251results about How to "Improve hit rate" patented technology

The invention relates to the technical field of computer storage, and discloses a computer memory data processing system, which comprises the following modules: an intelligent cache prefetching module, which is used for predicting a future memory access address based on deep learning and reinforcement learning technologies and loading data into a cache in advance; the heterogeneous storage management module is used for dynamically classifying the data based on the data popularity value and realizing migration of the data among different storage media; and the storage consistency guarantee module is used for managing the data consistency among different storage layers through the improved MES I protocol. Cache prefetching is optimized by introducing deep learning and reinforcement learning technologies, and by predicting a future memory access address in real time and loading data in advance, the system effectively reduces the cache missing rate, improves the cache hit rate, remarkably reduces the memory access delay, and improves the memory access efficiency. And the system can keep more efficient data access performance in a large-scale data processing scene.

The invention provides a front-end performance optimization method based on cache management, and the method comprises the steps: S1, collecting and processing user behavior data, and obtaining target data; s2, analyzing a user behavior based on the target data, constructing a user behavior model, and predicting an access demand of a user by using a prediction algorithm to obtain a prediction model; s3, generating a caching strategy based on the user behavior model and the prediction model; and S4, storing resources on the client, the server and the content distribution network node based on the caching strategy. The cache hit rate is improved; the caching strategy is dynamically adjusted according to the actual use condition of the user, and the user requirement is more accurately met; the cache preloading mechanism can shorten the response time of the user request, can adapt to different user behaviors, and improves the user experience. The system and the electronic equipment provided by the invention have corresponding advantages.

The invention discloses a cache memory and a corresponding cache processing method thereof. A plurality of cache groups are stored through a label unit in a cache memory, each cache group comprises a plurality of cache lines, each cache line comprises a label bit and an index bit, the label bit is used for storing a label address, the index bit is used for storing an index, a data unit is used for storing data, an identifier of the data is the index, and after an access request is received, the label address is stored in the data unit. And obtaining a matched existing label address according to a target label address in the access request, and accessing corresponding data in the data unit according to a target index corresponding to the matched existing label address. Therefore, the label address and the data are associated through the index, dynamic mapping between the data and the label address is achieved, the capacity proportion between the data and the label address can be flexibly adjusted, the capacity of the label unit can be increased, and meanwhile the capacity of the data unit can be kept unchanged; to improve hit rate without increasing area and power consumption of a cache memory.

The invention relates to the technical field of semiconductor material preparation, and provides a method for controlling the resistivity of a fifth main group element doped crystal bar, crystal pulling time is distinguished in different crystal pulling stages, the volatilization coefficient of doped elements in different time periods is adjusted, the volatilization amount of the doped elements in each time interval is calculated more accurately, and the resistivity of the doped crystal bar is controlled. Therefore, the amount of doping elements required for controlling the resistance can be calculated more accurately, the resistance of the crystal bar can be adjusted more effectively and more accurately, the resistance difference between the head and the tail of the single-crystal crystal bar can be better reduced, meanwhile, through timely adjustment of the control process of the grade and the ending volatilization amount, the supplementary doping condition in the crystal pulling process is defined, and the crystal pulling efficiency is improved. And the resistance hit rate of the crystal bar is improved, and the head and tail resistance difference is reduced, so that the resistivity of each part of the crystal bar is accurately controlled, and the consistency of subsequent silicon wafer product quality is ensured.

The embodiment of the invention provides a multi-core network-on-chip transaction processing system, method, device, medium and product, the multi-core network-on-chip transaction processing system comprises a plurality of clusters, each cluster comprises four processing cores, a slave node and a master node, each processing core is provided with a cache, each slave node is provided with a slave node, and each slave node is provided with a slave node; the slave node is provided with a private cache, and the slave node comprises a four-stage pipeline architecture; the slave node is used for processing a request transaction through the four-stage pipeline architecture and generating a target response message or target response message data under the condition that the request transaction is received, and the four-stage pipeline architecture is constructed based on different clock periods and is used for periodically processing the request transaction.

The invention discloses a cache management method, electronic equipment, a storage medium and a program product, and relates to the technical field of data storage, and the method comprises the following steps: obtaining current operation data of a storage system in a current period, determining current access popularity corresponding to a plurality of data blocks according to the current operation data, and storing the data blocks according to the current access popularity. The resource allocation of the target cache is dynamically adjusted, the problem that the data access efficiency of a storage system in the related technology is low is solved, and the technical effect of improving the data access efficiency of the storage system is achieved.

