147results about How to "Reduce the number of reads" patented technology

A method and system for generating a decision-tree classifier from a training set of records, independent of the system memory size. The method includes the steps of: generating an attribute list for each attribute of the records, sorting the attribute lists for numeric attributes, and generating a decision tree by repeatedly partitioning the records using the attribute lists. For each node, split points are evaluated to determine the best split test for partitioning the records at the node. Preferably, a gini index and class histograms are used in determining the best splits. The gini index indicates how well a split point separates the records while the class histograms reflect the class distribution of the records at the node. Also, a hash table is built as the attribute list of the split attribute is divided among the child nodes, which is then used for splitting the remaining attribute lists of the node. The method reduces I/O read time by combining the read for partitioning the records at a node with the read required for determining the best split test for the child nodes. Further, it requires writes of the records only at one out of n levels of the decision tree where n>/=2. Finally, a novel data layout on disk minimizes disk seek time. The I/O optimizations work in a general environment for hierarchical data partitioning. They also work in a multi-processor environment. After the generation of the decision tree, any prior art pruning methods may be used for pruning the tree.

The invention discloses a machine translation method and device. The method comprises the steps that a sentence to be translated is received; at least one phrase fragment in the sentence to be translated is replaced by a preset character string, and a template which is matched with the sentence after replacing is searched from a template library after replacing operation every time; for the sentence matched with the template, a constant translation result of the part, corresponding to the constant of the template, of the matched sentence is acquired according to the template, a variable translation result of the part, corresponding to the variable of the template, of the matched sentence is acquired by decoding, and the constant translation result and the variable translation result are spliced; for the sentence which is not matched with the template in the template library, the translation result is acquired by decoding. According to the machine translation method and device, calculation can be reduced, and translation quality is improved.

The invention relates to a rapid synchronization method of logical volumes based on data difference, which belongs to the technical field of computer data storage. The method comprises steps of finding out used data blocks on each logical volumes according to file system information of source volumes and destination volumes, comparing corresponding data block pairs on the source logical volumes and the destination logical volumes, then computing two eigenvalues of a data block pair, synchronizing the data block pair if the two eigenvalues are varied, or not synchronizing if not. The invention only compares the used data blocks in case of data synchronization, thereby obviously reducing the range of data comparison, and simultaneously only the changed data blocks are synchronized during synchronization, thereby effectively reducing the quantity of synchronized data.

The invention discloses a data recovering method and device, which can solve the problem that the performance of a database is low as a data file is frequently read from a magnetic disk in a database recovering process, and the problem that a database recovering speed is slow as continuous logs correspond to discontinuous data pages. The method comprises the following steps: obtaining at least one log from a recovering starting point in a log file; generating a virtual data page according to each Redo log in the at least one log, wherein the virtual data page is used for representing a data page obtained by redoing according to the Redo logs; and combining each virtual data page with a corresponding original data page to be recovered in the magnetic disk according to a storage sequence of the original data page to be recovered to obtain the recovered original data page. The data recovering method and device disclosed by the invention are mainly applied to the database recovering process.

The invention discloses an acquisition method of parameters of a technique process, which is used for acquiring the change condition of the parameters of the technique process with the time; the acquisition method comprises the following steps: when a lower computer detects that the parameters of the technique process pre-subscibed by an upper computer are changed, the upper machine is notified; and the upper computer acquires the technique process parameters and corresponding time information according to the notification, and stores the technique process parameters and corresponding time information in a database. The invention can reduce the acquisition of garbage data, guarantee the validity of the acquired data, and can effectively save system resources.

The invention discloses a data operating method, which operates data stored in each memory heap in the following steps: receiving a hybrid address, wherein the hybrid address comprises a logic address and a first physical address, and the first physical address is used as the low order byte of the actual physical address of the data operated in the current heap; according to the mapping relation between the stored logic address and physical address, acquiring the physical address corresponding to the logic address as a second physical address, and determining the high order byte of the actual physical address according to the second physical address; according to the low order byte of the actual physical address and the high order byte of the actual physical address, generating the actual physical address; and linking to the corresponding memory space by using the actual physical address, and operating the data stored in the memory space. The invention also discloses a data operating device, an addressing method and an addressing device. The invention can solve the problem of high time consumption in the prior art since data searching and operation need to read the EEPROM for times.

A non-volatile memory includes a plurality of cells each individually capable of storing multiple bits of data including bits of multiple physical pages including at least a first page and a second page. A controller of the non-volatile memory determines a first calibration interval for a first read voltage threshold defining a bit value in the first page and a different second calibration interval for a second read voltage threshold defining a bit value in the second page. The second calibration interval has a shorter duration than the first calibration interval. The controller calibrates the first and second read voltage thresholds for the plurality of memory cells in the non-volatile memory based on the determined first and second calibration intervals.

The present invention relates to a method for a page-level address mapping based on flash memory and a system thereof. A method for a page-level address mapping based on a flash memory according to an embodiment of the present invention includes the steps of: receiving a write operation from a file system; generating condensed mapping information using a size of data information of the write operation and a start logical address of sequentially allocated logical addresses of the write operation; and storing the condensed mapping information as a first mapping table in a memory of a flash translation.

