20results about How to "Improve cache hit ratio" patented technology

This invention discloses a cache adjustment method, apparatus, device, and computer-readable storage medium, comprising: collecting access characteristic data from the client layer, object storage daemon layer, and device layer respectively; determining the object access mode of the system based on the access characteristic data; and adjusting the cache priority of each access task, the cache capacity ratio between the client layer and the object storage daemon layer, and the data residence time in the cache layer according to the object access mode; wherein the cache layer includes the client layer and the object storage daemon layer. This invention improves cache hit rate and overall read / write throughput performance, and reduces access latency.

This invention discloses a joint optimization method for mobile edge computing and service caching assisted by multiple UAVs, comprising: constructing a multi-UAV-assisted edge computing system model; constructing a communication model; dividing the system decision time into a set of large-timescale periodic sequences and a set of small-timescale time slot sequences, each large-timescale periodic sequence containing multiple consecutive small-timescale time slot sequences; deriving expressions for the total task processing latency and total energy consumption of the system in any time slot; establishing a joint optimization objective function with the goal of minimizing the total system latency and total energy consumption; constructing a Markov decision mechanism; employing the SAC deep reinforcement learning algorithm to implement the dual-timescale decision-making of the system; and solving the joint optimization objective function to obtain the optimal joint optimization scheme. This invention can significantly improve the edge service cache hit rate, optimize UAV load balancing, and reduce task processing latency and overall system energy consumption.

This application provides a method, system, device, and medium for virtual terminal resource preloading and caching management based on the ISOBUS standard. This application innovatively constructs a multi-level caching architecture and combines a Markov chain-based resource prediction engine with adaptive memory management technology to achieve intelligent preloading and efficient caching of interface resources. Addressing the pain points of existing ISOBUS VT systems, such as high resource loading latency, low cache hit rate, and insufficient memory utilization, this invention provides a targeted technical solution. Experimental data verifies that this technical solution can reduce resource loading time by 50% to 70%, increase cache hit rate to over 92%, improve memory utilization by 40%, and reduce system crash rate by over 80%.

The present disclosure provides a blockchain contract processing method and device, electronic equipment and storage medium, comprising: loading N candidate smart contracts in batches from the to-be-processed smart contracts; in response to identifying that the candidate smart contract needs to perform a target external data read-write operation, judging whether the target external data exists in the local cache; in response to the target external data not existing in the local cache, performing saving processing on the first execution context data of the candidate smart contract, and determining a target contract from the to-be-processed smart contracts except the N candidate smart contracts; replacing the target contract with the candidate smart contract, and loading the target contract to the execution slot of the execution memory corresponding to the candidate smart contract, and executing the target contract. When it is found that the candidate smart contract needs to wait for the database to return, the context is saved and the execution is switched, avoiding the idling of the execution unit, improving the hardware utilization rate, and at the same time, by supporting batch preloading, the cache hit rate and execution continuity are improved.

The application relates to an active cache routing method and system of a satellite content distribution network, the satellite content distribution network comprising at least one satellite node. The method comprises: extracting first semantic information features of distributed content and second semantic information features of historical popular content by using a deep learning model; calculating semantic similarity between the first semantic information features and the second semantic information features; calculating a satellite semantic cache priority value of a satellite node corresponding to the distributed content according to the semantic similarity; and planning an active cache routing path of the distributed content according to the satellite semantic cache priority value of each satellite node. The application can enable the satellite to automatically make a cache decision of the distributed content in the content distribution process, and improve the efficiency of active cache by combining content distribution and cache deployment.

The invention discloses a strategy optimization method for cache garbage collection, relates to the technical field of computer cache management, and is used for solving the problem of low cache collection efficiency. According to the method, a cache storage area is abstracted into a plurality of cache partitions, a partition state vector composed of capacity, access and fragmentation degree operation indexes is constructed for each partition, and a strategy evaluation thread periodically reads the partition state vectors to calculate garbage recycling pressure and recycling priority. The upper limit of the number of scanning objects and the upper limit of time slices are distributed for the target cache partition by combining the idle time proportion of the processor, and a recycling execution thread conducts incremental recycling in a batch mode according to a strategy list and records the recycling progress; after recovery, index weight and scanning configuration are adjusted according to release capacity, fragment degree change, hit times and back-to-source request change, so that garbage recovery gives consideration to release space and overhead, the hit rate and the memory utilization rate are improved, and pause and invalid recovery caused by centralized recovery are reduced.

