1166 results about "Reading (computer)" patented technology

The invention discloses a request allocation method and system based on a heterogeneous computing system, equipment and a medium, relates to the technical field of computers, and determines reading time required for using key value cache when a plurality of heterogeneous computing power equipment accesses a memory under the condition of preferentially considering the use condition of a key value cache mechanism. Requests are reasonably distributed by considering the performance difference of the computing power information of the heterogeneous computing power equipment, the memory information corresponding to the memory extension and the characteristics of the reasoning task information. Furthermore, the reasonability of request allocation is improved according to a comparison relationship between the preset allocation request quantity and the concurrent request quantity and different strategies for judging whether allocation conditions are met or not. Therefore, the technical problem that more computing power resources are wasted due to the fact that the allocation balance of the reasoning request is poor when a separated memory or a local memory is called can be solved, and the technical effects that the reasoning task information is reasonably allocated to the heterogeneous computing power equipment, so that the allocation balance is improved, the computing power resources are saved, and the cost is reduced are achieved.

The invention provides a server voltage monitoring system and method, electronic equipment and a storage medium, and relates to the technical field of computers, the system comprises a complex programmable logic device CPLD, and an initial timing module, a voltage acquisition module, an abnormal data storage module, a CPU processor and a baseboard management controller BMC module which are respectively in communication connection with the CPLD. The initial timing module activates the monitoring function of the CPLD through a startup signal; the voltage acquisition module acquires voltage data under the control of the CPLD so as to read the voltage data of the server mainboard; the CPLD obtains voltage data, compares the voltage data with a reference voltage range of the abnormal data storage module, triggers an interrupt signal after identifying abnormal fluctuation voltage, and feeds back the interrupt signal to the CPU and the BMC, so that the CPU and the BMC quickly output an alarm. The system covers full-cycle monitoring from startup to BMC startup, supports efficient acquisition, dynamic detection and low-delay response of multi-path voltage, relieves the voltage pressure of a BMC monitoring mainboard, and provides reliable guarantee for power supply safety of a server.

The invention discloses a data reading method and device, computer equipment and a storage medium, and relates to the technical field of flash memories, and the data reading method comprises the following steps: before reading data, firstly obtaining a performance index value of a current flash memory page at a current moment, and firstly carrying out detection operation on the current flash memory page to obtain an error code index value of the current flash memory page; and determining a target risk level according to the performance index value and the error code index value, then obtaining a target pre-charging duration corresponding to the target risk level, and performing a reading operation on the current flash memory page based on the target pre-charging duration to read data stored in the current flash memory page. As the pre-charging duration can influence the accuracy of data reading, the accuracy of reading the data for the first time can be improved by reading the data by adopting the target pre-charging duration corresponding to the target risk level before data reading, so that the number of times of re-reading is reduced, and the waste of resources is reduced.

The invention relates to the technical field of computers, and discloses a data management method and device, computer equipment and a storage medium, and the method comprises the following steps: obtaining target data of a disk array; grouping the target data to obtain corresponding grouped data under each preset category; writing the grouped data into the corresponding storage positions; after the storage fault corresponding to the disk array is recovered, data to be recovered are obtained in the storage position according to the preset reading sequence, the data to be recovered are utilized to recover the data of the disk array, and the data to be recovered are contained in the target data. The problems that when RAID data are stored, write-in operation needs to be repeatedly carried out, so that storage space loss is caused, and the storage process is very time-consuming due to log introduction are solved.

The invention relates to a data processing method and system of an NAND FLASH storage array, computer equipment and a storage medium. The method comprises the following steps: receiving a data writing or reading instruction; and analyzing the instruction through the system-on-chip, and distributing the data to a target module in the plurality of NAND FLASH modules connected in parallel according to a preset strategy. And writing the data into the target module or reading the data from the target module through the control and communication line. And executing data management operation through the system-on-chip, specifically, writing data into different NAND FLASH modules in a dispersed manner by adopting a dynamic wear leveling algorithm so as to balance the erasing times of each module. And detecting and marking bad blocks in the NAND FLASH module in real time, and avoiding the bad blocks in calling operation after data storage. By the adoption of the method, the storage performance, reliability and expandability of satellite-borne storage can be comprehensively improved, and the method is suitable for task scenes with high reliability requirements.

