51results about How to "Speed up writing" patented technology

The invention provides a data indexing method for streaming time sequence data, which comprises the following steps of: 1, establishing an integral data indexing structure which is formed by adding aB * tree pointing to a sibling pointer to non-root and non-leaf nodes of a B + tree; Step 2, carrying out batch resource application and index optimization; And step 3, carrying out trigger type treestructure segmentation and combination.

The invention belongs to the field of designing of integrated circuits and provides a read-write bandwidth improvement memory apparatus. Particularly, a broadband data read-write port, besides a normal read-write port, is arranged on a memory in four modes, namely a wide bandwidth read mode, a normal bandwidth read mode, a wide bandwidth write mode and a normal bandwidth write mode. Different columns of memory units, which share word lines, of a memory array can pass through a broadband data write-in port and perform write-in operation simultaneously; and after values in the memory units corresponding to an input address is read out during read-out, values inside all other columns of memory units which share the word lines can be read out through a broadband data read-out port. Write-after-read operation can be performed on different memory rows of the memory array in the same clock period. A plurality of memory arrays can share a group of read-out apparatuses, and read-write operation in different arrays can be performed simultaneously.

The invention provides a solid state disk writing method. The solid state disk writing method comprises the following steps: starting a task of writing to-be-written data into a cache region; judgingwhether the residual space of the cache region is smaller than a first threshold value or not; if the residual space of the cache region is smaller than the first threshold value, starting a garbage collection mechanism, and reserving the hot data in the cache region in the cache region; and moving the cold data in the cache region to a flash memory region. In addition, the invention further provides a solid state disk writing device, a solid state disk and a computer readable storage medium. According to the technical scheme provided by the invention, high-performance writing of the solid state disk can be kept for a long time.

The invention provides a data writing method, a memory controller and a memory storage device. The method is used for controlling a rewritable nonvolatile memory module, which comprises two storage memory cells. The method comprises the following steps of: configuring a plurality of logical addresses to be mapped to at least part of entity erasing units in the two memory cells; receiving a writing command from a host system, which indicates that the data are written in one logical address; writing the data in one entity erasing unit in the two memory cells; judging which memory cell does the entity erasing unit belong to; if the entity erasing unit belongs to one memory cell, erasing one entity erasing unit in the other memory cell when the data are written in. therefore, the speed that the data are written in the memory storage device by the host system can be increased.

The invention discloses a data updating method, wherein, bumping blocks in preset quantity are arranged. The data updating method includes the following steps: updating data corresponding to any one block data is written to the bumping block corresponding to the block data, and the correlation between the position of the updating data in the bumping block and the position of the data to be updated in the corresponding block data; when any one bumping block is overwritten, the updating data in the bumping block is entirely written to the corresponding block data according to the correlation. The invention also discloses a data updating device at the same time. The data updating device comprises a bumping block definition module, a bumping block operation module and an updating module. The data updating method and the data updating device of the invention avoids the repetitiveness operation caused by rewriting the entire data after a corresponding web page in the block data is updated when updating the date of the web page in the block data, and increases the efficiency of data writing and updating.

The embodiment of the invention provides a method, a device and a system for reading and writing a magnetic disk file of a distributed file system, wherein the system comprises a metadata processing module cluster, and the metadata processing module cluster comprises at least two metadata processing modules; a write-in method comprises that the metadata processing module cluster receives a file write-in request from a client side, distributes all the magnetic disk files to be written-in to each metadata processing module, and each metadata processing module respectively sends the distributed magnetic disk file to be written-in to a magnetic disk; and a read-out method comprises that the metadata processing module cluster receives a file read-out request from the client side, distributes all the magnetic disk files to be read-out to each metadata processing module, and each metadata processing module respectively sends the distributed magnetic disk file to be read-out to the client side, so that data processing stress is balanced to each metadata processing module, the data processing stress of each metadata processing module is reduced, and speed and reliability for the client side in the distributed file system to write-in and / or read-out the magnetic disk file are increased.

The embodiment of the invention discloses a cache processing method and device. The method can comprise the following steps that: Cache receives a main memory address sent from a CPU (Central Processing Unit), wherein the main memory address comprises a Tag and an Index 1; and a Set1 corresponding to the Index1 is searched, and a Set2 corresponding to an Index2 is searched; and a matched Cache Line is searched in the Cache Line of the Set1 and the Set2. When the method is adopted, power consumption expenditures during Cache deficiency can be lowered, and Cache performance is improved.