A non-volatile memory includes a plurality of physical pages each assigned to one of a plurality of page groups. A controller of the non-volatile memory performs a first calibration read of a sample physical page of a page group of the non-volatile memory. The controller determines if an error metric observed for the first calibration read of the sample physical page satisfies a calibration threshold. The controller calibrates read voltage thresholds of the page group utilizing a first calibration technique based on a determination that the error metric satisfies the calibration threshold and calibrates read voltage thresholds of the page group utilizing a different second calibration technique based on a determination that the error metric does not satisfy the calibration threshold.

A system and method for efficiently accessing operands in a datapath. An apparatus includes a data operand register file and an execution pipeline with multiple stages. In addition, the apparatus includes a result bypass cache configured to store data results conveyed by at least the final stage of the execution pipeline stage. Control logic is included which is configured to determine whether source operands for an instruction entering the pipeline are available in the last stage of the pipeline or in the result bypass cache. If the source operands are available in the last stage of the pipeline or the result bypass cache, they may be obtained from one of those locations rather than reading from the register file. If the source operands are not available from the last stage or the result bypass cache, then they may be obtained from the data operand register file.

A method of maintaining and providing information relating to file deduplication. A first portion of a first file and a second portion of a second file that contain a first content are identified. A first header associated with the first portion is created. The first header identifies the first portion and the second portion containing the first content. The first header is appended to a storage location of the first content of the first portion to form a first data structure for the first file. The first data structure is stored. The first data structure is provided to an application requesting the first file so that duplicate data processing can be avoided by the application. The first data structure is updated when the first file or the second file are altered. A similar process may occur to generate a data structure for the second file.

The invention discloses a data backup method and system based on a submodule model, and belongs to the technical field of computer storage. The method comprises the steps that partitioning is conducted on data flow, and different data segments are formed; a fingerprint of each data block of the corresponding data segment is calculated, fingerprint indexing is inquired, and container information quoted by repeating data blocks is obtained; a submodule function maximization model is established for selecting containers containing more quotable and non-redundant data blocks; the data blocks quoting the containers are subjected to duplication eliminating, and fragment data blocks quoting other containers are rewritten. The invention further provides the data backup system based on the submodule model. According to the technical scheme, more duplicating data blocks can be removed, consumption of redundant and non-quoting data blocks on bandwidth recovery is reduced, and the recovery performance is improved while the high duplication eliminating rate is guaranteed.

The invention provides a hardware circular processing method and system of a processor. Circular buffering is increased in an original fetch unit; first N commands in a corresponding circular body are directly output to a subsequent decoding unit through the circular buffering, and a condition that an extra waiting period caused by the fact that the reading of data in a program memory is delayed during each skipping to a first command of the circular body from a final command of the circular body in a circular processing process is eliminated, so that zero-delaying skipping of hardware circulation is realized. The method provided by the invention has a simple design, and only one hardware circular buffering and a corresponding selection module need to be increased in an original system to realize the zero-delaying skipping of the hardware circulation. Furthermore, the access of the fetch unit on the program memory can also be reduced by utilizing the method so that the power consumption of the processor is reduced.

The invention belongs to the technical field of high-definition digital video compression encoding and decoding and particularly relates to a filtering method of a block-removing filter suitable for the HEVC standard. Each image is set to comprise a luminance component Y which corresponds to two chrominance components Cb and Cr. Processing is performed in a block-removing filter module based on one quarter-LCU. Pixel blocks of the quarter-LCU needing to be processed are set to be provided with vertical edges and horizontal edges. The filtering sequence of the filtering method is that filtering is performed on the vertical edges and then on the horizontal edges; the vertical edges are sequentially filtered from left to right, and the pixel blocks on the two sides of each vertical edge are sequentially filtered from top to bottom; the horizontal edges are sequentially filtered from top to bottom, and the pixel blocks on the two sides of each horizontal edge are sequentially filtered from left to right. According to the filtering method of the block-removing filter suitable for the HEVC standard, the number of times of reading of an external storage device is effectively reduced, the processing time is shortened, the processing capability of a chip is improved, and therefore real-time encoding of a high-definition video is efficiently achieved.

One embodiment of the present invention sets forth a mechanism to schedule read data transmissions and write data transmissions to/from a cache to frame buffer logic on the L2 bus. When processing a read or a write command, a scheduling arbiter examines a bus schedule to determine that a read-read conflict, a read-write conflict or a write-read exists, and allocates an available memory space in a read buffer to store the read data causing the conflict until the read return data transmission can be scheduled. In the case of a write command, the scheduling arbiter then transmits a write request to a request buffer. When processing a write request, the request arbiter examines the request buffers to determine whether a write-write conflict. If so, then the request arbiter allocates a memory space in a request buffer to store the write request until the write data transmission can be scheduled.

