water flow to thoroughly remove residual contaminants within the micropores of the polyurethane. This invention integrates liquid supply and deep cleaning functions into the same motion mechanism, eliminating the need for machine shutdown or manual intervention. It offers advantages such as non-contact, non-destructive, thorough cleaning, and high automation, extending the life of the polishing pad and improving the stability of the CMP process.

This invention belongs to the field of intelligent agent large language model service technology, and discloses a chip computing system, device and application for intelligent agent large language model. Through the collaborative design of hardware architecture and software scheduling, it systematically solves the core technical challenges faced by intelligent agent large language model services. Specifically, it includes a hardware and software collaborative implementation scheme for resource partitioning, key-value cache management and runtime scheduling of pre-filling, decoding and tool call related workloads in a multi-chip encapsulation and inconsistent memory access environment.

This invention discloses a communication method and system for anti-interference wireless ad hoc networks, relating to the field of wireless networking technology. The method includes: a master node collecting historical interference feature sequences transmitted by each slave node (not part of the ad hoc network) and predicting predicted interference frequency band sequences that may affect each node within a preset future time range; analyzing and quantifying the interference confidence of each predicted interference frequency band sequence by combining historical transmission interaction data between nodes; using the interference confidence as weight, jointly and dynamically allocating transmission time slot frequency bands of the ad hoc network within the preset future time range using an iterative optimization algorithm to generate multiple optimized time slot frequency band sequences capable of actively avoiding high-confidence predicted interference; and precisely controlling the communication behavior of each node within the future time period based on the multiple optimized time slot frequency band sequences. This invention transforms traditional passive interference response into active prediction and avoidance, improving the communication reliability and overall performance of wireless ad hoc networks in complex interference environments.

The utility model discloses a multi-bin paper-based micro-fluidic chip for visual detection of bacterial drug resistance. The multi-bin paper-based micro-fluidic chip comprises a hydrophilic paper base, the detection mechanism is a hydrophilic channel established on the hydrophilic paper base; the detection mechanism comprises a sample adding bin and an anti-overflow bin, and the appearance of the anti-overflow bin is circular; one end of the anti-overflow channel is communicated with the sample adding bin, and the other end of the anti-overflow channel is communicated with the anti-overflow bin; the appearance of each reaction bin is an isosceles trapezoid; the plurality of reaction bins are uniformly distributed in a fan shape by taking the circle center of the sample adding bin as a central original point; one end of each diffusion channel is communicated with the reaction bin, and the other end of each diffusion channel is communicated with the sample adding bin; wherein the diffusion channels and the reaction bins are arranged in a one-to-one correspondence manner; the hydrophobic barrier structure is a hydrophobic area formed on the hydrophilic paper base through wax spraying printing along the outer edge of the detection mechanism. According to the multi-bin paper-based micro-fluidic chip for visual detection of bacterial drug resistance, provided by the utility model, rapid, high-throughput and low-cost detection of multiple bacterial drug resistance can be realized.

The invention discloses a high-temperature-resistant solvent-resistant polyurea-based organic solvent reverse osmosis membrane as well as a preparation method and application thereof. The organic solvent reverse osmosis membrane comprises a cross-linked polyimide porous support layer and a polyurea separation layer, wherein the polyurea separation layer is formed by an interfacial polymerization reaction of a water-phase mixed monomer and an organic-phase monomer. According to the invention, Cyclen is introduced as a comonomer to regulate and control the microstructure of the polyurea separation layer, so that the separation performance and solvent permeation flux of the membrane are remarkably improved. The retention rate of the prepared membrane on NaCl in an aqueous solution reaches 97.9%-99.3%, and the water flux reaches 13.7 LMH / MPa; and the retention rate of ibuprofen (206 Da) in an ethanol solution reaches 92.6%, and the ethanol flux reaches 4.52 LMH / MPa. The membrane has excellent solvent resistance and high temperature resistance, the retention rate of NaCl is still kept above 95.8% after the membrane is soaked in DMF (Dimethyl Formamide) at 25 DEG C for 30 days, and the separation performance is not obviously reduced after the membrane is soaked in ethanol at 70 DEG C for 12 days. The preparation method is simple and has a wide application prospect in separation of low-molecular-weight organic mixtures in the fields of chemical engineering, pharmacy and the like.