The invention provides a real-time dynamic data management system based on HBase. The real-time dynamic data management system comprises data reading optimization, a composite index mechanism and parallel query optimization. According to the invention, a persistent data storage mechanism based on HDFS and HBase is designed, so that the problem of mass storage of multi-structure data is solved; a composite index storage model is designed based on an HBase table structure, and realizing mapping management of index cache addresses and index table design of cache data storage; queue division is performed on the data based on a multi-source buffer area structure to realize parallel storage, and the concurrent writing pressure of mass data is relieved by combining strategies such as a consistency hash algorithm, a multi-thread technology, a line key optimization design and the like; And a multi-query parallel mechanism is designed based on an HBase query real execution module and a multi-line technology, so that the query speed is increased. According to the real-time dynamic data management system based on the HBase, the data writing and reading performance is effectively improved, theproblem of mass storage of multi-structured data is solved, and the high-dynamic and real-time environmental requirements are met.

The invention provides a method for optical storage medium writing. The method comprises: transferring raw data from a host to a primary memory; reading the primary memory to obtain read-out data; carrying out a series of encoding steps simultaneously to generate formatted data according to the read-out data; and directly writing the formatted data to the optical storage medium without accessing the primary memory again; wherein the raw data must be converted in advance to the formatted data through the encoding steps before the formatted data is written to the optical storage medium. Some encoding steps are performed in synchronism based on data obtained by once memory accessing, thus the delay time of memory accessing is reduced to quicken the write of the data, therefore the system effectiveness is increased.

The embodiment of the invention provides a driving method of a display panel, a driving chip and a display device in the technical field of display, and is used for improving the ghost shadow phenomenon and improving display performance. The driving method comprises the steps of monitoring a static pattern displayed by the display panel in a first driving mode and when the display brightness value of the area where a static pattern is located and display brightness value of an area where the surrounding pattern is located meet preset conditions, defining the area where the static pattern is located as a target area; driving a pixel circuit in the target area in a second driving mode, in a second drive mode, enabling the grid line to scan the pixel circuit at a second frequency; wherein inthe first driving mode, the second frequency is greater than the first frequency of scanning the pixel circuit by the grid line, wherein in the second driving mode, the driving period of the pixel circuit comprises a signal switching period, an initialization period, a data signal writing period and a light emitting period, and in the signal switching period, the data line writes different data signals into the pixel circuit.

The invention discloses a data read-write buffer method and a buffer device; the method includes the following steps of writing reserved section data in storage medium to a buffer zone and writing updated data to the buffer zone by a system to form whole buffer zone data for being written to the storage medium. By adopting the method and device of the invention, during the data buffer reading / writing, the non-updated data in the storage medium is written / read, namely when writing the data, the data is read when needed, so as to avoid most of unnecessary reading / writing switching; under the condition of not increasing the memory usage, the data writing or burning speed can be greatly improved.

A trap type memory element in which inflow of charges into a charge storage region from the outside, external diffusion of stored charges and outflow caused by a defect are suppressed. A gate conductor (6) is formed on a silicon substrate (1) through a multilayer insulating film consisting of a first gate insulating film (3), a charge storage layer (4), and a second gate insulating film (5). The multilayer insulating film (3-5) is projecting to the outside of the gate conductor (6) and extending below the outer end of a sidewall (8). The charge storage layer (4) has a high trap surface density region (4a) directly under the gate conductor, and a low trap surface density region (4b) on the outside of the gate conductor.

The invention discloses a log writing method and device, a system and a computer storage medium. The method comprises the steps that an end point page number which is written into a log is read from avariable area of a log disk; the log is written into a data area page corresponding to the end point page number to form a log page; the writing position corresponding to the log page and a data section number which the log page belongs to are acquired and stored into a bitmap area of the log disk; in the variable area of the log disk, one is added to the end point page number which is written into a log, and whether or not the log is written to a tail disk of the log disk is judged; if yes, an end point which is written into the log is set to a disk head.

The invention discloses a data transmission method and device based on cloud storage, computer equipment and a storage medium. The method comprises the following steps: receiving full data uploaded bya Hive database, and storing the full data; obtaining the number of pre-partitions in an HBase database; partitioning the total data according to the number of the pre-partitions and the row keys ofthe data in the total data to obtain corresponding partition data; performing ascending sorting on each part of partition data according to the column and row keys in sequence to obtain correspondingsorted partition data; and sending the sorted partition data to a partition server corresponding to the Hbase database for storage. According to the method, the sorting process is completed at the cloud before the total data is written into the Hbase database, so that the efficiency of writing the data into the Hbase database is improved.

The embodiment of the invention discloses method and terminal for writing image data into a file. The method comprises the steps of creating a target interface while receiving an image data processing instruction; transferring a target function through the target interface; processing the original image to obtain target image data; dividing the target image data into a plurality of logic blocks; writing the target image data into a target file in match with a target identifier by means of logic blocks. With the adoption of the method and terminal, the efficiency of writing the image data into the target file can be increased.

