36 results about "Dynamic memory management" patented technology

A computer system implements a unified framework integrating an adaptive elastic funnel (AEF) with a convergent intelligence fabric (CIF) for multi-agent AI collaboration. The system provides a universal multi-modal key-value subsystem for sharing partial computations, implements hybrid placement strategies for dynamic memory management, and incorporates quantum-resistant secure enclaves. The architecture integrates hardware acceleration through GPU-FPGA hybrid caching and neuromorphic processors, applies adaptive energy and thermal management across hardware generations, and implements autonomous flash resource orchestration with multi-dimensional wear management. The system orchestrates tensor workflows using hierarchical scheduling, enables cross-agent collaboration with privacy preservation, and supports continuous learning without catastrophic forgetting. This integration delivers unprecedented computational efficiency and security in high-dimensional decision-making environments while supporting incremental adoption through modular interfaces.

A computer system implements a unified framework integrating an adaptive elastic funnel (AEF) with a convergent intelligence fabric (CIF) for flexible and contextualized multi-agent AI and human collaboration at scale. The system provides a universal multi-modal key-value subsystem for sharing partial computations across agents, implements a hybrid greedy / non-greedy placement strategy for dynamic memory management, orchestrates dynamic computational workflows and tensor workflows using hierarchical tensor-fragment scheduling, enables cross-agent orchestration with policy-based privacy preservation, and incorporates quantum-resistant secure memory enclaves. The architecture supports continuous learning without catastrophic forgetting, compositional reasoning across modalities, and secure task execution in distributed environments. This integration enables unprecedented computational efficiency, secure collaboration, and adaptive intelligence in high-dimensional decision-making environments while supporting incremental adoption through modular interfaces.

The invention discloses a large model training dynamic memory management method based on a multilayer hierarchical cache mechanism, and relates to the technical field of computer system architecture and storage. The method is managed on the basis of a dynamic heterogeneous memory unloading system, and comprises the following steps of: in an initial stage of a training task, constructing a user state access path, establishing a TensorID-LBA mapping table, and performing hierarchical caching and threshold presetting; in the training task running stage, the access popularity of each tensor is evaluated in real time; when occupation of the GPU video memory cache layer exceeds a threshold value or cold data is detected, a dynamic migration strategy is triggered, and the low-frequency tensor is unloaded to a CPU memory cache layer or an SSD storage layer; when the tensor access popularity is improved, automatically migrating back to a high-performance layer through a prediction mechanism; in the process, tensor self-adaptive migration is executed by a user mode I / O thread pool. According to the method, the high-performance I / O stack is constructed in the user mode, and the adaptive tensor unloading and hierarchical caching mechanism of the multi-layer heterogeneous memory is combined, so that the universal deployment of the system is realized.

The invention discloses an interactive response method, device and equipment for multi-modal sensing data and a storage medium, and relates to the technical field of intelligent interaction.The interactive response method for the multi-modal sensing data comprises the steps that the multi-modal sensing data is obtained, and a user mental evaluation result is constructed based on the multi-modal sensing data; identifying user identity information according to the multi-mode sensing data, and determining short-term memory information and long-term memory information according to the user identity information; when a demand instruction of a user is received, performing retrieval in the short-term memory information and the long-term memory information according to the demand instruction to obtain a retrieval result; and generating interaction response content according to the retrieval result and the user mental evaluation result, and performing interaction. Through multi-modal perception and dynamic memory management, long-term continuous personalized interaction is realized, and the user interaction experience is improved.

The application discloses a large language model offline reasoning task reasoning acceleration method and system under limited resources, a model construction module, a dynamic batch processing reorganization module, a dynamic memory management module and an unloading strategy adjustment module; based on the lexGen large language model offline reasoning system, dynamic batch processing reorganization is performed on the reasoning task, dynamic memory management design is combined, reasoning tasks are managed in an iterative level fine-grained manner, occupied memory of completed reasoning tasks is released in real time, and dynamic arrangement is performed on uncompleted reasoning tasks, so that the waste of reasoning resources in the reasoning process is reduced, and the offline reasoning system throughput based on the unloading technology is improved. The application efficiently utilizes the idle memory of hardware, improves the throughput of offline reasoning, reduces the influence of unloading on the reasoning process, and transfers part of tensors in a device with low bandwidth to the idle memory, so that the I / O overhead during transmission is reduced.

The invention discloses a method and a system for realizing a security acquisition software plug-in with dynamic characteristics for an electric power monitoring system, and aims to solve the problems of difficulty in function extension, insufficient security and poor resource isolation of a traditional electric power monitoring system. According to the system, through a plug-in architecture design, a plug-in function module is realized in a dynamic library form, and dynamic loading, unloading and security isolation operation are supported. The plug-in is started through the sub-process, the main program ensures the legality of the plug-in through SM2 certificate verification, and efficient interaction between the main program and the plug-in is achieved through a callback function mechanism. The plug-in supports dynamic memory allocation, and potential safety hazards caused by predefined field length are avoided. Core rights points of the method comprise plug-in expansion capability, sub-process isolation, SM2 certificate verification, a transparent transmission mechanism and dynamic memory management, and the function expansibility, the safety and the operation stability of the power monitoring system can be remarkably improved.

