34results about How to "Improve lookup performance" patented technology

The invention discloses an index structure based on a dictionary tree, a data query optimization method and a main memory manager, and belongs to the field of data storage. The index structure is composed of containers, wherein the containers store T nodes and S nodes; complete keys are divided into partial keys; the T nodes store the front m / 2 bits of the partial keys, and the S nodes store the rear m / 2 bits of the partial keys; and the partial keys passing through a certain node path from the root node of the dictionary tree are connected to form the complete keys. And the index structure ismore compact, so that the main memory space utilization rate is improved. The T node is directly skipped to a subsequent brother node by using a skipping relay, so that the comparison frequency is reduced, and the search performance is improved; a T node skip list is used for skipping a T node to an S node, a container skip list is used for skipping the S node to the T node, and the S node directly skips to the nearest target, so that a large number of unnecessary entries are skipped, and the searching performance is improved; and the index structure and the main memory address are completelydecoupled by a pointer P which is analyzed by the main memory manager and contains each hierarchical structure ID.

The invention discloses a navigation method with one-button operation, which comprises the following steps: dividing the service range of the navigation center into several areas, and establishing a central database; the navigation terminal obtains GPS information and base station information, and uploads them to the navigation center; the navigation center judges If the GPS information is valid, the location of the vehicle is located; if it is invalid, the location of the vehicle is estimated according to the base station information; the central database is searched according to the location of the vehicle, and the customer service number of the center in the user's area is obtained, and sent to the navigation terminal; the navigation terminal dials the customer service number of the center, The agent finds out the relevant destination information for the user and confirms it; after the call ends, the navigation terminal connects to the navigation center through the mobile network and requests to download the destination list information. This kind of navigation method can ensure that the navigation center can locate the vehicle to the greatest extent and meet the needs of users as much as possible; and can effectively improve the communication effect between the user and the seat personnel, and improve the search efficiency.

The present invention discloses a method and device for adding, searching for and deleting a key in a hash table, wherein, the method for adding a key in a hash table includes: calculating both the bucket address corresponding to the key to be added and the address extension of the key to be added in each hash sub-table according to a hash function; generating the fingerprints of the key to be added corresponding to each hash sub-table, wherein the fingerprint of the key to be added corresponding to one hash sub-table includes: the address extension of the key to be added in the hash sub-table, and the bucket address corresponding to the key to be added and the address extension of the key to be added in other hash sub-tables; and when determining that there is no conflict, storing the fingerprint of the key to be added corresponding to the selected hash sub-table into the selected hash sub-table. The solution of the present invention enables that the fingerprint of each key can only match one fingerprint in one hash sub-table, thus when searching for a key, at most once is required to visit an original key table, thereby improving the search performance, and simplifying the search process.

The invention discloses a method for checking a null pointer in a source code, comprising: a computer system construction referencable variable type dictionary, a computer system receiving an instruction for scanning the null pointer, formatting the source code to be checked; scanning the global code zone of the source code to be checked and obtaining an uninitialized global variable list corresponding to the variable statement type in the dictionary; scanning the method zone of the source code to be checked and obtaining the uninitialized local variable list corresponding to the variable statement type in the dictionary; acquiring uninitialized global variable list in method based on the uninitialized global variable list; scanning the method zone of the source code to be checked, searching the called variable name, if the called variable name is in the uninitialized local variable list or uninitialized global variable list in method, the found called variable name is the null pointerin the source code. The invention also discloses a computer system.

The embodiment of the invention provides a method and system for positioning an interference source based on a single stationary orbit satellite. The system comprises a signal collecting device, a reference source signal emission device, a signal processing device and an interference source positioning device, wherein the signal collecting device is used for collecting an interference signal of the stationary orbit satellite, the reference source signal emission device is used for sending a reference source signal to the stationary orbit satellite, and receiving a reference source return signal sent by the stationary orbit satellite according to the reference source signal, the signal processing device is used for conducting error removal on actual Doppler frequency shift of the interference signal and the reference source return signal, and the interference source positioning device is used for determining the position of the interference source according to the interference signal and the reference source return signal. Accurate positioning on the interference source of the satellite can be achieved through the only one stationary orbit satellite, the finding capability of the satellite on the interference is greatly improved, and a plurality of restraint conditions of interference finding are reduced.

A classification management and index method for the extensive package box type warehouse is characterized by that a three-layer structure is used to manage the box-type warehouse and the database technique is used for classification management and index and increasing the additional description information about package boxes. Its advantage is high management and index efficiency.

The invention provides a method and a system which temporarily store and process charging tickets and comprises the following steps that the first step is that whether key fields of parts of the charging tickets exist in a memory is judged after receiving the part of the charging tickets from a charging gateway, and the second step is that if no key fields exist, a position of an available disk file is obtained from the memory and the part of the charging tickets is stored in the available disk file which is corresponding to the position of the available disk file, otherwise, the existing charging tickets and the part of the charging tickets which are corresponding to the key fields are combined to generate a new charging ticket. Therefore, by adopting the method and system of the invention, main key data is stored in the memory, ordinary data is stored in the disk, the search of the part of the charging tickets can be realized in the memory, so the property of searching performance ration databases is greatly improved, thus reducing the cost of the system and also improving the property of the system.