The invention discloses a memory cell and a content addresable memory body. A memory cell includes a set of storage switch units, a set of memory units, a set of comparison switch units and a discharge switch unit. The storage switch units are turned on by a turn-on signal transmitted by a word line. The memory units receive and store write data transmitted by a bit line or a source line when the storage switch units are on under a write mode. The comparison switch units are turned on by comparison data transmitted by comparison lines under a search mode. The discharge switch unit is turned on by a detection voltage under the search mode when the comparison data transmitted by the comparison lines is different from the write data stored in the memory units, so that the reference signal transmitted to the comparator is redirected to a reference voltage. A content addressable memory using the memory cell is also provided.

The invention provides an information management method and system of a distributed system, a storage medium and equipment. The method comprises the steps of presetting a user system and a cache system, wherein the user system conducts information synchronization by penetrating through all nodes in the distributed system, and the cache system is used for providing data cache service for all the nodes in the distributed system; after the user connection gateway node obtains authentication, receiving user identification information, and taking the user identification information as an ID identification of a user system; enabling the user system to perform hash calculation and direct remainder calculation according to the ID identifier to obtain the position of the user node; deciding whetherto route information to a service node according to a load value provided by a user system operated by the user node; and routing the information to the service node to process the logic relationshipbetween the service and the data, so that the consistent transaction processing is completed, and the cross-process access of the information routing data in the distributed system is more convenient.

A method for the performance of a function by a microcircuit, includes:at least one step of determining (205) whether an anomaly is detected or whether the operation of the microcircuit is normal;when it is determined that an anomaly is detected, a step of performing (210) a protection function;when it is determined that the operation of the microcircuit is normal, a step of performing (215) a decoy function simulating the protection function by being perceptible, from the outside of the microcircuit, in a manner more or less identical to the protection function;the method being characterized in that it includes an interruption (250) of the performance of the decoy function by a timer.

The invention provides a data management method and system based on a pandas library and an InfluxDB database, a terminal and a medium, adopts a technology of combining the pandas library and the InfluxDB database, and aims to perform data processing and analysis on mass data of the InfluxDB database by utilizing the pandas library and have good application to mass data refreshed at high frequency. Due to the fact that the InfluxDB database abandons the data updating and deleting functions, data writing and querying are intensified in a centralized mode, the InfluxDB database quite conforms tothe database interface characteristics of the Panda database, the data analyzing and processing functions of the Panda database can be brought into full play, and the time for connecting, reading andwriting with a traditional database in the past is saved.

The invention discloses a high-performance data packet caching method and a high-performance data packet caching device based on hierarchical storage. The high-performance data packet caching method comprises the following steps that: a data receiving unit of a network processor receives data packet caching data in a network link; an off-chip memory stores cached data of the data packet; when thedata packet is stored, the space needs to be reserved for the head of the data packet instead of starting to store the data packet from the initial address of the space allocated to the off-chip memory by the network processor system; when the network processor modifies the header of the data packet, the header of the data packet is stored in the reserved space, and the whole data packet is storedat the same time; and the data sending unit transmits the stored data packet to a sending network link. According to the invention, in a zero-copy data buffer structure, a large amount of data does not need to be copied and copied, and only the head of the data packet is modified and forwarded before the payload of the data packet, so that the high-speed data processing capability is ensured.

The invention discloses an object storage additional writing implementation method based on logs. The method comprises the steps that a client sends an additional writing request; an object storage Web end processing layer checks the legality of the request parameters; an object storage processing layer performs object legality check, and judges whether a target object exists or not if the check is passed; if the target object exists, the metadata of the target object is analyzed, and to-be-added data is written in a corresponding mode according to the type of the target object; if the targetobject does not exist, the strip information of the target object is initialized, and to-be-added data is written by taking a new strip as a starting position; the metadata is written into a head object of the target object; after metadata updating is completed, the object storage system returns response information to the client. On the premise that the original object file does not need to be uploaded repeatedly, the data additional synthesis is achieved, and meanwhile, logs are recorded during data uploading to avoid data loss caused by accidental interruption.

The invention discloses a live broadcast data scheduling method, a live broadcast data access method, a live broadcast data scheduling device, live broadcast data access equipment, a medium and a product, and the scheduling method comprises the steps: calling a regression prediction model, and obtaining an anchor identifier corresponding to an online anchor client in a current time period in a live broadcast platform predicted by the model; querying live broadcast data corresponding to the anchor identifier stored in a hierarchical database, determining the live broadcast data as hotspot data, floating the hotspot data to a storage medium with the highest performance level of the hierarchical database for storage, and dispersedly storing the data of the hierarchical database in a plurality of storage mediums with different performance levels; and responding to a data collection event pushed by the live broadcast service server, querying live broadcast data corresponding to the event in each storage medium according to the performance level from high to low, and pushing the live broadcast data. According to the method, the source live broadcast data is stored in a layered manner, and the predicted hotspot data is floating and stored in the high-performance storage medium, so that the transmission efficiency of the live broadcast data is effectively improved.

