108 results about "Dynamic storage" patented technology

The invention belongs to the field of computer system structures and artificial intelligence accelerators, and discloses a software and hardware collaborative heterogeneous storage computing accelerator oriented to full-chip DLRM reasoning. A control domain composed of a static random access memory and an embedded dynamic random access memory, an on-chip storage domain, an in-memory computing domain and an on-chip computing domain are integrated in the same chip, an embedded table is compressed by introducing a hierarchical clustering quantization frame, and a dynamic storage management strategy of access frequency perception is combined. The embedded access path is shortened, and the bandwidth requirement is reduced. According to the method, the reasoning process of the whole set of DLRM can be independently completed, and the memory access overhead in a traditional terminal SoC is remarkably reduced by reducing cross-domain data migration between the calculation unit and the storage unit, so that the model reasoning performance and energy efficiency are improved.

The invention discloses an intelligent storage automatic control optimization method and system based on dynamic environment perception, and relates to the technical field of intelligent storage control. According to the method, a specific scene type of a specified warehouse is firstly obtained, a corresponding dynamic perception allocation scheme is determined through scene matching, and then a collaborative operation association object is determined, so that accurate adaptation of dynamic environment perception and a warehouse scene is realized; then, collecting dynamic environment sensing data of a storage area where a collaborative operation association object is located, and after collaborative association processing is carried out, generating a structured fusion data set adapted to different time sequence scenes; dynamic storage parameter calibration is carried out, an initial control strategy is generated, and effectiveness evaluation is completed; whether the path planning scheme and the operation execution sequence are dynamically adjusted or not is decided according to the evaluation result, dynamic adaptation and accurate regulation and control of intelligent storage automatic control are achieved, and the adaptability of storage control, the smoothness of process operation and the stability of operation execution in a complex dynamic scene are effectively improved.

The invention relates to the technical field of data processing, and discloses a compiling method and system for generating hardware accelerator executable codes facing a multi-mode large language model, a terminal and a storage medium. The method comprises the following steps: constructing a computational graph according to the structure of a to-be-deployed large language model and operators supported by a hardware platform; traversing the computational graph to deduce output tensor type information of each operator, and verifying the type consistency between adjacent operators; based on the verified type information, executing static memory allocation to determine a static storage address of the weight data, and executing dynamic memory allocation to determine a dynamic storage address of the state data; then combining the verified type information and the static and dynamic storage addresses to generate an intermediate representation instruction sequence for configuring a hardware acceleration platform register; and finally, converting the intermediate representation instruction sequence into a target code file which can be directly loaded and executed by a hardware acceleration platform. The compiling efficiency, the deployment reliability and the resource utilization rate are improved.

This invention provides a data storage method and system for tuberculosis detection, relating to the field of tuberculosis detection technology. The data storage method for tuberculosis detection includes the following steps: S1. Collecting whole-genome sequencing data of Mycobacterium tuberculosis, host serum IgG titer, and drug sensitivity test results; S2. Calculating genetic distance D based on a reverse evolution model to generate four-dimensional spatiotemporal coordinates (t, x, y); S3. Performing data partitioning and storage based on the drug target barrier value β; S4. Generating dynamic metadata using a host-pathogen dynamics model and compressing and storing it. This invention implements a dynamic storage entropy adjustment algorithm at the hardware and software collaborative level, continuously optimizing the matching efficiency of data compression and physical storage. This results in an intelligent data hub that can perceive the evolutionary pulse of pathogens and autonomously optimize resources, providing support for clinical tuberculosis prevention and control decisions with temporal depth, spatial correlation, and risk evolution.

The embodiment of the invention discloses a video data management method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining a video code stream from each video channel, and compressing the obtained video code stream; according to the historical access data and the scene label corresponding to each video channel, configuring a dynamic storage period corresponding to each video channel; according to the dynamic storage period corresponding to each video channel, storing the compressed video code stream into a target storage area; detecting the residual storage space of the target storage area; when the residual storage space is smaller than a preset early warning storage threshold value, the target video code stream is deleted according to the access frequency of each video code stream and the dynamic storage period corresponding to each video channel, and according to the scheme, storage resource distribution can be optimized, storage occupation can be reduced, and key videos are prevented from being deleted by mistake.