The invention discloses a batch data processing method and device suitable for a database, equipment and a medium, relates to the technical field of data processing, and is applied to a processor. Monitoring whether a target instruction meeting a preset stride condition exists in a data processing instruction stream corresponding to the preset database in the target assembly line or not, if yes, entering an instruction pre-execution mode, vectorizing the target instruction and a subsequent instruction having a dependency relationship with a register of the target instruction, and outputting the vectorized instruction; performing instruction pre-execution based on the vectorized instruction and a first execution unit, and storing an instruction pre-execution result to a first physical register file and a result cache region; and when the data processing instruction stream in the target assembly line reaches the input end of the second execution unit, matching the instruction pre-execution result in the result cache region, and completing batch data processing operation based on an information matching result. According to the invention, the efficiency and accuracy of batch data processing are improved.

This invention relates to a lock-free I / O scheduling method, apparatus, and storage medium for cloud PCs. Responding to a client session request initiated by a client, the method controls a metadata coordinator to allocate storage nodes to the client and, within those storage nodes, allocates I / O shards for the client, obtaining allocation information. The metadata coordinator then returns the allocation information to the client, and the client caches this allocation information to facilitate the use of the allocation information to fulfill I / O requests. Each I / O shard includes an I / O thread bound to a first CPU core, a data plane development toolset driver, a storage performance development toolset driver, a first lock-free circular buffer, and a second lock-free circular buffer. By binding the CPU core and using lock-free circular buffers, thread switching, lock contention, and most memory copying are eliminated, reducing I / O processing latency from milliseconds to hundreds of microseconds and significantly reducing latency jitter.

The invention provides a chip calculation reasoning method and system for a hybrid expert model, and relates to the technical field of artificial intelligence calculation, and the method comprises the steps: obtaining input data, calculating an expert activation weight, carrying out the time sequence tracking, carrying out the clustering analysis, recognizing an expert combination, and preloading the weight to a shared cache; selecting a target expert based on the activation weight, and loading the weight; fragmenting the data, constructing an assembly line level, and mapping the data to different computing units; and each unit is controlled to execute calculation and weighted summation according to an assembly line to obtain a result. Through expert activation mode prediction and pipeline parallel processing, the reasoning efficiency is improved, and the memory access overhead is reduced.

Embodiments of the present application provide a data processing method, device and equipment, readable storage medium and program product, the method comprises: obtaining the starting address of the data block to be translated; determining the starting address of the target data block based on the starting address of the data block to be translated, the target data block is the data block obtained by translating the data block to be translated; if it is determined that the target data block is not stored in the preset special memory based on the starting address of the target data block, the data block to be translated is translated into the target data block; the target data block is stored in the preset special memory, and the target data block is executed in the preset special memory; in this way, dynamic binary translation is performed based on the preset special memory, and the cache hit rate of the target data block in dynamic binary translation is improved.

The present disclosure relates to a structured query statement processing method, device, equipment and readable storage medium. By obtaining a plurality of to-be-processed structured query statements, a corresponding target plan tree is generated for each structured query statement, the target plan tree corresponding to each structured query statement is mapped to obtain a vector plan tree of each structured query statement, based on the vector plan tree of each structured query statement, a special data processor is called to process each structured query statement to obtain a processing result of each structured query statement. Compared with the prior art, by calling the special data processor to process each structured query statement based on the vector plan tree of each structured query statement to obtain the processing result of each structured query statement, the processing functions inside each structured query statement can be batch processed, the processing efficiency is improved, the number of function calls is reduced, the cache is effectively utilized, and the cache hit rate is improved.

The invention discloses a method and system for reducing KV storage and loading expenditure in a retrieval enhancement generation system, and the method comprises the steps: S1, compressing a key value cache block of a pre-calculated external document to reduce the storage and loading expenditure of the key value cache block; s2, when the requested key value cache block does not hit the cache, searching and replacing the key value cache block with similar semantics in the cache according to the similarity of retrieval results; and S3, performing comprehensive value evaluation on the key value cache blocks in the cache based on the frequency, the recency and the first retrieval score, and executing cache replacement according to an evaluation result to maximize a cache hit rate. Through an innovative two-stage compression technology, storage occupation and data loading overhead of mass external document pre-calculation key value cache are remarkably reduced. Based on a key value cache reuse strategy of semantic similarity, high-delay disk input / output access is converted into rapid memory access, and the performance bottleneck of data loading is effectively relieved.

The application relates to a federated learning assisted edge caching method based on an AE-DDPM model, and comprises the following steps: an edge computing network system model is established, including a base station, a remote cloud server and users; an AE-DDPM model based on federated learning is used for training; global prediction content popularity is obtained, and the most popular N contents are cached according to the cache capacity of the base station; the application extracts a user data potential feature vector through an AE model, reduces data dimension and sparseness, and then learns data distribution through a DDPM model to generate high-quality content popularity prediction; the application deploys a cache unit on an edge node, enables users to quickly obtain the pre-cached contents on the node, effectively improves the cache hit rate, reduces the time delay of the users in obtaining the contents, significantly improves communication efficiency, and simultaneously reduces the risk of user privacy leakage.