The invention provides a data prefetching method and device. The data prefetching method comprises the steps that parameters of a read request for a storage device in a prefetching time window are obtained, and the parameters of the read request comprise features of the read request and logic block address information of the read request; according to the characteristics of the read requests, determining the prefetch number of subsequent read requests for the storage device; according to the logic block address information of the read request, determining the prefetch heap block addresses of the prefetch quantity of the subsequent read request; and pre-fetching the data of the pre-fetching heap block addresses of the storage device according to the pre-fetching number.

The embodiment of the invention relates to the field of artificial intelligence, and discloses an edge semantic caching method and device based on a content distribution network, and a storage medium, and the method comprises the steps: obtaining a current prompt inputted by a user, and judging the correlation between the current prompt and the context of the current prompt; after it is judged that the current prompt is related to the context of the current prompt, the current prompt and the context of the current prompt serve as retrieval conditions, and a first target answer is retrieved in a semantic cache pre-deployed in an edge node to which the user belongs; if the first target answer is retrieved, returning the first target answer to the user; and if the first target answer is not retrieved, submitting a current prompt to the large language model to obtain the answer, returning the answer to the user, and storing a prompt of dialogue information determined according to the dialogue information of the user and the large language model, a context and the corresponding answer in a semantic cache pre-deployed in an edge node to which the user belongs. According to the scheme, the hit rate and the accuracy rate of semantic caching can be considered and improved at the same time.

The invention discloses a cloud storage acceleration method and system based on edge computing, and relates to the field of electrical digital data process.The method comprises the steps that a file access track ring is constructed; extracting historical access files as a first associated file set; calculating the access time difference of each historical access file; determining a plurality of association clusters and two association file sets according to the access time difference; performing cyclic displacement in the second associated file set to obtain a displacement associated cluster; extracting a third associated file set; determining a prefetching file; setting an annular cache queue, and determining a prefetching position according to the position information; when the space of the annular cache queue is smaller than a preset value, the prefetched file corresponding to the prefetched position in the preset interval of the annular cache queue is reserved, and a final prefetched file is obtained; and when it is detected that the user accesses the target file, loading the final prefetched file from the cloud storage server to the edge server. The method and device are used for solving the problem that the cache hit rate of the associated files is reduced, and the cloud storage acceleration speed is increased.

The invention relates to the technical field of cache optimization, and discloses a cache optimization method and device, equipment, a storage medium and a computer program product.The method comprises the steps that data information of all cache nodes in different cache hierarchies is obtained, a cache strategy is updated based on the data information, and an updated cache strategy is obtained, the different cache levels comprise a local cache level, a regional cache level and a global cache level; on the basis of the updated caching strategy, each caching hierarchy is optimized, and the optimized caching content is obtained; in response to a content request of a user, carrying out dynamic path selection at different cache hierarchies to obtain a target request path; and obtaining target cache content from the optimized cache content based on the target request path. The method can effectively adapt to data requests of different scales and types, efficient utilization of storage resources and optimization of access performance are achieved, the cache hit rate is effectively increased, and the content transmission cost is reduced.

The invention discloses a flow-based local message classification method and device, and relates to the field of message classification, and the method comprises the steps: carrying out the parallel comparison of a keyword of a to-be-classified message with a keyword temporarily stored in a shift register, obtaining a comparison result, and generating an index matching result through an arbiter according to the comparison result; according to the locality principle of a message flow, historically queried table item information is cached into a dynamic cache table, an arbiter arbitrates a comparison result of a shift register, calculates an offset address of the dynamic cache table and obtains a matching result, and whether an original rule table needs to be queried and whether a dynamic cache table memory needs to be updated are determined according to a search result. Under the condition that the matching is unsuccessful, the dynamic cache table memory can guide the same message to directly use the offset address to obtain the table item information in the next search without searching from an original rule table, so that the hit rate is increased, the matching efficiency is improved, and the storage resources are saved.