The application relates to a system for realizing rapid mass memory data persistence, data double writing and data pushing in a securities core transaction system, wherein the system comprises a file storage engine, a data analysis engine, a data routing engine, a data persistence engine, a data double writing engine and a data pushing engine, and a plurality of nodes are arranged in the system, and the file storage engine, the data analysis engine, the data routing engine, the data persistence engine, the data double writing engine and the data pushing engine are arranged in the corresponding nodes. The application also relates to a corresponding method. The system and the method for realizing rapid mass memory data persistence, data double writing and data pushing in the securities core transaction system have the advantages that the system is deployed in a cluster mode, the node state is monitored in real time through a master-slave mode, the system reliability is improved, high throughput and low delay are achieved under the premise of high availability, and the data integrity is ensured.

This application relates to the field of interface gateway management technology, and discloses an open application programming interface (API) gateway management system and method. The method includes: responding to a received API call request, extracting the caller's identity information and request context parameters, performing dynamic authentication on the call request, and obtaining the authentication result; obtaining the corresponding rate limiting rules, performing dynamic mixed-mode traffic control on the call request, and obtaining the traffic control result; forwarding the call request to the corresponding backend service, and monitoring the backend service's operating status data in real time during the forwarding process, performing circuit breaker analysis based on the operating status data and preset business semantic rules, and obtaining the circuit breaker control signal; executing the corresponding request processing action according to the circuit breaker control signal, and collecting full-link observable data such as call chain tracing data and network performance data in real time to manage interface calls; this application can effectively intercept replay attacks, abnormal logins from different locations, etc., and improve gateway throughput.

The present application relates to the technical field of GPU computing optimization, and in particular to a sparse attention computing method, device and medium for GPU, wherein the method realizes the high efficiency of long context reasoning through the geometric perception sparse attention framework of ball hashing, and combines a large-scale parallel hashing optimization algorithm and a load adaptive computing kernel. Compared with the existing sparse attention methods based on heuristics or gradient learning, the present application realizes higher retrieval recall rate, lower preprocessing overhead and efficient hardware adaptation to irregular sparse patterns.

The application discloses a method for non-contact testing of quantum efficiency (EQE) and short-circuit current (ISC) of a solar cell, relates to the technical field of photovoltaic device testing, and aims to solve the problems that the existing quantum efficiency and short-circuit current measurement must contact electrodes, is slow and cannot cover new gateless or process pieces; a non-contact testing method of "multi-wavelength LED excitation + local shielding + carrier lateral diffusion" is provided; the method blocks the excitation light by using a shielding filter, and only transmits the electroluminescence signal of the shielding area; the EL power of each wavelength is extracted through phase-locked amplification and deep learning denoising, the continuous EQE curve is calculated by using the calibrated incident photon flux ratio, and the ISC is obtained by convolution of the AM1.5 spectrum.

The application relates to a high-strength high-separation-performance carbon molecular sieve hollow fiber membrane and a preparation method and application thereof, and the preparation method comprises the following steps: (1) mixing cellulose acetate, cellulose, an organic solvent and an ionic liquid to prepare a casting solution, and stirring and heating the casting solution; (2) vacuumizing and defoaming the casting solution, and preparing a hollow fiber cellulose acetate membrane by using a dry-wet spinning method; and (3) drying the hollow fiber cellulose acetate membrane, and preparing the high-strength high-separation-performance carbon molecular sieve hollow fiber membrane by carbonizing under temperature rise in a protective atmosphere. Compared with the prior art, the carbon molecular sieve hollow fiber membrane has good gas separation performance and high strength. The carbon molecular sieve membrane is used for He / CH4 separation, has high He permeation performance and a high He / CH4 separation factor, has good gas separation performance under high pressure, and provides a new choice for natural gas helium extraction membrane materials.