Embodiments of the present disclosure provide a memory system and an operating method thereof, a storage medium and a memory controller, wherein the memory system comprises at least one nonvolatile memory device and a memory controller coupled with the at least one nonvolatile memory device; wherein the memory device stores a multi-level mapping table for implementing mapping of a logical address to a physical address; the memory controller comprises a cache storing part of the multi-level mapping table; and the memory controller is configured to perform a random read operation on data stored in the memory device; wherein, in response to a random read range corresponding to the random read operation satisfying a preset condition, the capacity of the cache for storing different levels of mapping tables is adjusted.

This application provides a system memory management method, circuit, and electronic device. The method includes: during the initialization phase, storing a first number of consecutive conversion relationships extracted from a memory module into a cache module, wherein the conversion relationship refers to the relationship between the virtual address and the physical address of video data; when a second number of preceding conversion relationships among the first number of consecutive conversion relationships stored in the cache module are used up, extracting a second number of consecutive conversion relationships from the memory module; wherein the extracted second number of consecutive conversion relationships have a linear temporal progression relationship with the conversion relationships currently stored in the cache module; and using a cross-replacement strategy, replacing the used second number of preceding conversion relationships in the cache module with the second number of consecutive conversion relationships extracted from the memory module.

The application relates to a data caching method, system and device, computer equipment and a storage medium. The method comprises the following steps: in response to an access request for target data, determining a caching priority of the target data; storing the target data in a pre-set dynamic caching area according to the caching priority, obtaining target dynamic caching data of the dynamic caching area; at a pre-set updating time, updating original static caching data in a pre-set static caching area according to the target dynamic caching data, obtaining updated static caching data; and the updated static caching data is used for providing repeated data access for the target data. The method can improve the hit rate of data access.

The embodiment of the invention provides a memory priority double-layer LRU cache management method oriented to limitation of a swan sandbox, which belongs to the technical field of data processing, and specifically comprises the following steps: step 1, creating a double-layer LRU cache architecture, initializing a memory cache and a disk cache, and setting a memory priority space allocation strategy based on 200MB total space limitation; 2, a double-layer LRU cache architecture is initialized, disk cache constructs an LRU queue by scanning a cache directory and matching access time records, and therefore the LRU state is recovered from a disk; 3, when the data needs to be stored in the cache, preferentially judging whether the memory cache is available or not; 4, deleting the oldest disk cache file according to an LRU strategy, and updating a cache state; and 5, adding the file information to the tail of the disk LRU queue, and updating the access time of the disk LRU queue. Through the scheme of the invention, the caching performance, the flexibility and the real-time performance are improved.

This invention discloses a method and system for accessing medical image data in a Kubernetes cluster environment, comprising: receiving an access request from a user client including a DICOM identifier and a unique request identifier; determining the target Pod based on the unique request identifier; sending the DICOM identifier to the target Pod and querying its local cache; if a match is found, returning the matched target medical image data to the user client; if a match is not found, querying a preset Redis distributed shared cache based on the DICOM identifier; if a match is found, filling the target medical image data back into the target Pod's local cache and returning it to the user client; if a match is not found, retrieving the target medical image data corresponding to the DICOM identifier from a preset medical image database, storing it in the Redis distributed shared cache, and returning it to the user client. This invention achieves improved overall cache hit rate while ensuring low-latency access.

This application provides a front-end interface request management method, electronic device, storage medium, and program product. The method includes: obtaining context features of interface requests initiated by a front-end application; the context features are used to characterize the real-time attributes, execution environment, and user intent of the interface request; classifying the interface requests based on the context features and determining processing strategies corresponding to different categories; and performing corresponding processing operations on the interface requests according to the processing strategies, the processing operations including at least one of request aggregation, cache reading, or real-time forwarding. Different categories of requests can be matched with the most suitable processing strategies, and differentiated scheduling of request traffic is achieved, reducing network resource waste caused by redundant requests, improving cache hit rate and response speed, and constructing a full-link request management system with context awareness, dynamic decision-making, and strategy coordination capabilities.

The invention discloses a distributed system oriented to a deep learning recommendation model and a data caching method in the technical field of distributed deep learning system communication, and the method comprises the steps: obtaining a label of each query request data stream passing through a central switch in each measurement period; according to the method, the size of the corresponding data flow is calculated, the hot query request of each distributed node is determined based on the size of each data flow, cluster communication flow can be measured and analyzed in real time, global embedded table access hotspots are counted, and hotspot data are pushed to each node for caching, so that the cache hit rate is effectively improved, the All-to-All communication traffic is reduced, and the user experience is improved. According to the method, the requirement for accessing a GPU video memory is reduced, and the problems of dense data throughput and high access delay caused by the fact that the scale of a recommendation model is too large and embedding operation involves frequent and random table item access at present are solved.