11results about How to "Reduce read" patented technology

The application relates to electronic data processing of engineering cost, and discloses a method for sharing and canceling nodes in the scope of engineering cost equivalence request. A processor parses a candidate variable into a normalized scope semantic tuple containing a project version, a hierarchical node range, a source field path, filtering, aggregation, transformation, a unit of measurement and precision, to request a summary to locate a sharing bucket, and compare complete tuples in the bucket, and only merge requests that are completely equivalent in semantics, have certainty and have no external side effects. A shared node maintains a node generation number and an active candidate reference set; when a candidate is canceled, only the reference is removed; only when the reference is empty, the node is to be executed and the generation numbers match, the node is canceled. A node executes a snapshot of an immutable variable once, writes a barrier to prevent distribution to canceled candidates, and invalidates affected nodes according to a change set, so that range missharing, cross-candidate cancellation interference and old snapshot reuse are avoided.

The embodiment of the present application discloses an information processing method, device and computer-readable storage medium. The embodiment of the present application obtains a first file to be processed in a first preset format; parses the first file to be processed, and converts the first file to be processed into a second file to be processed in a second preset format; reads the number of feature dimensions and corresponding feature information of each training sample information in the second file to be processed, and stores each training sample information in a first preset memory device; reads each training sample information in the first preset memory device into a graphics processor for processing. In this way, by adding the number of feature dimensions to the training sample information, the corresponding feature information can be quickly determined through the feature dimension data, and only two reading operations are required to store each training sample information in the first preset memory device, and then transfer it to the graphics processor for processing, reducing information calculation and reading, and greatly improving the efficiency of information processing.

The application relates to a fingerprint library reconstruction method and device, equipment, a readable storage medium and a program product, and relates to the technical field of data recovery backup. The method comprises the following steps: reading the head part of a data file; acquiring the head part offset of each data block in the data file according to a bit set carried in the head part of the data file; reading the head part of each data block in parallel according to the head part offset of each data block, and obtaining the data block fingerprint and metadata carried in the head part of each data block; and inserting the data block fingerprint and metadata corresponding to each data block into a fingerprint library. The method can reduce the reading of invalid data and read the head part of each data block in parallel, so as to improve the fingerprint library reconstruction efficiency.

This invention relates to the field of multimedia video processing and playback control technology, specifically a video segment-level content management and control method and system, comprising: a time-series discrete mapping step: obtaining a server-side rule set and mapping it to a time-series state index structure with video durations divided into discrete-time graticles; an asynchronous look-ahead detection step: calculating the look-ahead time based on the current playback time using an independent asynchronous thread and reading the corresponding graticle data from the bitmap; an instruction generation and caching step: generating instruction objects based on graticle control identifiers using preprocessing logic and pushing them sequentially into an instruction circular queue; and a frame-level synchronous application step: checking the queue during main playback thread rendering, and applying the instruction if the current time matches the instruction trigger time. This invention reduces playback CPU usage, eliminates micro-stuttering caused by main thread blocking, and ensures accurate triggering of control logic at the frame level.

The present invention relates to the intersection of deep learning and high-performance computing, and in particular to a three-dimensional convolution parallel computing method, device and equipment based on a multi-core processor. The method includes: according to the available space size of the memory of the slave core array of the multi-core processor, the input feature map is divided into multiple sub-input feature maps in the direction parallel to its third dimension; the convolution kernel is obtained, and the sub-input feature map plane matrix equal to the number of convolution kernel plane matrices contained in the convolution kernel is taken from the sub-input feature map, and the plane matrix is ​​sent to the slave core array together with the convolution kernel to perform three-dimensional matrix convolution calculation; the three-dimensional matrix convolution calculation is divided into a number of computing units of equal number corresponding to the number of available slave cores in the slave core array, and each slave core is responsible for the calculation of a computing unit. During the calculation process of the computing unit, data exchange is performed in the memory of the slave core array, and the slave cores are calculated in parallel. After all slave cores have completed the calculation, the results are written back to the main memory.