The invention belongs to the technical field of extensible computer architecture, and particularly relates to a two-level cache coherence protocol based on a many-core computing architecture. Comprising a many-core computing architecture, the many-core computing architecture comprises a plurality of computing clusters, each computing cluster is formed by interconnection of a plurality of core grains and an I / O and memory controller, and the I / O and memory controller is interconnected with at least two memory DDRs; each core particle is formed by interconnection of a plurality of computing cores and a shared cache LLC; each calculation core is provided with a private cache, and when the data is not in the private cache, the memory access request continues to go to the next-level shared cache LLC or the memory DDR until the latest data is obtained; the shared cache LLC can access the memory DDR only after the shared cache LLC is in butt joint with the memory controller through high-speed interconnection and I / O; and the two-level cache consistency protocol directory is composed of a first-level directory and a second-level directory. According to the invention, the delay of re-reading from the storage system is omitted, the performance is further improved, and the congestion at the storage position is reduced.

The invention relates to the technical field of computer storage, in particular to an intention perception data prefetching method and device of solid-state storage equipment and a storage medium. The method comprises the following steps: collecting user behavior data, wherein the user behavior data comprises an application use frequency, file access time, a system call log, a hard disk reading track and network flow information; classifying, filtering and preprocessing the collected user behavior data to generate a time sequence and a behavior sequence, and mapping the time sequence and the behavior sequence into multi-dimensional feature vectors; according to the intention perception data prefetching method of the solid-state storage device, dynamic and efficient data prefetching is achieved in combination with real-time analysis of the user operation intention, and the data access speed and the overall performance of the solid-state storage device are improved.

The invention relates to the technical field of data processing, and provides a BMC fault recording method and device, computer equipment and a storage medium, and the method comprises the steps: responding to a kernel fault of a target operating system of a target BMC, and printing fault log information; obtaining a pre-configured memory technology device partition; transmitting the fault log information to an annular fault cache region to form cache log data; storing the cached log data in a memory technical equipment partition, and generating a fault log file; and in response to a user reading instruction, reading and outputting the fault log file of the memory technical equipment partition. According to the technical scheme provided by one or more embodiments of the invention, the fault log can be effectively recorded, maintenance personnel can timely and accurately know the fault position and the fault reason, and BMC fault diagnosis and repair can be efficiently completed.

The invention provides a training method, a reasoning method, a device and equipment for a large language model. Relates to the technical field of computers. The method comprises the steps of obtaining a training text, wherein the training text comprises a first text fragment which is not labeled and a second text fragment which is labeled as a block text; inputting the training text into a large language model, and processing the second text fragment through a block attention mechanism module in the large language model; processing the first text fragment through a standard attention mechanism module in the large language model; calculating a first loss based on a prediction result of the first text fragment output by the large language model; calculating a second loss based on a prediction result of a second text fragment output by the large language model; determining total loss according to the first loss and the second loss; and optimizing parameters of the large language model based on the total loss to obtain a trained large language model. According to the method, the memory reading frequency is reduced, and the efficiency of large language model reasoning is improved.

Methods, systems, and computer-readable storage media for receiving, from a first entity, a first call, the call being received by a middleware client of a second entity, storing a set of headers associated with a workflow in a request context holder of the middleware client, and determining that a second call is to be executed from the middleware client of the second entity for the workflow, and in response: reading, by a client of the middleware client, the set of headers associated with the workflow, including the headers in a request of the second call, determining an endpoint for the second call, and transmitting the second call for receipt at the endpoint.

Disclosed in the embodiments of the present disclosure are a data reading method for a chip, and a chip, a computer device, a storage medium and a computer program product. The method comprises: acquiring a data reading instruction, wherein the data reading instruction is used for instructing data reading to be performed according to reading information and data prefetching to be performed according to prefetching information; on the basis of the data reading instruction, acquiring a first data reading request, wherein the first data reading request includes the reading information and the prefetching information; sending the first data reading request to a cache module, so that in response to the first data reading request, the cache module reads and returns first target data according to the reading information, and pre-fetches and caches second target data according to the prefetching information; and receiving the first target data returned by the cache module. The embodiments of the present disclosure can reduce data request pressure between a request source and a cache module, compress a request bandwidth, improve the performance of a processor, and facilitate flexible control over a prefetching mode on a per-task basis, thereby better satisfying data reading requirements of tasks.