The application discloses a high-density stereoscopic warehouse dynamic storage space allocation method for WMS, and relates to the technical field of artificial intelligence. Stereoscopic warehouses are arranged along the direction of the roadway, and stackers, shuttles and hoists are arranged in the stereoscopic warehouses. Field programmable gate array acceleration cores are arranged in the edge nodes of the stereoscopic warehouses. The acceleration cores are interconnected with the WMS through a time-sensitive network. The method comprises the following steps: step 1, the acceleration core constructs a beat dependency graph based on the current carrying instruction issued by the WMS, wherein the carrying motion is taken as an event node, and the action connection relationship is taken as a directed edge; step 2, the acceleration core constructs a colored Petri net; and step 3, the acceleration core constructs a graph attention reinforcement learning decision module and takes a heterogeneous graph as input, and the WMS drives the corresponding carrying equipment to perform the actions of warehousing, moving and replenishment. The application can significantly improve the system throughput, reduce the equipment idle rate and beat jitter under the sub-millisecond end-to-end decision delay.

One or more techniques and / or systems are provided for dynamically provisioning logical storage pools of storage devices for applications. For example, a logical storage pool, of one or more storage devices, may be constructed based upon a service level agreement for an application (e.g., an acceptable latency, an expected throughput, etc.). Real-time performance statistics of the logical storage pool may be collected and evaluated against the service level agreement to determine whether a storage device does not satisfy the service level agreement. For example, a latency of a storage device within the logical storage pool may increase overtime as log files and / or other data of the application increase. Accordingly, a new logical storage pool may be automatically and dynamically defined and provisioned for the application to replace the logical storage pool. The new logical storage pool may comprise storage devices expected to satisfy the storage level agreement.

This invention discloses a dynamic data area sharing system and method for Flash chips, belonging to the field of computer firmware technology. The system includes a BIOS core framework, a shared storage component, and a dynamic shared storage area. The BIOS loads and runs the shared storage component through the BIOS core framework. The shared storage component can dynamically use the shared storage area according to the needs of the current computer boot, performing BIOS upgrade backups and BIOS log recording. This invention solves the problem of BIOS backup and BIOS log storage by dynamically dividing the SPI Flash memory into a storage area when the BIOS space on the computer motherboard is limited.

The invention discloses an intelligent warehouse dynamic storage location optimization system and method based on AI and digital twinning, and relates to the technical field of storage location optimization. Obtaining current inventory state information of the goods shelf; establishing a nearest goods storage distance constraint function, a goods shelf bearing capacity constraint function and a goods shelf stability constraint function based on the distance between each storage location and the warehouse entrance and the weight of the unit of goods stored in each storage location, and establishing a target function; initializing a particle swarm based on the number of units of goods to be stored and the current inventory state information of each goods shelf; solving the particle population by using a genetic algorithm to obtain an optimal particle individual; based on the optimal particle individuals, determining a storage location allocation scheme of each to-be-stored cargo unit; and establishing a digital twinborn model of the warehouse, and highlighting the storage location model of each to-be-stored cargo unit in the digital twinborn model based on the storage location allocation scheme of each to-be-stored cargo unit. The warehouse-in and warehouse-out efficiency can be ensured, and meanwhile, the goods storage safety is considered.

The invention relates to the technical field of software defined storage, and discloses a dynamic storage distribution system and method of a radio frequency signal broadband stream disk, comprising: a parameter configuration module analyzes radio frequency data flow rate and storage array physical parameters, and an address mapping module, based on a data frame time index, stores the data frame time index; executing modular operation by utilizing a step length factor which is co-prime with the total storage stripe number, and generating a target logic address of discrete jump; the strip assembly module constructs a standard write-in unit according to a physical alignment size, and the direct memory access control module constructs a straight-through transmission path and dispersedly writes data into different physical channels. Through a modal co-prime mapping mechanism, full-space deterministic discrete distribution of write-in addresses is realized on a pure logic level; physical single-point congestion during multi-channel concurrent writing is eliminated, I / O long tail delay is reduced, and zero frame loss and time sequence stability of a high-bandwidth radio frequency flow disc are ensured.