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.

This invention discloses a heterogeneous acceleration system and method for analog modulation recognition of large-scale multi-channel signals, comprising: a CPU host processing end and a GPU device computing end; the CPU host processing end is used for asynchronous pipelined processing of data reading, task scheduling, and result distribution, realizing parallel execution of data reading, GPU computing, and result distribution on the time axis; the GPU device computing end completes the entire process of signal preprocessing and deep learning model inference within the GPU memory after the original signal is transmitted to the video memory via the PCIe bus, and does not exchange data with the CPU host processing end until the recognition result is generated. This invention eliminates data transport and I / O blocking through a three-stage asynchronous pipeline architecture and a closed-loop processing link in the entire video memory, and combines parallel preprocessing operators, inference acceleration engines, multi-stream scheduling mechanisms, and dynamic video memory pools to achieve high throughput and low latency recognition of large-scale multi-channel analog modulation signals.

This invention discloses a hardware acceleration method and accelerator for Transformer based on mixed-precision quantization and Huffman coding. The acceleration method includes: using a genetic algorithm to obtain several configuration schemes for mixed-precision quantization of Transformer network layers; performing mixed-precision quantization on each Transformer network layer based on each quantization configuration scheme to obtain a corresponding KL divergence; training a multilayer perceptron to obtain a quantization configuration prediction network using the quantization configuration scheme and the corresponding KL divergence as the output label and input feature, respectively; receiving a user-set target KL divergence value, using the quantization configuration prediction network to obtain the corresponding quantization configuration scheme, and performing mixed-precision quantization on each network layer based on the quantization scheme; and using Huffman coding to encode and compress all quantization weights before on-chip storage. This invention can reduce storage and computational overhead while maintaining model accuracy.

The application discloses a SSR molecular marker primer related to Chinese chives dormancy and a detection method and application thereof. In view of the problems that traditional Chinese chives dormancy trait identification relies on field morphological observation for 4 months, is inefficient and lacks special molecular markers, 7 specific SSR primers (SSR005 / 170 / 249 / 325 / 423 / 478 / 610) are developed based on cross-species genome resequencing (taking Chinese onion as a reference genome, and the alignment rate is 18.76%-18.89%) for the first time. Through the establishment of a matching detection system: CTAB method for rapid DNA extraction, optimization of the PCR system (20 muL contains 30 ng of template DNA, 0.4 U of Taq enzyme, 0.2 muM of primer, 0.2 mM of Mg 2+ 2.0 mM of dNTPs, 57 DEG C annealing), 8% polyacrylamide gel electrophoresis combined with silver staining development, the dormancy trait can be accurately identified within 7 days. For example, the SSR423 primer has no band at 177 bp, and the non-dormant / dormant germplasm cannot be distinguished. The detection results of 52 germplasms (such as dormancy germplasm 21-1 and non-dormancy germplasm 22-7) are consistent with the field phenotypes at a rate of 100%.

This invention provides a primer-probe combination, reagents, methods, and applications for detecting Brucella vaccine strain RB51, belonging to the field of pathogen detection technology. The primer-probe combination for detecting Brucella vaccine strain RB51 of this invention includes a forward primer RB51-F, a reverse primer RB51-R, and a probe RB51-P; the forward primer RB51-F is shown in SEQ ID NO:1, the reverse primer RB51-R is shown in SEQ ID NO:2, and the probe RB51-P is shown in SEQ ID NO:3. The fluorescent PCR and digital PCR detection methods established based on the primer-probe combination of this invention have high sensitivity, good specificity, fast detection speed, and high throughput, enabling rapid detection of Brucella vaccine strain RB51, making differential diagnosis more accurate, economical, efficient, and convenient.

