47results about How to "Avoid running out of memory" patented technology

The invention is applicable to the technical field of communication and provides a backup method of an application program in a communication terminal and the communication terminal. The backup method comprises the following steps of: detecting the current residual memory space of the communication terminal before downloading a new application program; obtaining the memory occupation value of the new application program; judging whether the current residual memory space is more than the memory occupation value of the new application program; if so, executing an application downloading step; otherwise, executing an application backup step; backing up at least one application program in the communication terminal into a server, automatically deleting the backup application programs in the communication terminal, and returning back to a memory judging step; and downloading the new application program to the communication terminal. Therefore, the problem that the installation is abandoned due to insufficient memory during or after the process of downloading the application program can be avoided, so that the time and communication flow of a user are effectively saved; and meanwhile, the user can selectively back up the application programs which are not usually used into the server, so as to facilitate the continuous use after the application programs are recovered.

The invention discloses a backup method and a backup device of a link-state advertisement. The backup method comprises the following steps of: receiving LSA (Link-State Advertisement) relevant information transmitted by a main plate by a backup plate; and generating the LSA by the backup plate according to the LSA relevant information. In the invention, a backup plate of a distributed type devicegenerates the LSA, and a main plate is not need to backup great amounts of routing information to the backup plate through plate communication, thereby avoiding the problem of memory exhaustion due to more routing information overstocked by the plate communication.

An embodiment of this invention solves memory shortage that occurs in enlargement processing of an image because a web application holds data of an image or the like not as binary data but as character string data. In the embodiment, a processor in an information processing apparatus executes a program including a first program layer with an instruction set to be interpreted and executed by the processor and a second program layer with an instruction set interpreted in advance by a unit other than the processor in the following way. That is, the first program layer draws an image, and divides and enlarges the drawn image using a memory. Image data of each of a plurality of images obtained by the division and enlargement is output from the second program layer to an external device.

A method for preventing off-line request forgery includes: enabling a Portal Web server to store a fixed secret key, generate a random secret key for each user passing through authentication, generate a final encrypted string by means of two-time encryption according to the random secret key and the fixed secret key, and transmit the encrypted string to the user by carrying the encrypted string in a response message; and performing back-stepping decryption for the encrypted string to analyze a source IP (internet protocol) address corresponding to the encrypted string if an off-line request of the user carries encrypted string information, comparing the analyzed source IP address with a source IP address of the user in the message, allowing the user to be off-line if the comparison is consistent, and if the comparison is inconsistent, the request message is considered to be forged, and making no response. By the aid of the technical scheme, internal memory of the Portal Web server is released, and performance and operating efficiency of the Portal Web server are improved.

An embodiment of this invention is directed to an information processing apparatus capable of performing high-speed processing and preventing memory shortage even when executing a hybrid application. According to the embodiment, an information processing apparatus that executes a program including a first program layer with an instruction set to be interpreted and executed by a processor and a second program layer with an instruction set interpreted in advance by a unit other than the processor includes the following arrangement. That is, the processor includes a plurality of interpretation units configured to interpret the first program layer. A first interpretation unit is provided in an operating system that operates in the processor, and a second interpretation unit is provided in the second program layer.

The invention discloses a mobile robot navigation and positioning system based on GNSS/INS/LIDARSLAM information fusion, and belongs to the field of robot navigation. According to the system, an extended Kalman filter is adopted to fuse a GNSS positioning result, IMU data and a relative pose of the mapping positioning of a laser radar, so as to obtain a positioning result of a mobile robot. For anSLAM algorithm with a large calculation performance requirement, the system adopts a 3D laser radar SLAM technology based on feature extraction, and the calculation amount of point cloud matching isreduced through feature pre-extraction, so that the navigation and positioning system can operate on an embedded platform.