The invention discloses a data processing method, electronic equipment, a medium and a computer program product, and relates to the technical field of computers, and the data processing method comprises the following steps: firstly, arranging feature data blocks by taking batches as elements, arranging weight data blocks by taking output channels as elements, and accessing feature data and weight data more orderly in convolution calculation; then, a proper storage mode is selected by judging the scale of network data and a set threshold value; when the data scale is smaller than a set threshold value, convolution calculation is carried out after data are loaded in an on-chip storage mode, data reading delay can be reduced, and the calculation response speed can be increased; when the data scale is large, an off-chip memory access mode is adopted to load data while convolution calculation is carried out, so that idle calculation resources caused by waiting for all the data to be loaded are avoided. In this way, the convolution calculation part for feature extraction of the reinforcement learning front end can be cooperatively deployed to the block matrix calculation architecture, resource reuse is achieved, and hardware implementation complexity and hardware resource consumption are effectively reduced.

The invention discloses a memory fault processing method and device, electronic equipment and a readable storage medium, and relates to the technical field of computers.The method comprises the steps that when a memory line fault exists in a memory of a target system, an isolation operation is triggered for a target physical address, the target physical address at least comprises the physical address of the fault line with the memory line fault in the memory, the technical problem of system downtime caused by the fact that the memory line fault cannot be rapidly processed in the related technology is solved, the isolation process is started for the target physical address in time when the memory line fault is detected, and the system downtime is reduced. Performance reduction and data error accumulation caused by the fact that the fault line continues to participate in data reading and writing are avoided, and then the technical effect of reducing the risk of system downtime is achieved.

Embodiments of the present invention provide a storage die, a data processing method, and a computer system. The storage die includes a die interconnection interface, a computing module, and a storage module. The die interconnection interface is configured to receive a data processing command sent to the storage die and pass the data processing command to the computing module. The computing module is configured to read target data from the storage module based on the data processing command, perform an operation on the target data based on the data processing command to obtain an operation result, and store the operation result in the storage module based on the data processing command. The storage module is configured to store data. The technical solution provided by the embodiments of the present invention realizes data reading, operation, and storage of operation results inside the storage die by setting a computing module inside the storage die, avoiding the realization of both data reading and storage by the die interconnection interface of the storage die, so that the bandwidth pressure and power consumption of the die interconnection interface of the storage die can be reduced.

The invention relates to a database migration method, device and system, computer equipment, a storage medium and a computer program product. The method comprises the following steps: starting a first double-writing single-reading state of a target data server corresponding to a target service; the first double-writing single-reading state is used for indicating to write data to the corresponding source database and target database and read data from the source database; under the condition that it is determined that the data writing operation for the target database is normally executed, switching from a first double-writing single-reading state to a second double-writing single-reading state; the second double-writing single-reading state is used for indicating data writing to the source database and the target database and data reading from the target database; switching from a second double-writing single-reading state to a second single-writing single-reading state under the condition of determining that the data writing operation and the data reading operation aiming at the target database are normally executed; the second single-write single-read state is used for indicating to read and write data from the target database. By adopting the method, the database migration efficiency can be improved.

The invention discloses a memory fault prediction method and device, a chip, equipment and a storage medium, and belongs to the technical field of computers. The method comprises the steps that error position distribution of an error correction unit associated with a memory is obtained, the memory comprises at least one memory particle, the error correction unit comprises a plurality of data bits obtained from a plurality of storage units of the memory through continuous multiple burst reading, and each data bit corresponds to one storage unit in the memory; the error position distribution is used for indicating the position of an error data bit in a plurality of data bits, and the error position distribution causes the occurrence of CE in the memory; a memory fault model is obtained, the memory fault model comprises at least one reference position distribution, and the reference position distribution is associated with the UCE of the memory; according to the similarity between the error position distribution and the at least one reference position distribution, a prediction result is output, and the prediction result is used for indicating the probability that the UCE occurs in the memory. The method is beneficial to improving the timeliness of memory fault prediction.

The invention provides a pointer type instrument automatic reading and verification method based on deep learning, and relates to the technical field of computer vision, and the method comprises the steps: 1, obtaining various types of pointer type instrument images; 2, performing data enhancement on the images in the training set; step 3, performing dial target detection and image segmentation on the instrument image; 4, performing tilt correction on the pointer type instrument picture in combination with an image segmentation result; 5, key point detection is carried out on the instrument image; 6, performing reading calculation by using an angle method to obtain the reading of the instrument; and step 7, obtaining the reading of the standard digital meter in the verification process. According to the invention, the precision of instrument dial detection in a complex environment is improved through a deep learning technology, and manual intervention is reduced; and meanwhile, the automation level and the intelligence level of instrument reading and verification are improved, higher detection efficiency and standardization are achieved, and wide application value is achieved.

