7 results about "Average memory" patented technology

The invention discloses a reinforcement learning-driven block chain consensus algorithm dynamic selection method and system, and relates to the technical field of block chains, and the method comprises the steps: carrying out the performance evaluation of a target block chain, and determining the performance state information of the target block chain, including the transaction throughput, the average CPU utilization rate, the average memory utilization rate and the performance score; determining an optimal consensus algorithm by adopting a consensus algorithm selection method based on reinforcement learning according to the performance state information; and switching the current consensus algorithm to the optimal consensus algorithm based on the optimal consensus algorithm. According to the method, the optimal consensus algorithm can be selected and switched for the block chain system in real time, evaluation and optimization of the block chain consensus algorithm are realized, and the optimal performance of the block chain system under continuously changing fault scenes and workload conditions is ensured.

The invention discloses a heterogeneous BIM model lightweight server intelligent scheduling method and system. The method comprises the steps that a data storage structure is constructed, and an engine information table, a server information table and a task information table are designed; according to the engine identification code, obtaining unit task memory consumption and server memory configuration information from the engine information table and the server information table, counting the total number of server tasks and the number of lightweight task success from the task information table, calculating a lightweight success rate, and counting the number of tasks in processing from the cache structure; based on the unit task memory consumption, the server memory configuration, the lightweight success rate and the number of tasks in processing, calculating a memory usage amount, a memory occupancy rate and a residual computing power, and determining a weight factor according to the lightweight success rate to obtain a weighted residual computing power; and realizing server scheduling based on the calculated memory occupancy rate and weighted residual computing power. By adopting the method, model resource consumption and server residual computing power can be dynamically matched, and the average utilization rate of a cluster memory is improved.

The embodiment of the application provides a kind of Java probe memory occupation amount determination method, equipment, medium and program product, the method includes: in response to target method being called, based on Java probe, the target span instance corresponding to target method is created, target method is any one in the multiple methods of application program and provides different functions, target span instance is used to record the running state data of target method.After creating target span instance, control span instance counter to carry out plus one processing, to determine the total number of current living span instance.Further, according to the total number of current living span instance and the average memory occupation of span instance, the total memory occupation corresponding to Java probe is determined.Java probe memory occupation is determined by the average memory occupation of the number of current living span instance and span instance in Java probe, so that Java probe memory occupation can be accurately determined.

This application relates to a method and apparatus for testing memory performance degradation. The method includes: monitoring user processes of a cloud host and a host machine, and terminating user processes when the user process load exceeds a first threshold, so that the cloud host and the host machine are in an idle state; repeatedly executing a first file, and sequentially executing a synchronization command and a cache clearing command after each execution of the first file to release slab objects and page caches of the cloud host; obtaining environment variables to limit the number of parallel threads and CPU affinity of the host machine; obtaining a control group to limit the memory usage of user processes on the host machine; repeatedly executing a first command, and sequentially executing a synchronization command and a cache clearing command after each execution of the first command to release slab objects and page caches of the host machine; statistically analyzing the execution results of the first file and the first command, and calculating the average memory bandwidth of the cloud host and the host machine; and calculating the memory performance degradation ratio of the cloud host based on the average memory bandwidth of the cloud host and the host machine.

The invention discloses a WiFi feature data structure optimization method for a mobile terminal, and the method comprises the steps: dividing WiFi feature data into AP, RATIO and MAX unit data for storage, loading MAX units in batches during the APP analysis of the mobile terminal, dispersing the memory use time, and reducing the memory peak value; the assembled unit data follows a fixed byte occupation rule, so that the size of the unit is predictable, excessive memory is prevented from being reserved, and the peak value is further reduced; data are assembled according to a specific sequence rule, a compact structure is formed, redundant intervals are reduced, the storage density is improved, a mobile terminal can efficiently access the data, the average memory overhead is reduced, meanwhile, the data are stored in a fixed sequence, analysis can be accessed in a memory address sequence, the memory utilization efficiency is improved, and the average overhead is also reduced; in addition, a fixed rule and an ordered structure are convenient for accurately distributing and releasing the memory, timely releasing the unit memory, avoiding invalid occupation leakage, reducing memory garbage, improving the memory utilization rate, reducing the management complexity and reducing fragment-related garbage.

The invention particularly relates to a high-speed routing table lookup method, and aims to solve the problems of limited lookup efficiency and high dynamic updating cost in the prior art. The method comprises the following steps: firstly, constructing a binary search tree which takes prefix length as a keyword and optimizes node positions according to use frequency; secondly, configuring a hash table for each node in the tree for storing a routing prefix and a jumping mark capable of pointing to a better search path; in the searching process, matching items or marks are searched in parallel in the node hash table from the tree root, rapid sub-tree jumping is achieved through the marks, and step-by-step traversal is avoided. When a routing prefix is inserted or deleted, only relevant marks need to be locally updated along a tree path, and global reconstruction is not needed. According to the method, the average memory access times and the query time delay are remarkably reduced, meanwhile, the dynamic updating of the routing table is supported at relatively low overhead, and the method is suitable for a network forwarding scene with high requirements on throughput, delay and updating frequency.

The present application relates to the field of network communication technology, and in particular to an auditing method and apparatus. The method comprises: upon detecting an audit of a target network device, determining whether the number of network devices currently being audited is less than a first threshold; if it is determined that the number of network devices currently being audited is less than the first threshold, obtaining the average CPU usage and the average memory usage of the controller within a preset time period before the current time; determining whether the obtained average CPU usage of the controller is less than a second threshold, and whether the average memory usage is less than a third threshold; if it is determined that the obtained average CPU usage of the controller is less than the second threshold, and the average memory usage is less than the third threshold, then auditing the target network device.