The invention discloses a method for establishing a multi-level virtual maintenance training system of a complex equipment vehicle. The method comprises the following steps: 1, acquiring the information of the whole complex equipment vehicle and each part thereof; 2, constructing a rough mold for visual display and a fine mold for virtual disassembly and assembly of the whole complex equipment vehicle and all parts thereof; 3, optimizing a visual display three-dimensional coarse mold and virtually disassembling and assembling a three-dimensional fine mold; 4, establishing a virtual disassemblyand assembly planning sequence of the optimized virtual disassembly and assembly three-dimensional precision mold; 5, constructing a multi-level virtual maintenance self-learning training mode of thecomplex equipment vehicle; 6, constructing a multi-level virtual maintenance guide training mode of the complex equipment vehicle; 7, constructing a multi-level virtual maintenance autonomous training mode of the complex equipment vehicle; and 8, integrating and packaging the multi-level virtual maintenance training system of the complex equipment vehicle. According to the method, the rough modeland the fine model are established, training is smooth, the truth and immersion are good, the virtual decomposition and assembly object model and the hierarchical relationship thereof are established, and later virtual maintenance training is facilitated.

The invention provides a flow control method and system for a data link layer in a mobile communication system. The flow control method comprises the following steps: initializing flow control threshold values, comprising a cell-grade flow control threshold value and a bearing-grade flow control threshold value of a PDCP layer and an RLC layer,; when a data message from a GTPU layer is received, judging a load condition of a base station by the PDCP layer according to a current congestion zone bit and processing the load condition; carrying out PDCP downstream flow processing; meanwhile, and carrying out pre-judgment on the cell-grade and bearing-grade congestion zone bits by the PDCT layer according to the length of a current downstream message queue and entity quantity downstream buffering zones; carrying out RLC downstream flow processing; meanwhile, and carrying out pre-judgment on the cell-grade and bearing-grade congestion zone bits by the RLC layer according to the length of a current downstream message queue and entity quantity of downstream buffering zones; utilizing pre-judgment results for flow control processing of subsequent new data messages. The flow control method and system in the technical scheme can be used for flow control on a data link efficiently and reasonably and ensure the stability of a base station system.

The invention provides a picture resource recycling method and apparatus. The picture resource recycling method comprises the steps of packaging a picture resource recycling program to a picture resource loading catalogue; packaging a unified resource management interface for the picture resource loading catalogue; and calling the picture resource recycling program through the resource management interface according to preset recycling moment for recycling useless picture resources. Therefore, when the recycling moment is triggered in a loading process, the picture resource recycling program can be called to recycle the useless picture resources, so that it is ensured that the useless picture resources can be recycled in time, so as to release memory occupied by the useless picture resources, avoid insufficient memory of Native Cache, and further avoid an out-of-memory (OOM) problem.

Embodiments of the invention provide a sample serialization method and apparatus, and relate to the technical field of machine training. The method comprises the steps of obtaining character strings in to-be-serialized samples; according to corresponding relationships between the character strings and management servers, determining the management servers corresponding to the character strings; sending the character strings to the corresponding management servers, thereby enabling the management servers to convert the received character strings into corresponding serialization IDs according to mapping tables maintained by the management servers, wherein the character strings in the mapping tables maintained by different management servers are different; receiving the serialization IDs corresponding to the character strings and returned by the management servers; and according to the received serialization IDs corresponding to the character strings, converting character strings in sample data into corresponding serialization IDs. According to the method and the apparatus, the query time of the serialization IDs of the character strings is shortened, so that the sample serialization time can be shortened and the serialization efficiency can be improved.

The invention provides a video monitoring system. The system comprises: a loading unit, which is used for loading channel data of all image obtaining units to a memory during the initialization process of the video monitoring system; an obtaining unit, which is used for obtaining corresponding channel data of any of the image obtaining units from the memory according to a checking instruction of a user; and a display unit, which is used for displaying a video corresponding to the channel data and displaying a video corresponding to the channel data according to a selection instruction of the user. In addition, the invention also puts forward a video monitoring method. According to the technical scheme, all data of a camera can be loaded to the memory by one time, thereby preventing abnormal phenomena that the memory is exhausted or the camera channel is occupied and the like during frequency video switching.

The invention discloses an entrance guard voiceprint identification system and device. The entrance guard voiceprint identification system comprises a camera, a voice module, a voiceprint acquisitionmodule, a voiceprint signal converter, a central processing module, a safety module, a voiceprint data storage module and a voiceprint identification module, wherein an output end of the safety moduleis connected with input ends of the camera, the voice module, a warning unit and a building security monitoring chamber, the warning unit is used for reminding a security worker of a failure person who attempts to open a door for many times, the building security monitoring chamber is used for viewing persons outside an entrance guard by the camera, an output end of the voiceprint acquisition module is connected with an input end of the voiceprint signal converter, and an output end of the voiceprint converter is connected with an input end of the central processing module. The entrance guardvoiceprint identification system is reasonable in design and is convenient to use, the security worker is convenient to manage the entrance guard very well by the voiceprint identification system, and meanwhile, the safety of the entrance guard system is improved.