This application discloses a blockchain-based data processing method, device, and readable storage medium. The method includes: acquiring a set of state message pairs X. y ; Status message pair set X y Including status messages for A b Status message for A b Including consensus node Z b The first status message C of the broadcast b and consensus node Z y The generated second state message C y Determine consensus node Z b In status message pair A b The state score in the set X; based on the state message, the set X... y The consensus nodes are sorted according to the state score corresponding to each state message pair to obtain the set X of state message pairs. y The corresponding node sorting results; based on the node sorting results corresponding to the set of N state messages, the node states corresponding to the N consensus nodes are determined. Using this application, node states can be accurately identified and the throughput of the blockchain network can be improved. This application can be applied to various scenarios such as cloud technology, artificial intelligence, smart transportation, and assisted driving.

The application discloses a kind of unmanned vehicle logistics directional relay network system and its coordination method, belong to intelligent logistics and unmanned vehicle application technical field, the system includes: multiple cellular airspace units, cellular airspace unit is constituted continuous grid in service area;Deployed in the center of each cellular airspace unit one micro hub, as the ground logistics node of corresponding unit;Wherein, each micro hub includes one terminal nest and at least two relay nests;Terminal nest is used to provide the storage and withdrawal interface of user goods;Each relay nest is arranged around terminal nest, and each relay nest is pre-allocated and fixedly responsible for the bidirectional goods transfer task between the unit where the micro hub is located and a particular adjacent unit;Terminal nest and each relay nest are connected by internal logistics transmission device 103.The application fundamentally simplifies scheduling and management complexity by deep coordination of cellular airspace grid, heterogeneous node division of labor, directional routing rules and high automation.

The utility model relates to sewage treatment technical field, a kind of coal chemical industry wastewater ceramic flat sheet membrane water reuse system, including ceramic membrane pool, and ceramic flat sheet membrane is inside in ceramic membrane pool;Water production pool is connected to the filtration side of ceramic flat sheet membrane by water production pipe;Water production pump is set on water production pipe;Backwash pump input is connected with the water production pool by pipeline, and the output end of backwash pump is connected the sewage side of ceramic flat sheet membrane;Ceramic membrane pool bottom is equipped with sludge discharge pipeline, and sludge discharge pump is arranged on the sludge discharge pipeline.This technology handles coal chemical industry wastewater effect is good, and effluent quality water quantity is stable.

This invention discloses a layered decoding method and apparatus for LDPC frames based on mutual information, relating to the field of digital signal transmission in the physical layer of wireless communication. The method includes the following steps: 1) Calculating the LLR value of a given codeword to be decoded, and performing Q-ratio specific-point quantization to maximize mutual information on the entire frame's LLR value; 2) Calculating the PP value of the Q-ratio specific-point quantized LLR value according to Bayes' formula; 3) Layering the PP values ​​of the LLR values ​​of the data to be decoded according to the coefficients of the parity check matrix H; 4) Calculating the information within the information node; 5) Performing parity check node update calculations according to the PMS decoding algorithm to obtain the information value outside the parity check node; 6) Updating the information node using the information value outside the parity check node; 7) Symbol decoding; 8) Calculating the syndrome according to H. If the result of the syndrome is 0 or the maximum number of iterations is reached, outputting the three-frame codeword result; otherwise, returning to step 3) for the next decoding iteration. This invention can effectively improve the decoding processing speed of each frame.

This invention proposes a multi-committee parallel Byzantine fault-tolerant consensus method, GP-BFT. This method groups system replicas into multiple independent committees based on the minimum Byzantine fault tolerance parameter. Each committee executes consensus in parallel, effectively reducing intra-group communication. A two-level leader election is employed: each committee elects a first-level leader responsible for intra-group consensus, and a second-level leader is elected from among the first-level leaders using a threshold coin to coordinate globally. A dual timeout handling mechanism is designed: when less than half of the committees time out, cross-path collaborative commit is used, leveraging committees that have completed consensus to drive the timed-out committees to quickly resubmit; when more than half of the committees time out, aggregated timeout certificates are used to advance the global round. This method significantly reduces consensus latency and communication overhead in large-scale replica scenarios while ensuring security and liveness.