The invention provides a sniping type arresting shooting planning method based on an electromagnetic gun, and the method comprises the steps: 1, obtaining target motion state information through a target state information filtering algorithm; wherein the target motion state information comprises position information, speed information and acceleration information of the target; 2, the main vehicle customizes an array bullet screen according to the target motion state information and the vehicle body motion state information; the arrayed bullet screens are divided into bullet facing face bullet screens and flyswatter bullet screens; 3, the main vehicle predicts the position information and the required movement time when the target reaches the specified interception slant distance according to the target movement state information, so that the maneuvering target is regarded as a fixed point and sniping interception is carried out; and step 4, information intercommunication among multiple cannon vehicles is carried out, a launching time sequence is determined, asynchronous launching is carried out, a plurality of cannon vehicles are shot at the same time, an array bullet screen is formed, and sniping is carried out on a maneuvering target. The method has the capability of actively sniping the maneuvering target.

The invention provides a bad root cause analysis method and system based on a comprehensive analysis framework, and relates to the technical field of bad root cause analysis, and the method comprises the steps: obtaining bad data and processing process data; performing comprehensive analysis on the bad data and the processing process data according to the comprehensive analysis framework so as to position intermediate link factors related to the bad product; performing root cause positioning analysis on the processing process data according to the bad data so as to preliminarily position a root cause station and root cause machine equipment; root cause positioning analysis is carried out on the processing process data according to the intermediate link factors related to the bad products so as to further position root cause machine equipment; and performing fusion judgment on the root cause site and the root cause machine equipment which are preliminarily positioned and the root cause machine equipment which is further positioned to obtain a comprehensive root cause positioning result. The root cause positioning of the processing equipment is carried out based on the comprehensive analysis framework, and the problem that comprehensive analysis and judgment cannot be realized due to isolation between different bad root cause analysis means is solved.

The embodiment of the invention discloses a method and system for improving the caching performance of an online transaction platform, and belongs to the technical field of computers. According to the embodiment of the invention, the cache resource allocation strategy of the online transaction platform is optimized through a series of steps. The method comprises the following steps: firstly, acquiring comprehensive cache performance statistical data to provide a basis for accurate decision making; and determining a candidate strategy and an initial weight by using reinforcement learning decision knowledge learned from the association platform, adjusting the weight by means of an aging demand level, and finally determining an optimal strategy. The process can dynamically adapt to the data access mode and the cache state change of the platform, the cache hit rate is improved, the cache resource utilization rate is optimized, the system response delay caused by improper cache strategies is reduced, and the overall performance and the user experience of the online transaction platform in the Internet / supply chain financial scene are enhanced.

The invention discloses a data storage and cache optimization method and system based on Alluxio, and the method comprises the steps: obtaining a to-be-written file, and calculating the priority of the file according to the four attributes of the to-be-written file; the four attributes of the file to be written comprise file access frequency, file capacity, file importance level and file freshness; according to a file transmission strategy corresponding to the file priority, storing the file to be written into an Alluxio cache and a bottom storage; obtaining data to be read, judging whether the data to be read is in the cache or not, and if yes, returning the data in the cache to the user; if not, returning the data in the bottom layer storage to the user; and according to the association rule, performing prefetching operation on the to-be-read data and the data blocks associated with the to-be-read data, and if the cache utilization rate exceeds a set threshold value, executing cache replacement operation.

The invention provides a data caching method and device, electronic equipment and a storage medium, and the method comprises the steps: if the space occupied by persistently cached data on local storage equipment is greater than a persistent cache threshold value, deleting part of data in the persistently cached data on the local storage equipment, sending the deleted data to a local storage device so that the space occupied by the persistent cached data on the local storage device is not greater than a persistent cache threshold value, and transferring metadata of the deleted data from a cache region in the memory to a standby cache region; if the space occupied by the persistent cached data on the local storage device is smaller than the persistent cache threshold value, caching the data to which part or all of the metadata in the standby cache region in the memory belong to the local storage device, and transferring the metadata of the cached data to the cache region from the standby cache region in the memory; wherein the persistent cache threshold value is a space threshold value which is distributed by the database node on the local storage device and is used for persistent cache.

The embodiment of the invention relates to a question answering method and device, electronic equipment, a storage medium and a system, and the method comprises the steps: obtaining a question text, converting the question text into a vector text, and carrying out the word segmentation of the question text, and obtaining a word segmentation text; retrieving a preset knowledge database according to the vector text to obtain a plurality of pieces of first knowledge data, and retrieving the knowledge database according to the word segmentation text to obtain a plurality of pieces of second knowledge data; retrieving a preset keyword knowledge database according to the word segmentation text to obtain a plurality of pieces of third knowledge data; sorting the first knowledge data and the second knowledge data to obtain a preset number of fourth knowledge data, the fourth knowledge data being the first knowledge data or the second knowledge data; and determining a response text of the question text according to the third knowledge data and the fourth knowledge data. Therefore, the efficiency and the accuracy of searching the knowledge data can be improved, so that the question can be quickly and accurately responded.