The present invention relates to the field of data management technology, and in particular to a platform data processing management method and system based on big data. A platform data processing management method based on big data, comprising the following steps: S1: using NLP to obtain a unique semantic label for each data item field in a data warehouse, and adjusting the usage quantity of each data item in all downstream systems according to the unique semantic label; S2: based on the adjusted usage quantity in the downstream system, using a comprehensive weight formula to obtain the comprehensive weight of each data item, using a weighted time calculation formula according to the comprehensive weight, obtaining a weighted average waiting time, and obtaining a cache adjustment formula according to the weighted average waiting time. The present invention adapts to the complex needs of different scenarios by adjusting the cache size and data management rules in real time according to the dynamic needs of the downstream system, while providing good support for future system expansion.

The invention relates to a block chain data storage method and device and computer equipment, and the method comprises the steps: determining a storage position of historical data under the condition of modifying the historical data stored in a current data storage structure, and storing the current data obtained by modifying the historical data into the current data storage structure; if the storage position is a temporary storage area of the current data storage structure, storing historical data in the temporary storage area into a historical data storage structure; if the storage position is a permanent storage area of the current data storage structure, determining a data identifier corresponding to the historical data; and storing the historical data extracted from the permanent storage area based on the data identifier into a historical data storage structure. By adopting the method, the read-write efficiency of the block chain data can be improved.

The embodiment of the application discloses a kind of parameter calculation methods under asymmetric short-circuit fault condition. Among them, the method comprises: under symmetric short-circuit fault condition, obtaining the expected current ac component oscillogram of symmetric short-circuit;By expected current ac component oscillogram, the current effective value and the phase angle difference of current voltage of multiple time are measured;Under asymmetric short-circuit fault condition, the voltage phase angle of multiple time is measured;According to current effective value, the phase angle difference of current voltage and voltage phase angle, attenuation coefficient is calculated;According to attenuation coefficient, direct current time constant is calculated;According to current effective value, the phase angle difference of current voltage, voltage phase angle and the direct current time constant determine asymmetric short-circuit current expression, so current expression under asymmetric short-circuit fault condition can be simply and accurately obtained, and the operation amount and error rate are reduced.

The invention discloses an implementation method for filtering and pushing down a storage engine of a time sequence database, and relates to the technical field of time sequence databases, which comprises the following steps of: designing the storage engine of the time sequence database into five levels including a time sequence table data main directory, Entity Groups, SubEntity Groups, time partitions and data subfolders, each Entity Group comprises Tag table data, Metric data and a Data data table, and each Entity Group comprises the Tag table data, the Metric data and the Data data table; designing data structures, filtering push-down elements and filtering push-down processes of the Tag table and the Data table; based on the above design, filtering push-down is implemented in each level of a time sequence database storage engine, a query execution plan is optimized, and query is executed to obtain a summarized result. According to the method, effective filtering and pushing-down can be realized, and the data reading and transmission quantity is reduced, so that the query performance is improved.

The application provides a data processing device, a data processing method and related devices, comprising a neural network processor, first, according to the channel number n of the first data, the first data is divided into n first sub-data, the first data represents data to be stored, n is a positive integer less than or equal to M / 2; each first sub-data is written into a circular cache space composed of INT(M / n) ring cache space groups, and any one circular cache space is used for storing any one first sub-data. The area of the cache control circuit in the neural network processor can be reduced by a specific storage space architecture, and the power consumption of data writing and reading is reduced.

The invention discloses a spark-based spatial incremental data processing method, and belongs to the field of spatial data processing. Comprising the following steps: establishing an equalization space intelligent partitioning algorithm based on a spark parallel analysis principle; constructing a distributed storage format for the spatial data; during data reading, filtering the original partition data file by using the spatial index file; a spatial analysis mode is constructed based on an intelligent partitioning algorithm and a distributed storage format, and is used for executing the following steps: reading newly-added spatial data, and obtaining the data volume and the spatial range of the newly-added spatial data; carrying out space partitioning operation on the newly-added space data by utilizing an equalized space intelligent partitioning algorithm, and reducing each partition according to a data range; establishing a spatial secondary index in each partition; and writing the processed partition data into a database in parallel. The working efficiency of space incremental data processing can be improved, the labor cost is reduced, it is ensured that the processing process adopts a unified standard, human errors are reduced, and the accuracy is improved.