The invention discloses a method and system for brushing an MAC address forwarding table entry. The method comprises the following steps: judging whether a to-be-brushed first MAC address forwarding table entry is the same as a second MAC address forwarding table entry blocked in a brushing chain table; if so, judging whether the first MAC address forwarding table entry performs counteraction operation, updating operation or coverage operation on the second MAC address forwarding table entry according to an operation mode of the first MAC address forwarding table entry and the second MAC address forwarding table entry, and performing corresponding operations according to a judgment result; and otherwise, caching the first MAC address forwarding table entry in the brushing chain table. In the case of blockage of the brushing MAC address forwarding, by adoption of the method and system, the number of accumulative MAC in the brushing chain table can be maximally reduced, and it is ensuredthat no repeated MAC address forwarding table entry exists in the brushing chain table, and only one MAC address forwarding table entry exists, the situation of memory use up is avoided, and thus thestability of network performance is improved.

The invention discloses a high-performance k-mer frequency counting method and system based on a clustering algorithm, and relates to the technical field of bioinformatics, and the method comprises the steps: reading a genome format file through a CPU, and obtaining a gene sequence reads with the length being L; transmitting the file to a GPU (Graphics Processing Unit) through a CUDA (Compute Unified Device Architecture) stream, splitting a gene sequence by adopting a character string splitting method, and carrying out k-mer sequence base coordinate value conversion; then the mass center of the coordinate array is obtained through Kmeans clustering, and the k-mer sequence and the mass center of the k-mer sequence are returned to the CPU; and finally, the CPU counts the occurrence frequency of the k-mer and outputs a k-mer frequency distribution result. According to the method, the k-mer frequency counting speed is greatly improved, the occupation of computing resources is reduced, and the method is helpful for raw signal analysts to obtain an accurate analysis result at a faster speed in a shorter time.

The invention provides a memory processing method and system. The method comprises the following steps of acquiring the current memory information of a configuration program in real time; if the allocated memory is smaller than the current required memory, applying to a memory manager and obtaining an additional memory; and if the allocated memory is greater than the current required memory, releasing the redundant memory. In the scheme, if the allocated memory of the configuration program is not enough, the additional memory is applied to the memory manager, and if the allocated memory is greater than the currently required memory, the redundant memory is released. By adjusting the memory allocation of the configuration program flexibly, the situations that the memory of the configurationprogram is insufficient and the memory is not fully utilized are avoided, and the utilization rate of the memory is increased. Meanwhile, the memory manager determines whether to allocate the additional memory or not according to all allocated total memory and the memory upper limit, so that the situation that the configuration programs cannot be downloaded and compiled due to the fact that all allocated total memory exceeds the memory upper limit is avoided, and the user experience is improved.

The invention belongs to the technical field of simulation display, and discloses an oblique photography modeling data generation method applied to immersive display equipment, which comprises the following steps: according to a set path of oblique photography measurement, carrying out omnibearing visual content recording on flight simulation visual content in six surfaces including upper, lower,left, right, front and back at each node on the path; synthesizing the visual contents in the six directions into a corresponding node flight simulation visual panorama through a panorama texture mapping technology; and loading the flight simulation visual panorama to immersive display equipment. The invention aims to solve the problems of blockage and delay generated when oblique photography modeling data is viewed in the immersive display equipment.

The invention relates to a base station upgrading method and device, equipment and a storage medium. The method comprises the steps of receiving an upgrading request sent by a server, deleting a professional mirror image file, wherein the professional mirror image file is a file obtained by decompression when the base station is powered on and started; sending a downloading request by the server,wherein the downloading request is used for indicating to obtain an upgrading file required for upgrading the base station; and receiving an upgrading file returned by the server according to the downloading request, and upgrading the base station according to the upgrading file. According to the invention, the base station deletes the professional mirror image file before receiving the upgradingfile returned by the server, so that the memory space corresponding to the professional mirror image file is released, the situation of insufficient memory space is avoided, and the problem of upgrading failure caused by insufficient memory space is further avoided.