The utility model discloses a kind of feeding devices of elastic belt material, the device includes rack, dynamic storage unit and deviation rectifying unit, dynamic storage unit includes feed drive roller and discharge drive roller, and elastic belt material is transported forward with natural overhanging state between feed drive roller and discharge drive roller, by adjusting the relative speed of feed drive roller and discharge drive roller, to control the overhanging height of elastic belt material between feed drive roller and discharge drive roller, make the tension of elastic belt material after material roll output can be fully released, avoid material in subsequent processing due to tension instability causes size leg, lining curling etc. It can be widely applied to the processing field of elastic flexible material such as sanitary product, clothing etc.

The invention relates to a trackless railway wagon wheel set storage selecting and matching system and method. The system comprises an income scanning area, a dynamic storage area and a selecting and matching expenditure area which are longitudinally and linearly arranged. A plurality of transversely arranged turntables are arranged at the starting point of the incoming scanning area, and a plurality of longitudinal routes corresponding to the turntables and parallel to one another are arranged in the incoming scanning area; a scanning recognition device is arranged at the starting point of the dynamic storage area and located beside the extension line of each longitudinal route of the collection scanning area, and a truss manipulator assembly and a ground cross fixed wheel pair device are arranged in the dynamic storage area. The end point of the selective distribution expenditure area is also provided with a plurality of turntables which are transversely arranged, and a plurality of longitudinal routes which correspond to the turntables and are parallel to one another are arranged in the selective distribution expenditure area. The core shackle of traditional track storage is broken through, the three areas of income scanning, dynamic storage and selective distribution expenditure are planned and arranged, the AGV and the truss manipulator are combined, and a basic hardware facility is provided for income, storage, distribution and selective distribution of wheel sets.

The invention discloses an intelligent warehouse management method and platform for pericarpium citri reticulatae quality tracing based on Ai, and relates to the technical field of project management and the technical field of pericarpium citri reticulatae warehouse management. A dynamic parameter output module inputs a structured data set into a pericarpium citri reticulatae aging quality-warehouse environment matching model pre-constructed based on a deep learning algorithm; the dried orange peel aging quality-storage environment matching model fuses dried orange peel raw material attributes and storage environment parameters and outputs two types of dynamic storage regulation and control parameters, and the strategy generation module generates a regulation and control strategy according to the output two types of dynamic storage regulation and control parameters in the actual storage process of dried orange peel. And the internal environment autonomous balance system and the external environment precise intervention system are respectively driven to cooperatively operate. According to the management platform, collaborative operation of autonomous balance of the internal environment of the warehouse and precise intervention of the external environment is realized, and the problems of mildewing, fragrance loss or uneven aging and the like caused by empirical regulation and control of traditional warehouse are effectively solved.

The invention provides a hierarchical storage management method based on land observation satellite data value, which belongs to the field of application in a land observation satellite data processing system, and comprises the following steps of: calculating an initial comprehensive weight score during data archiving by creating seven types of core data tables such as a satellite parameter table and a data value parameter table; determining the number of storage days of the online or near-line cluster in combination with a dynamic storage strategy table; obtaining user order update data return times in real time, and recalculating the weight score at regular time to dynamically adjust the storage strategy; and finally, realizing data cleaning in combination with the water and fire wire threshold of the storage device. The problems of low resource utilization efficiency and slow user data return caused by making a storage strategy only based on the number of storage days in the prior art are solved, and the storage resource utilization rate and the user data response efficiency are remarkably improved.