A display control system includes a first wireless device attached to a management target and storing first identification information for identifying the management target, a second wireless device installed at a predetermined position and storing second identification information, a reading apparatus, at least one processor, and at least one memory. The processor and memory are configured to estimate a located position of the management target based on reading results of the first identification information at a first reading time and the second identification information at a second reading time and cause a display apparatus to display an estimated area within which the management target is estimated to be located based on the located position. A magnitude of the estimated area is based on an amount of movement of the reading apparatus between the first reading time and the second reading time.

The invention provides a data processing method and device of a storage device, a medium and computer equipment. The data processing method comprises the following steps: analyzing a reading request initiated by a user, and obtaining a logical addresses to be read; determining c L2P mapping tables associated with the a logical addresses according to the range of the a logical addresses, wherein c is a positive integer greater than or equal to 1; backing up an original mapping relation of the c L2P mapping tables; b latest mapping relations of a temporary information layer are updated to the c L2P mapping tables, and the original mapping relations of the c L2P mapping tables are covered; querying the updated c L2P mapping tables according to the reading request, obtaining physical addresses corresponding to the a logical addresses, and reading data; and after the reading is completed, restoring the backed-up original mapping relation to the c L2P mapping tables. Through the mapping management process of the scheme, redundancy check can be avoided, accurate locking of an updating range is realized, the check times are greatly reduced, physical limitation of bandwidth is released, and delay is greatly reduced.

The invention discloses a method for providing integrity guarantee during state data reading, a method for realizing state index recovery and computer equipment. The method for providing integrity guarantee during state data reading comprises the steps that in the process of reading state data from a persistent storage medium to a memory, the state data corresponding to a version number and version end meta-information are read; the state data comprises a state key-value, a version number and an operation mark, and the operation mark is used for representing a writing sequence of the corresponding state data written into the persistent storage medium; the operation marks are compactly and progressively increased according to the sequence of the write-in state data; the version ending meta-information comprises a version number of the block and a range of an operation mark of writing state data of the corresponding block into the persistent storage medium; and filtering the state data corresponding to the read version number based on the version number and the operation marking range in the version ending meta-information.

The invention provides a method and a device for adjusting read voltage by recording programming positions, computer equipment and a storage medium. The method comprises the following steps of: constructing a nearest programming position table according to a nearest programming page number of an unprogrammed block, and pre-constructing a distance-state conversion table and a state-voltage conversion table; receiving a host read command, and judging the position information of a target block in the nearest programming position table; calculating and converting the position information through the distance-state conversion table and the state-voltage conversion table to obtain a read voltage; and performing a read operation based on the read voltage. According to the invention, the problems of high bit error rate of unprogrammed blocks and waste of power consumption caused by a fixed voltage strategy in a traditional storage system are solved.

The invention discloses a method, a device and equipment for displaying a register address in a programmable logic controller, and relates to the technical field of computers, the method comprises the following steps: in response to a received register address display request, executing a reading operation on a register address storage area of the programmable logic controller to obtain read data, the read data comprises a plurality of byte data arranged continuously, and a part of byte data is used for identifying a starting point position of a register address; executing analysis operation according to different analysis data lengths from a starting point position of a register address in the read data to obtain a plurality of pieces of analysis data; one of the multiple pieces of analysis data corresponding to each starting point position is a register address; and performing distinguished display on the analysis data.

The invention provides a DSL (Digital Subscriber Language)-based public dimension data conversion method and system, and the method comprises the following steps: reading an input script as a file stream by using a business language script reading module; segmenting characters in the file stream into tokens by utilizing a lexical analysis module; using a grammar parser module to construct the token into a grammar tree; traversing all nodes of the syntax tree by using a syntax traversal module, and triggering a conversion rule monitor module at each node; and extracting information by using the conversion rule monitor module, and assembling the information into a computer language program. According to the technical scheme, modeling is carried out on the service function through the configuration data conversion assembly of the DSL, the model is formed, an existing model or a newly-built model is introduced during modeling, and model verification is carried out on a system, so that the condition of repeated modeling is avoided, and a bloated model can be simplified.

The invention discloses a key product data acquisition method and device, equipment and a storage medium, and relates to the technical field of computers. The method comprises the following steps: reading newly-added key product data from a backboard of cluster equipment according to a first path through a platform module of a cluster, and storing the newly-added key product data to a cache; the first path comprises a platform module, a monitoring module, a storage case service module and a backboard in sequence; reading initial key product data from a backboard of cluster equipment according to a second path through a case management module of the cluster, and reading the newly added key product data from the cache through the case management module; the second path sequentially comprises a case management module, a storage case service module and a back plate; and obtaining complete key product data based on the newly added key product data and the initial key product data. The reading time of key product data can be shortened, and the case management module is prevented from reading the product data overtime.