The invention discloses a massive small file storage performance optimization method based on time sequence prediction, and belongs to the field of information storage. The massive small file storageperformance optimization method comprises the following steps: collecting historical file access records with time information to obtain a data set; after the data set is preprocessed into discrete time series data, using a time window to rolling on the data set to generate a training data set, wherein the training data at any t moment takes the data at the t-n-t moment as input data, and takes the data at the t + 1 moment as label data; establishing a time sequence prediction model based on a recurrent neural network, and performing training, verification and testing in sequence by utilizinga training set, a verification set and a test set obtained by dividing the training data set, thereby obtaining a target model; predicting the change trend of the file size by utilizing the target model so as to identify a large file and a small file in the file size; and directly storing the large files, and aggregating and storing the small files based on a time sequence. According to the massive small file storage performance optimization method, the storage performance of massive small files in the distributed storage system can be optimized.

The invention discloses a magnetic composite film for magnetic storage as well as a preparation method of a magnetic composite film. According to the magnetic composite film, a non-magnetic silicon wafer serves as a substrate, and a soft magnetic layer, an alloy hard magnetic layer and a barium ferrite hard magnetic layer are sequentially arranged on the surface of the substrate; and the alloy hard magnetic layer is a [FePt / FeCo]N composite film layer with vertical magnetic anisotropy, wherein N is the number of periods and is a natural number between 2 and 5, the molar ratio of Fe to Pt in the FePt alloy layer is 1:(0.5-1.1), and the molar ratio of Fe to Co in the FeCo alloy layer is 1:(1-1.5). The BaFe12O19 hard magnetic layer serves as a protective film of the [FePt / FeCo]N alloy magnetic layer, so that high oxidation resistance, heat stability and wear resistance are achieved; and meanwhile, the BaFe12O19 hard magnetic layer and the [FePt / FeCo]N hard magnetic layer are both verticalmagnetic recording materials and are subjected to interlayer coupling to produce interface vertical magnetic anisotropy to form a magnetic tunnel junction, thereby further increasing storage densityand avoiding demagnetization phenomenon caused by the traditional metal protective film.

The invention provides a manufacturing method of a semi-floating gate memory. The manufacturing method comprises the following steps: providing a substrate of a first doping type; generating a first semiconductor and a second semiconductor on the substrate, wherein the first semiconductor and the second semiconductor are opposite to the substrate in conductive type; and generating a tunneling layer and a barrier layer at the same time, wherein the tunneling layer is arranged on the substrate and adjacent to the first semiconductor in parallel, the second semiconductor covers the tunneling layer and the first semiconductor, the barrier layer covers the second semiconductor, and the tunneling layer and the barrier layer are generated at the same time so that the process steps are reduced, the process complexity is reduced, and the production efficiency is greatly improved. Besides, when a diode structure formed by the first semiconductor and the substrate is conducted, writing of data is accelerated, the rapid storage function of the data is achieved, and due to the diode structure and the tunneling layer, the storage time of the data is greatly prolonged. In addition, the invention also discloses the semi-floating gate memory.

An improved method and an improved device for the contact-free transmission of data from and / or to a plurality of data or information carriers (9) are characterized by the following features: After the exchange of information between a read and / or write device (1) and a data or information carrier (9), said carrier is switched to a standby impedance mode (Z4) in order to reduce undesired interaction with data or information carriers (9) adjoining the antenna field, said mode differing from an initial impedance mode (Z1) and communication impedance modes (Z2, Z3) during the emission of an information signal from the data or information carrier (9).

A driving circuit for a display panel comprises a single-inductor double-voltage generator, a first multiplexer, a multi-voltage generator and a second multiplexer, wherein the single-inductor double-voltage generator generates and outputs a positive voltage and a negative voltage according to an input voltage and an inductor; the first multiplexer sequentially drives a plurality of scanning lines of the display panel according to the positive voltage, the negative voltage, input data and a first clock pulse generated by a clock pulse generator; the multi-voltage voltage generator generates a plurality of groups of driving voltages according to a second clock pulse, the positive voltage and the negative voltage which are generated by the clock pulse generator; and the second multiplexer generates a plurality of data voltages to a plurality of data lines of the display panel according to the plurality of groups of driving voltages and the input data.