The invention relates to a data storage method and device, computer equipment and a storage medium. The method comprises the following steps: acquiring a data storage request, and determining a target storage area and a data block index of a target data block from candidate storage areas of a memory according to the data storage request; if the target storage area is the static data storage area, determining a physical address of the target data block according to the data block index, and storing the to-be-stored data into the target data block based on the physical address of the target data block; and if the target storage area is a dynamic data storage area, determining an initial write-in address in the dynamic data storage area based on the data block index, and storing the to-be-stored data into the target data block based on the initial write-in address. According to the scheme, the security of data storage can be ensured, unnecessary operation codes are reduced for both dynamic storage and static storage, and the stability of data storage is improved.

Embodiments of the present application disclose a method and device for taking out and placing goods, a warehousing robot and a warehousing system. For the warehousing system based on a dynamic storage location storage mechanism, the method includes: moving, according to an operation instruction for target goods, to a target position, where the target position is a position corresponding to a dynamic goods storage space for the target goods; acquiring first state information and second state information, and determining, according to the first state information and the second state information, available spacing between the target goods and adjacent objects, where first state information is used for representing a spatial position feature of the target goods and the second state information is used for representing spatial position features of the adjacent objects; and taking out or storing, according to the available spacing, the target goods.

The invention relates to a cloud database computing cache system and an implementation method, and the system comprises a decentralized cache architecture module which is used for eliminating a centralized bottleneck, supporting ten-thousand-level concurrency and adapting to a big and small file mixed scene; the memory object sharing module is used for reducing intermediate result processing delay through intermediate result full-life-cycle zero-copy circulation and dynamic storage decision based on a memory pool and metadata service of the memory object sharing module; the self-adaptive asynchronous write-in module is used for connecting remote storage, multiplexing memory object sharing module resources and combining mechanisms such as delay confirmation, so that the write response speed is increased, remote requests are reduced, data integrity is guaranteed, and meanwhile, persistent support is provided for intermediate results. Through cooperative operation of the three modules, the cloud database computing cache system and the implementation method can effectively adapt to the requirements of high concurrency, large and small file mixing and elastic capacity expansion and contraction of a cloud native OLAP scene, and the integrity of data transmission and the stability of system operation are guaranteed.

This invention provides a method for predicting the surface stress distribution of a three-dimensional dynamic storage capacitor, comprising: establishing a TCAD process simulation program according to the manufacturing process of the three-dimensional dynamic storage capacitor to obtain the three-dimensional simulation structure of the capacitor; performing stress simulation on each step of the manufacturing process based on the three-dimensional simulation structure of the capacitor, so that the stress in the structure is rebalanced after each step of the process is completed; obtaining stress distribution maps under different parameter combinations by changing the structural parameters, material parameters, and process parameters of the device; obtaining a dataset based on the stress distribution maps under different parameter combinations; dividing the dataset into a training set and a test set, and training it using a neural network to obtain a trained prediction network; and using the trained prediction network to predict the stress distribution of each layer under different input parameters. This invention can efficiently predict the surface stress distribution of 3D capacitors, significantly improve the performance of 3D capacitors, and is beneficial for optimizing device design and improving device stability.

The memory sub-system controller performs different memory operations based on a host profile. The controller accesses a host behavior model comprising an amount of data and timing information. The amount of data can represent how much data was written by a host system to a set of memory components in a past time period. The timing information can indicate when program requests were received from the host system to program the data during the past time period. The controller searches the timing information of the host behavior model based on a current time to identify a time point in the host behavior model that corresponds to the current time. The controller retrieves, from the host behavior model, a parameter that corresponds to the identified time point and performs one or more memory operations based on the parameter retrieved from the host behavior model.