The invention provides a request response method and device, equipment, a medium and a product, can be applied to the technical field of big data and the technical field of computers, and comprises the following steps: in response to a received data reading request, determining a target node set according to node states of a plurality of reading nodes in a server cluster and priorities of the reading nodes, the priority is determined according to a matching relationship between a request address of the data reading request and an operation address of each sub-cluster in the server cluster; according to the weight of each read node in the target node set, multiple accumulative weight intervals are determined, and the weights are obtained based on hardware configuration and performance data evaluation of the read nodes; matching a random value generated in the cumulative weight interval with each cumulative weight interval to determine a target weight interval corresponding to the random value; and executing a data reading operation by using a reading node corresponding to the target weight interval in the target node set.

Disclosed are a residual data writing method and apparatus, a computer device, and a storage medium. The method includes: determining a residual data reading mode comprising a reading sequence of each piece of residual data, and the quantity of residual data read simultaneously at each time; determining the number of buffer groups and the number of buffers in each buffer group according to the residual data reading mode, wherein each buffer is provided with a unique buffer index, and each buffer comprises a plurality of buffer addresses; acquiring a buffer index and a buffer address corresponding to to-be-written residual data; and writing the to-be-written residual data into a corresponding buffer according to the buffer index and the buffer address.

The invention discloses a multi-person online map access method and device, a computer storage medium and equipment, and the method comprises the steps: responding to a write-in request of a client, storing to-be-stored map data in a first storage area and a second storage area, and achieving the double-layer storage of the to-be-stored map data, quick storage of the to-be-stored map data can be ensured, and the map data can be ensured not to be lost through the second storage area; in response to a reading request of a client, obtaining position coordinates in the reading request; coding based on the position coordinates to obtain routing keys corresponding to the position coordinates; reading target map data corresponding to the routing key from the first storage area or the second storage area; therefore, the real-time access of the map data can be ensured during large-scale multi-person online, and the required storage cost is lower.

The invention relates to the field of computer hardware virtualization, and discloses a CAN network virtualization method, system, device and medium, a virtual CAN network is established through QEMU virtual machines, each QEMU virtual machine can be used as a node, all the nodes are connected through a Kvaser Virtual CAN Driver driver to form a virtual CAN network, the virtual CAN network can be used for simulation and testing, and the virtual CAN network can be used for simulating and testing. Therefore, logic analysis of the virtual device is separated from actual operation of the device, the front end is mainly used for reading, writing and the like of a register, and when hardware operation is involved, the operation is transmitted to the rear end, so that coupling is reduced, the flexibility of a model is improved, and the reliability of the model is improved. The problem that drivers and application software cannot be developed and verified in advance due to high CAN network hardware cost and long development period is solved.

The invention discloses an automatic calibration method based on deep learning multimeter reading identification and application thereof, and the method comprises the steps: obtaining calibration information, and transmitting the calibration information to a server; in response to the calibration request, sending a starting instruction to a server; acquiring a multimeter image in real time, extracting a real-time indicating value from the multimeter image based on a neural network model, and sending the real-time indicating value to an upper computer; the neural network model comprises a backbone network, a neck network and a head network, the backbone network comprises a low illumination enhancement module, a down-sampling layer and a fast spatial pyramid pooling layer, the down-sampling layer comprises alternative standard convolution modules and C2fGhostV2 modules, the C2fGhostV2 modules, and the fast spatial pyramid pooling layer is connected to the neck network through a simAM module. According to the method, the calculation amount and the parameter amount are effectively reduced, the calculation efficiency is improved, the robustness to a complex background is enhanced, full-process automation is realized, and the calibration efficiency, precision and stability are improved.

The invention discloses a data access control method and device, a storage medium and electronic equipment, and relates to the field of computers.The data access control method comprises the steps that a target reading request is received, and the target reading request is used for requesting to read data corresponding to target time from a storage space in a memory according to target address information; in response to the target reading request, positioning a target cache region corresponding to the target address information in a cache space in a memory; and extracting the target data of the target snapshot version corresponding to the target time from the reference snapshot version and the reference data which are stored in the target cache region and have the corresponding relationship, thereby solving the technical problems of relatively low data access efficiency and the like caused by the fact that historical snapshot data reading needs to depend on a rear-end disk. The technical effects of cache acceleration of historical snapshot data reading and improvement of data access efficiency are achieved.