The invention discloses an index structure based on a dictionary tree, a data query optimization method and a main memory manager, belonging to the field of data storage. The index structure is composed of containers. The container stores T nodes and S nodes. The complete key is divided into partial keys. The T node stores the first m / 2 bits of the partial key, and the S node stores the last m / 2 bits of the partial key. From the dictionary tree The partial keys passed on the path from the root node to a certain node are connected to form a complete key. The index structure is more compact, thereby improving the utilization rate of main memory space. Use jump successor to directly jump T node to subsequent brother node, thereby reducing the number of comparisons and improving search performance; use T node jump table to jump T node to S node, use container jump table to jump S node to T node, Jump directly to the nearest target, thereby skipping a large number of unnecessary entries and improving the search performance. The pointer P containing the IDs of each hierarchical structure parsed by the main memory manager fully decouples the index structure from the main memory address.

The invention discloses a method and system for searching NDN data names. The method includes: designing and implementing a data structure candidate alignment migration array for GPU-based data name searching; running a CATA construction algorithm and an update algorithm on a CPU; converting CATA from The CPU end is transmitted to the GPU end; the CATA-based data name search algorithm is run on the GPU; the operating parameters of the GPU are continuously adjusted to optimize the search performance; the search results are transferred from the GPU end to the CPU end. The present invention provides a data name search method based on the GPU data structure CATA, so that the data name search problem can be well solved, not only can realize the data name search at the line speed, but also greatly reduce the storage cost, and has important practical application value.

The embodiment of the invention provides a method and system for searching for a useless file. An older-version application program installation directory meeting an installation directory feature corresponding to the application program in a preset feature library besides the installation directory in the parent directory of the application program installation directory obtained according to a registry, and the useless file is searched for in the searched older-version application program installation directory. Due to the fact that the new-version and older-version application program installation directories generally have a certain installation directory feature, the older-version application program installation directory can be found through the installation directory feature, and therefore the useless file is searched for in the older-version application program installation directory. Therefore, according to the method and system for searching for the useless file, the useless file search effect is improved, the situation that the useless file in the older-version installation directory exists in terminal equipment for a long time and occupies the disk space is avoided, and efficient operating of the terminal equipment is effectively guaranteed.

The present invention relates to the technical field of route search, and discloses a method for IPv6 route search, comprising the following steps: 1. Component data structure: S1. Build a binary tree with 2001 as the root node and composed of the first 16 bits of the route prefix; S2. Build segment tables and routing buckets, and obtain the corresponding threshold under the balance between the search efficiency of the algorithm and the memory occupied by the algorithm. When the number of routing entries is less than the threshold, the routing bucket organization is used, and when the number of routing entries is greater than the threshold, the segment is used. Table organization; S3. Routing buckets linked on the segment table. The IPv6 route lookup method has the advantage of being able to realize IPv6 fast lookup, and can better meet the need for IPv6 lookup.

The embodiment of the invention provides a cache-coloring memory allocation method and device for search trees. The method includes: constructing the first search tree; determining that the first search tree has N layers of nodes; acquiring the number of the nodes which each layer in the N layers of nodes has; allocating corresponding colors in coloring colors of a memory for each layer of nodes according to the number of the nodes which each layer in the N layers of nodes has, wherein the corresponding colors allocated for all the layers of nodes are mutually different; and generating the second search tree after cache coloring according to the colors corresponding to each layer of nodes and the memory corresponding to the colors. Only the memory which can be mapped to the corresponding coloring colors is allocated when the memory is allocated for each layer of nodes according to the colors corresponding to each layer of nodes in the second search tree, mapping relationships exist between the memory and a Cache, thus an effect that the different layers of nodes cannot mutually compete for the Cache can be guaranteed, and thus search performance is improved.

The on-board fast route lookup system combined with the longest prefix match and exact match, including the exact match lookup module and the longest prefix match lookup module, the exact match lookup module includes a hash lookup circuit, a hash hash table and an aging update management circuit; The long prefix matching search module includes a compressed Trie tree search circuit, a Trie tree node storage area and a routing table item storage area. Compared with the traditional longest prefix matching search structure, the method of the present invention adds an exact match search as a high-speed cache of active routing table entries by establishing a route search structure combining the longest prefix match and exact match, greatly improving the IP packet rate. The routing lookup speed of the text stream effectively reduces the average lookup time. In addition, by building an aging update management circuit that accurately matches and looks up, the automatic learning and adding operations of active routing table items are realized, and the aging check and processing of routing table items are completed. Delete operation to ensure efficient management and lookup performance of active routing table entries.

The invention provides a data packet classification method and system based on a convolutional neural network. The method includes merging each rule set in the training rule set to form multiple merging schemes, determining the optimal merging scheme for each rule set in the training rule set based on performance evaluation; combining each rule set in the training rule set and the target rule set The distribution of the prefix combination is converted into an image, and the convolutional neural network model is trained with the image and the corresponding optimal merging scheme; the target rule set is classified based on the image similarity, and the corresponding hash table is constructed for packet classification. The method of the invention can remarkably improve the data packet search performance, increase the data packet search speed, and increase the update speed of the rules. Through the mutual cooperation between the online system and the offline system, the system of the present invention can ensure that the online system realizes efficient search of data packets and rapid update of the rule set, and can monitor the update of the rule set to always reflect the latest state of the network.