The utility model relates to the technical field of linear cutting machining, in particular to a linear cutting waste placing device. A wire cutting waste placing device comprises a support formed by splicing section bars, a liftable mounting plate is arranged in the support, one end of the mounting plate is movably provided with an overturning plate through a shaft, and a waste box is placed on the overturning plate. The wire cutting waste placing device comprises a support formed by splicing profiles, a mounting plate capable of vertically ascending and descending is arranged in the support, one end of the mounting plate is movably connected with an overturning plate through a rotating shaft, a waste box is borne on the surface of the overturning plate, the overturning function is achieved through the rotating shaft, and the waste box inclines along with the overturning plate to complete waste dumping. The whole structure is supported by the support, ascends and descends through the mounting plate and is movably matched with the overturning plate, and dynamic storage and transfer of waste are achieved.

The disclosure describes systems, devices, and methods for managing data storage environments. In an example implementation, a method of operating a controller in a data storage environment is provided. In performing the method, the controller identifies a change to a layout of drives in the data storage environment, and in response to identifying the change, takes a lock on instances of layout metadata stored on the drives. The controller then updates the instances of the layout metadata to reflect the change to the layout and releases the lock.

This invention discloses a spatially dynamic reconfigurable three-dimensional parking garage system and its control method. The internal space of the garage is defined as a reconfigurable resource library composed of several discrete location units, and a large number of intelligent transport units that can move autonomously and mechanically connect with each other are configured. When parking a vehicle, the required number of transport units is dynamically determined based on the vehicle's actual dimensions. These transport units are temporarily assembled at the entrance into an irregularly shaped transport platform adapted to the vehicle's contours. After the vehicle stops on the platform, the transport unit group disassembles and moves to its assigned discrete location unit for reassembly, thus realizing the dynamic generation and release of parking spaces. Retrieval is performed in reverse. This invention achieves conflict-free scheduling of a large number of transport units through distributed collaborative control, significantly improving space utilization and system fault tolerance, and realizing the transformation of three-dimensional parking garages from fixed storage to dynamic storage.

According to the hardware-network collaborative optimization FPGA digital recognition system and method, the LeNet-5 network architecture is improved, the number of channels is adjusted to be a multiple of 4 so as to be matched with an FPGA parallel computing array, a ReLU activation function is adopted to replace traditional Sigmoid / Tanh, and 8-bit integer quantization processing is implemented; a four-level assembly line architecture (convolution-quantization-pooling-activation) is adopted in the aspect of hardware, a snakelike movement strategy and a double-Block dynamic storage scheme are combined, the energy efficiency ratio and the reasoning speed are remarkably improved while high recognition precision is kept, an end-to-end real-time processing function chain from image collection, neural network reasoning to result display is achieved, and the method has the advantages of being high in recognition precision and high in practicability. The data reuse rate is obviously improved, the storage access bandwidth is reduced, a referenceable low-power-consumption and high-performance solution is provided for edge computing applications such as intelligent monitoring and industrial quality inspection, and high real-time performance and practicability are achieved.

This invention relates to the field of data encryption, and more particularly to a data encryption method and system based on the Glink bus. The invention extracts a local interface fingerprint feature set and transmits it to a neutral terminal. The neutral terminal extracts interface fingerprint features within a local time domain segment and sends them to the transceiver end. It dynamically determines the data storage link to be activated, generates corresponding interference fragments, inserts them into the interface fingerprint feature fragments to form an interference fingerprint feature sequence, and stores it at each node of the data storage link. The data at each node is switched according to a switching rule. The transceiver end uses the received partial interface fingerprint feature set to determine the node to be accessed, extracts the interface fingerprint feature fragments, and then determines the encryption and decryption algorithms. This invention utilizes the unpredictability of interface fingerprint features to dynamically store key materials, uses the acquisition process of the key materials itself as an encryption method, and updates the encryption and decryption algorithms in a correlated manner, thereby reducing the risk of key materials being cracked and hijacked, and improving encryption security.