The invention discloses a multi-level cache system and method based on REDIS. The multi-level cache system comprises N Node nodes, a plurality of application programs and proxy services are deployed on each Node node, and a first-level cache is configured on a driving layer closest to the application programs; simultaneously deploying a proxy cache database proxy service on a server where the application program is located, and configuring a second-level cache; and finally, the cache database Redis itself is used as a third-level cache. According to the method, by reasonably distributing cache responsibilities of different levels, the read load of the Redis database is reduced, the cache hit rate is increased, the network I / O overhead is reduced, and the overall performance and stability of the system are improved.

The invention discloses an intelligent reply method and device fusing a large language model and a knowledge graph. The method comprises the following steps: judging whether a target user has a second question or not; if yes, judging whether a time interval between the first question and the second question is within a first preset duration or not; if yes, extracting a first keyword set of the first question and a second keyword set of the second question; whether the keyword repetition rate between the first keyword set and the second keyword set is larger than a repetition rate threshold value or not is judged; if yes, screening out a target knowledge graph from a preset knowledge graph database; filling the preset cue word structure framework with the plurality of first retrieval results to obtain a first text; obtaining a second text; and inputting the second text into a preset large language model to obtain a target reply corresponding to the first question of the target user. The invention belongs to the field of intelligent customer service. According to the invention, the hit rate of large language model answering can be improved.

The invention relates to a cache data processing method and device, computer equipment, a storage medium and a computer program product. The method comprises the steps of obtaining task related data of a current running task, and inputting the task related data into a pre-trained prediction model; the prediction model is a machine learning model which is obtained based on self-attention mechanism training and is used for determining correlation among the operation tasks; a to-be-cached data list for the next task is output through the prediction model, the to-be-cached data list comprises respective caching probabilities of a plurality of data units, and each data unit comprises a data table and a data path; and determining target cache data according to the to-be-cached data list, and storing the target cache data in a preset cache. By adopting the method, the hit rate of the cache can be improved.

The utility model relates to an unmanned aerial vehicle air-drop simulation training bomb, which comprises a touch switch, a fuse, a warhead and an empennage, and is characterized in that the warhead is filled with explosives; the touch switch is fixedly arranged at one end of the fuze, the end, away from the touch switch, of the fuze is fixedly connected with the warhead, and the end, away from the fuze, of the warhead is fixedly connected with the empennage; the touch switch is coupled with a time delay circuit for delaying the trigger fuze; the fuse is provided with an electric detonator interface, and the electric detonator interface is coupled with the time delay circuit. According to the invention, trainees can better master the specific time efficiency of unmanned aerial vehicle air-drop ammunition percussion detonation in training, and the training effect is improved.

The invention provides a puncture outfit which comprises an angle rotator, a fixator and a position rotator, a plurality of mounting holes are formed in the periphery of the angle rotator, and the mounting holes limit and fix the head of a puncture needle; the fixator is connected with the angle rotator in a relative rotation mode, a plurality of guide groove holes are formed in the fixator, and the guide groove holes correspond to the mounting holes one to one; the position rotator is connected with the fixator in a relative rotation mode, a plurality of penetrating holes are formed in the position rotator, the penetrating holes correspond to the guide groove holes in a one-to-one mode so that the needle portion of a puncture needle can penetrate through the penetrating holes, and the puncture angle can be well adjusted and the puncture position can be well fixed in cooperation with the medical imaging technology before puncture. And the control on a puncture needle inserting point and a needle inserting angle is realized.

A trisection training rod is an auxiliary training instrument for trisection shooting, is a training instrument researched and developed according to the necessary conditions of accurate trisection shooting, can help an athlete to improve the sensing ability and the control ability of fingers and wrists on the gravity center of an object moving on the hands, and meanwhile increases the strength of the fingers, the wrists and the arms. Through the training of the three-point ball training rod, the hit rate of the three-point ball can be greatly improved. The whole body is in a short and thick rod shape, and the whole surface is made of damping materials. The device is characterized by being adjustable in weight, adjustable in length and adjustable in diameter of a middle hand grabbing part.