The present invention discloses an image brightness interpolation method of H.264/AVC based on serially input and paralleling output video and the device. A full pixel point in each line serially inputs a Wiener filter to produce a medium value of a level half pixel point, at the same time, the full pixel point is selected by utilizing the motion vector, thereby, the register array is reduced to 6 is multiplied by 8, and the filtering algorithm of the 1/4 pixel is optimized. Compared with the original paralleling input and paralleling output interpolation device, a 6 tap FIR filter in at least 3 horizontal directions are saved. Compared with the original paralleling input and serially output interpolation device, the register array is reduced from 6 is multiplied by 9 to 6 is multiplied by 8, and the complexity of the production chain controller is decreased. Compared with the original serially input and serially output interpolation device, and the bandwidth and the computation period are saved.

The invention discloses a data caching method and system in a Web network management client based on a RIA and relates to the data caching field in network management systems. The data caching method comprises the steps of establishing an independent storage area; marking areas for arranging cache data in a physical memory as a resident memory data area and a non-resident memory data area; placing resident memory data and non-resident memory data in the resident memory data area and the non-resident memory data area respectively; releasing the resident memory data and non-resident memory data in the independent storage area regularly when the data are not used by users, and placing the resident memory data and non-resident memory data in the physical memory when the data are reused by users; and deleting all cache data after being used by users. According to the data caching method and system, the cache data can be placed in the physical memory and the independent storage area respectively according to requirements of users, less physical memory is occupied, the reading frequency of discs is reduced, and performances of the Web network management system are improved.

The invention discloses a fast FAT32 disk partition traversal and file searching method and relates to an FAT32 file system and a magnetic disk data search and recovery technology. The method of the invention comprises the steps as follows: raw-disk-reading disk sector data of a root directory and subdirectories to internal storage, and performing directory entry search to all levels of directories: gradually adding 32 bytes in the same directory to perform horizontal search, and performing downward search to the searched subdirectory entry at a start position of the subdirectory, and returning to a search position stored in a parent directory to search by performing upward search after finishing the search of the current subdirectory, and performing traversing to finish the search of each directory entry sub-tree of the root directory in turn, and performing file attribute search, classification and ordered storage to a file directory entry. The method of the invention improves traversal and search speed, and improves overall performance for subsequent file search, read or download. The method of the invention is especially suitable for FAT32 file system partition traversal and file search in an embedded system.

According to the invention, a compact and inexpensive with low power consumption memory and low access speed can be used for a panel controller and a deterioration compensation circuit of a display device. In a display device of a digital gray scale method, a plurality of pixels of a display panel are divided into first to n-th pixel regions (n is 2 or more) and a format converter portion of a panel controller converts the format of only video data corresponding to one of first to n-th pixel regions and writes the data to one of first and second video memories in each frame period. A display control portion reads out video data that is converted in format and corresponds to one of first to n-th pixel regions in which video data is written to the other of the first and second video memories in the preceding frame period, and transmits the data to the display panel.

The invention provides a data reading method and device, and the method comprises the steps: receiving a data reading instruction which comprises an identifier of a target physical block; obtaining aread voltage identification table corresponding to the identification of the target physical block from a preset cache space; wherein the reading voltage identification table comprises a plurality ofreading voltage identifications arranged according to a reading sequence; based on the reading sequence of the plurality of reading voltage identifiers, selecting one reading voltage identifier from the plurality of reading voltage identifiers, and determining a reading voltage parameter corresponding to the reading voltage identifier; reading data from the target physical block by using the voltage reading parameter; if the data reading fails, returning to execute a process of selecting one reading voltage identifier from the plurality of reading voltage identifiers based on a reading sequence of the plurality of reading voltage identifiers; and if the data is successfully read, outputting the successfully read data. According to the technical scheme, the reading efficiency can be improved, the reading performance is improved, and the reading speed is increased.

The embodiment of the invention discloses an acceleration method and device of data operation, a terminal and a readable storage medium. The method comprises the steps of reading a data group to be operated in each storage unit, buffering the read data group to a GPU and calling the GPU to carry out parallel operation on the buffered data group, wherein each storage unit stores a data group formedby data elements at the same position in each matrix data to be calculated. Compared with the prior art, in the operation process of the embodiment of the invention, the number of times of data reading in the operation process is reduced, at the same time, the operation speed of the data to be calculated can be effectively improved, in addition, since the operation process is transferred from a CPU to a GPU in the embodiment of the invention, the occupation rate of the CPU also can be reduced, and the normal operation of the terminal is ensured.

The present invention discloses the core of a graph database ID caching technology, the dynamic caching is applied to the graph databases with millions or more of data, and the ID caching technology caches IDs in memory, although the graph database will still be queried. However, compared with the original query process of traversing millions of data, the speed of query is much faster, and the amount of query is very small, so that the query speed is also greatly improved, and the data dynamic, the data changes in the database and the cache do not conflict, greatly reducing the probability ofdirty data. The device is divided into the following parts of a special cache structure based on the structural characteristics of graph database, different matching strategies for simple query and complex query, updating cache to achieve consistency when the database is changed, and replacement strategy when the number of caches is too large.