The embodiment of the invention provides a storage resource switching system, method and equipment, a storage medium and a computer program product, and is applied to the technical field of computers. In the embodiment of the invention, in the process of switching the storage resources for the application program, the incidence relation between the PV and the storage resources is decoupled through the dynamic storage object, so that the incidence relation between the dynamic storage object and the storage resources can be dynamically maintained under the condition of ensuring that the incidence relation among Pod, PVC, PV and the storage resources is not changed, and the user experience is improved. Therefore, one PV can be bound with different storage resources according to needs, the capacity of dynamically switching the storage resources for the application program is achieved under the condition that Pod, PVC and PV are not reconstructed, the problem that Pod, PVC and PV need to be reconstructed for updating the storage resources for the application program is solved, a switching link for switching the storage resources for the application program is simplified, the switching process is made to be lighter, and the switching efficiency is improved. And the switching efficiency is improved.

The invention relates to the technical field of data storage, in particular to a large-scale data storage optimization method and system based on artificial intelligence. The method comprises the following steps: acquiring large-scale data, and performing multi-dimensional feature analysis and business influence evaluation based on artificial intelligence to obtain a business influence degree corresponding to each piece of sub-data; access frequencies and access time intervals corresponding to all the sub-data are obtained, importance level classification evaluation and storage device hierarchical mapping are carried out, and corresponding structured data and unstructured data in the large-scale data are stored in corresponding storage devices in a layout optimization mode; according to the method, the disk I / O rate and the storage utilization efficiency corresponding to the storage device are obtained, dynamic storage optimization adjustment is carried out, a large-scale data device dynamic storage strategy is generated, corresponding device dynamic storage layout adjustment work is executed, and the corresponding storage access speed is increased. The efficiency, reliability and space utilization rate of large-scale data storage can be improved.

The present application relates to the technical field of financial information processing, and is especially a kind of automatic delivery and single processing system of returning of bill fusion financial management platform, including single machine equipment, financial management platform and interface module;System is through the dynamic adjustment of the parameter of OCR model by the multi-factor weighted nonlinear optimization algorithm combined with resolution, contrast, noise level, illumination condition and image distortion degree, etc. Quality factor, realize the high-precision identification of low-quality bill image;Single processing module based on SVM model, introduce multi-dimensional features such as bill type, audit result, integrity, submission time and OCR quality factor into judgment, and automatically determine the return of single by combining historical data and business rules of financial management platform;System realizes the real-time synchronization of single machine equipment and financial management platform by temporary storage grid and capacity monitoring and dynamic storage strategy combined with RESTful API interface;The system can automatically optimize the bill processing flow, improve the accuracy of OCR identification and the efficiency of single determination, and ensure the safety of bill storage.

The application discloses a composite storage device and a fuel cell generated water utilization system. The composite storage device comprises a water storage unit and a dynamic storage unit. The water storage unit is internally provided with a water storage cavity used for storing generated water of a fuel cell system. The dynamic storage unit is arranged in the storage cavity and internally used for storing fuel. The dynamic storage unit can expand or shrink according to the increase or consumption of the internal fuel. The application can fully utilize the heat in the use process of the fuel cell system and the space gap in the internal water storage unit, and effectively improves the space utilization.

The dynamic memory array of a DRAM device operates using at least two voltages. A first voltage is the operating (i.e., switching) voltage of most of the digital logic circuitry of the DRAM device, which is used to power sense amplifiers during sense (i.e., read) operations and most other column operations (e.g., precharge, activate, write). A second voltage is greater than the operating (i.e., switching) voltage of most of the digital logic circuitry of the DRAM device, which determines the voltage written to the capacitor of a DRAM cell (i.e., the bit line voltage). The digital logic circuitry operates using a supply voltage, which is lower than the voltage written to the capacitor of the DRAM array. This allows lower voltage swing digital logic to be used for most of the logic on the DRAM device, while a larger voltage is written to the DRAM cell.

A dynamic memory system having multiple memory regions respectively storing multiple types of data. A controller coupled to the dynamic memory system via a communication channel and operatively to: monitor usage of a communication bandwidth of the communication channel; determine to reduce memory bandwidth penalty caused by refreshing the dynamic memory system; and in response, reduce a refresh rate of at least one of the memory regions based on a type of data stored in the respective memory region.