The invention provides a text retrieval enhancement generation method and device, equipment and a readable storage medium, and the method comprises the steps: obtaining document original data, and storing the document original data to a first database; performing text vectorization processing on the document original data to obtain document vector data, and storing the document vector data to a second database; retrieving a plurality of document related data from the first database and the second database based on question information input by a user; and generating a target keyword based on the related data of each document, and analyzing the target keyword through a large model to obtain a reply text for the question information. By adopting the method, the normal generation of the reply text is ensured, and the reliability, effectiveness and accuracy of the generation of the reply text are ensured.

The invention discloses an importance perception KV cache management system and method, and aims to solve the problems of long reasoning time and low throughput rate caused by low efficiency of a traditional K and V cache management strategy in a large model reasoning process. According to the method, the importance is introduced into the management strategy of the K and V caches, the cache hit rate on a relatively fast medium in storage is improved by tracking the importance of the K and V, combining the read frequency of the K and V and using the importance score as the basis of cache management, the time for reading the K and V is shortened, the overall time delay of an inference system is reduced, and the throughput is improved. The method is suitable for importance-based large model reasoning tasks with shared prefixes, the model reasoning precision can be maintained, meanwhile, the output time delay is shortened, and the throughput capacity is improved by about two times.

The invention provides a cache which comprises at least two layer structures, each layer structure comprises a first cache unit and a second cache unit, first cache access instructions of different layer structures enter different pipelines to access the first cache unit in the corresponding layer structure, and second cache access instructions of different layer structures enter different pipelines to access the second cache unit in the corresponding layer structure. The second cache access instructions of different layer structures enter different assembly lines to access the second cache units in the corresponding layer structures; the cache comprises a plurality of layer structures, the cache capacity can be expanded, the data hit rate is increased, and the overall performance of a CPU is improved; moreover, the cache access instruction of each layer structure independently accesses the respective cache unit by mistake, so that the cache processing efficiency can be ensured. The invention further provides a data processing device and a chip.

The present application provides a warehouse replenishment method, device and system, which is applied to a server, and includes: obtaining the current inventory information and vacancy identification of each shelf; the inventory information includes a location identification, and the type and quantity of goods corresponding to the location identification; based on the preset replenishment principle and the current inventory information and vacancy identification of each shelf, determining the target vacancy corresponding to each container; wherein the container is configured with goods to be replenished; the replenishment principle is used to characterize the product combination performance corresponding to each aisle based on the goods on the shelf after replenishment; the product combination performance corresponding to the aisle is proportional to the number of subsets in the product group set that can be combined with the goods on the shelf corresponding to the aisle; and allocating the container to the target vacancy for replenishment. This method can make the goods on the shelf corresponding to the aisle include more commonly used product combinations, thereby improving the product hit rate and picking efficiency.

The invention discloses a mobile robot performance system popularity racking method based on an improved TabNet algorithm. Comprising the following steps: collecting historical order data; the historical order data is preprocessed; a TabNet algorithm is improved, and a popularity model based on the improved TabNet algorithm is constructed; training a popularity model based on an improved TabNet algorithm by using the preprocessed data set, calculating a popularity score of each SKU, and performing real-time data updating and dynamic recalculation of the popularity score; and according to the SKU popularity score obtained through calculation, goods are put on different goods shelf areas. The invention provides a novel mobile robot performance system popularity shelving method based on an improved TabNet algorithm, so that high-popularity SKUs can be effectively identified and shelved in a concentrated manner through dynamic feature selection and a popularity scoring mechanism, thereby improving the order line hit rate of an RMFS system and optimizing the sorting efficiency.

The invention relates to the technical field of computers, and discloses a data prefetching method and device for a CDN edge node and computer device.The method comprises the steps that when an access request for a target channel is not monitored, a prefetching task is generated; determining a target channel matched with the prefetching task in a plurality of channels of the source station; acquiring real-time code stream data of the target channel based on a prefetching time length indicated by the prefetching task; and pre-storing the real-time code stream data to a target edge node in the CDN, the target edge node being used for returning the real-time code stream data in response to a data acquisition request. According to the method and the device, the real-time code stream data can be prefetched through the edge node of the CDN in a streaming media scene, so that the hit rate when a user requests to obtain a video file is improved, and the first screen duration when the user obtains and watches a video is reduced.