The application provides a resource allocation method and device, electronic equipment and a readable storage medium, comprising: obtaining a first target function and a constraint condition corresponding to the first target function, the first target function being associated with total throughput and transmission delay of a communication channel; determining a target channel ratio of the communication channel based on the first target function and the constraint condition; determining a first ratio of an orthogonal channel and a second ratio of a shared channel in the communication channel based on the target channel ratio; determining resource allocation parameters of at least one V2I link based on the first ratio and pre-acquired communication channel parameters; and determining target transmission power of each link in the at least one V2I link and the at least one V2V link based on the first ratio, the second ratio, the resource allocation parameters and the first target function, so that the value of the first target function is minimized under the target channel ratio, the resource allocation parameters and the target transmission power. The application can reduce data transmission delay.

This invention discloses a data protection method and system based on field splitting and distributed de-identification. The method includes: an authorized external application splitting sensitive information into several independent fields and sending each sensitive field to a corresponding de-identification service node. Each de-identification service node is independently configured and does not interact with any other data. Based on the complete randomness of the received plaintext data, each de-identification service node generates a locally unique and irreversible mask identifier (MID) using an incrementing algorithm and stores the mapping relationship between the MID and the plaintext data. The MID is then returned to the external application. The application database of the external system does not store any plaintext data of the sensitive fields, and all administrators have no right to access the complete dataset of each de-identification service node. This invention ensures that even if the application database or all de-identification service nodes are compromised, the complete sensitive information cannot be reconstructed, maximizing data security without increasing computational costs.

The application discloses a kind of alliance chain slice construction and consensus method based on dynamic clustering, comprising, the real-time computing power of alliance chain node, transaction frequency and transmission rate are collected;By verifiable random function, node is divided into Full_node of storing complete data and Slice_node of only storing slice data;DQN algorithm is used to determine the number of slices K, and the final consensus group is formed by selecting consensus representative nodes from Full_node based on dynamic performance threshold;The correlation matrix between the remaining nodes is constructed, and the nodes are evenly distributed to K slices by means of improved clustering algorithm;In each slice and consensus group, the Leader node is elected;When a new transaction arrives, execute the corresponding consensus process according to whether the transaction parties belong to the same slice, and write into the blockchain after verification, collect consensus performance data by Leader, and determine whether to trigger slice reconstruction according to performance data: if triggered, re-slice, otherwise continue to process transactions until completion;The application can improve the throughput of the overall system, relieve the data storage pressure, and improve the system response efficiency.

The application discloses a multi-agent material calculation consistency calibration method, belonging to the cross field of material science, artificial intelligence and computational physics. The method adopts a master-slave structure multi-agent system, containing one master agent and at least three slave agents. The master agent is responsible for overall coordination, embedding physical mechanism constraints, consistency calibration and output results. The method steps are: constructing a calculation system, introducing physical mechanism constraints and embedding slave agent calculation models, collecting calculation data, performing bias analysis, adjusting parameters or correcting results through iterative calibration, and finally verifying and outputting the final results. The application introduces physical mechanism constraints, combines bias analysis and iterative calibration, and solves the problems of poor consistency of existing multi-agent material calculation results and easy violation of physical laws. Experiments show that the method reduces the deviation of the calculation results and the experimental measured values to below 3%, improves the calibration efficiency by more than 40%, and improves the calculation accuracy by more than 35% compared with the prior art, thereby improving the accuracy and engineering practicability of high-throughput calculation of material performance.

This invention discloses a laboratory-grade ceramic membrane pressure filtration device, comprising a pressure ceramic membrane filtration unit, a pressure pump, a feed container, a valve, and a dialysate container. This invention uses an inorganic ceramic membrane as the filter medium, which is resistant to acids and alkalis, resistant to organic solvents, has a wide range of applications, high strength, and high operating pressure. It maintains high membrane integrity during installation, disassembly, and filtration, and offers high operating flux and high recovery rate. It can withstand both high-pressure steam sterilization and chemical sterilization. The sealing method is simpler than that of tubular ceramic membranes, and the dead volume is significantly smaller, requiring relatively less feed solution, making it suitable for laboratory operating environments.

The application discloses a method for preparing an antibacterial nanofiltration membrane by using amikacin as a water-phase monomer, and belongs to the technical field of composite nanofiltration membrane preparation. The method uses amikacin as a water-phase monomer and adopts an interfacial polymerization method to prepare an antibacterial nanofiltration membrane (polyamide nanofiltration membrane). The method has the advantages of simple operation, mild reaction conditions and relatively low raw material cost. In addition, the prepared polyamide composite nanofiltration membrane has higher selective permeation performance and good antibacterial performance. The nanofiltration membrane flux can be as high as 20 L / (m 2 ·h·bar), and the tetracycline rejection rate can be as high as 99.3%. In addition, the application can also provide a reference for the preparation of other antibacterial interfacial polymerization separation membranes.

The present application relates to the technical field of data communication, in particular to a kind of reverse scattering network uplink and downlink rate adaptive method;Utilize the change of RSSI and phase value and the frequency modulation of channel occurrence carry out mobility detection;Channel estimation is completed based on filter-based channel detection and lossrate estimation;Rate adaptation is carried out to uplink and downlink;In the present application, RSSI value, the change of phase value and the frequency modulation of channel occurrence are used as detection trigger;The selection command in the protocol is introduced to design filter-based detection scheme, which avoids collision in MAC, reduces detection time and improves channel detection accuracy;According to the channel condition obtained by detection, rate adaptation is carried out to uplink and downlink, which greatly improves system throughput.

This invention discloses the KASP molecular marker for the wheat metsulfuron-methyl resistance gene TaASADH-B1 and its application, relating to a SNP site associated with wheat metsulfuron-methyl resistance in the field of marker-assisted breeding technology. The SNP site is located on wheat chromosome 5B, corresponding to the gene number TraesCS5B02G403400. The SNP site corresponds to the 101st base from the 5' end of SEQ ID NO: 1. When this site is homozygous for AA, the corresponding genotype is TaASADH-B1. A When this locus is homozygous for CC, the corresponding genotype is TaASADH-B1. C Wheat tolerance to mesosulfuron-methyl is characterized by the following genotype: TaASADH-B1 C Wheat tolerance is greater than or candidate greater than genotype TaASADH-B1 A The KASP marker developed in this invention is associated with wheat tolerance to the herbicide mesosulfuron-methyl and can be used for screening and identification of phenotypic traits of wheat varieties to tolerate mesosulfuron-methyl.

The application discloses a heterogeneous parallel material removal calculation method and system. The method comprises the following steps: constructing a heterogeneous parallel calculation environment comprising at least a first processing unit and a second processing unit, wherein at least one processing unit supports data parallel execution for a large number of discrete elements; constructing a discrete representation of a workpiece, and generating state data for representing a material state based on the discrete representation of the workpiece; generating space constraint information based on tool geometric parameters, relative motion information of the tool and the workpiece, and the state data; decomposing material removal simulation calculation into a preprocessing task and a relationship evaluation task; mapping at least two calculation tasks to the first processing unit and the second processing unit for execution, so that the first processing unit and the second processing unit concurrently or flow water overlap execute in at least part of a time period; updating the material state based on the space relationship quantity and calculating the material removal quantity, and iterating until the trajectory is completed. Therefore, the calculation efficiency is significantly improved, and the real-time simulation requirement is met.