The invention relates to the technical field of memory management, in particular to a dynamic memory management method based on a bidirectional queue, electronic equipment and a medium. A first pointer and a second pointer are set to orderly control bidirectional insertion or deletion of queue elements, a queue main memory is started firstly, and when the storage requirement of the queue elements is not met, areas stored with the queue elements are reordered and then a queue auxiliary memory is started; when the queue main memory and the queue auxiliary memory are not enough to meet the storage requirements of the queue elements, updating the sizes of the queue main memory and the queue auxiliary memory to expand the queue memory, reordering discrete regions, in which the queue elements are stored, in the original queue memory, generating a continuous region, and writing the continuous region into a new queue main memory; the frequency of queue element copying and memory releasing in the dynamic change process of the bidirectional queue is reduced, and the memory utilization rate of the bidirectional queue is increased.

This invention relates to a dynamic memory management device and method for high-level synthesis. The device includes at least several search cache modules and several modification write-back modules. The search cache modules are connected to DRAM storage modules and BRAM cache modules, respectively. The modification write-back modules are also connected to DRAM storage modules and BRAM cache modules, respectively. The BRAM cache modules cache node information on the search path and temporarily store modification information for nodes. The search cache modules read node data from the DRAM storage modules and write it to the BRAM cache modules based on received operators and node addresses. The modification write-back modules read node data from the BRAM cache modules and write it back to the DRAM storage modules. Addressing the low execution efficiency of directly porting traditional operating systems to FPGAs, this invention leverages the large capacity of DRAM on FPGAs to achieve efficient dynamic memory allocation and reclamation, improving the ease of use and code reusability for high-level synthesis.

This application discloses a memory management method and electronic device, relating to the field of computer technology. The method includes configuring the memory of any controller in a target cluster based on its current memory management parameters, running preset test cases on the controller to obtain performance indicators under the current memory configuration, determining the fitness function value of the controller under the current memory configuration, and then determining the controller's historically optimal memory management parameters. If the current iteration count has not reached an iteration count threshold, the memory management parameters are updated based on the target update speed, and iteration continues based on the updated memory management parameters. If the current iteration count reaches the iteration count threshold, the controller is configured with appropriate memory based on its historically optimal memory management parameters. This achieves dynamic memory management of the controller, resulting in improved memory utilization.

The present application relates to a kind of coal mining operation multi-source time series data attention perception and dynamic memory method, belong to coal mine intelligent technology field.It includes: the three-dimensional feature dynamic extraction and scene information attention fusion of the multi-source time series data of coal mining face, generate scene multi-dimensional feature dataset;Multi-agent driven coal mining scene multi-dimensional dynamic information hybrid storage mode is used to store the multi-source time series data of coal mining face and scene multi-dimensional feature dataset, form data base;When obtaining coal mining scene task instruction, based on data base, in combination with equipment, process, environmental information, the related coal mining scene of it is mapped out, and the content indicated by the coal mining scene task instruction is completed by the cooperation of various types of Agent.The present application introduces Agent intelligent agent architecture and attention mechanism, in combination with dynamic memory management strategy, can provide core technical support for coal mining operation state real-time monitoring, fault intelligent diagnosis and production optimization decision.

The invention discloses a large model memory management and interaction method based on a dynamic cognitive map. The interaction method comprises the following steps: receiving information input by a user; performing dialogue generation, intention recognition and information extraction by adopting a large language model inference engine to obtain an information recognition result; and performing dynamic memory management on the information identification result. And through contradiction detection and an active verification mechanism of the reflection engine, continuous pollution of wrong memory is effectively avoided, so that the system memory becomes more and more reliable over time.

The invention relates to a coal mining operation multi-source time series data attention perception and dynamic memory method, and belongs to the technical field of coal mine intellectualization. Comprising the following steps: performing three-dimensional feature dynamic extraction and scene information attention fusion on multi-source time sequence data of a coal face to generate a scenarized multi-dimensional feature data set; the method comprises the following steps of: storing multi-source time sequence data and a scenarized multi-dimensional feature data set of a coal face by adopting a multi-agent driven coal mining scene multi-element dynamic information mixed storage mode to form a data base; after a coal mining scene task instruction is obtained, coal mining scenes related to the coal mining scene task instruction are mapped on the basis of a data base in combination with equipment, process and environment information, and contents indicated by the coal mining scene task instruction are completed through cooperation of various types of Agents. According to the method, an Agent agent architecture and an attention mechanism are introduced, and a dynamic memory management strategy is combined, so that core technical support can be provided for coal mining operation state real-time monitoring, fault intelligent diagnosis and production optimization decision making.

The embodiment of the invention discloses a memory management method and device for a secure container. The specific implementation mode of the method comprises the steps that a mounting point of the virtual file system is created according to preset mounting parameters, and the mounting parameters comprise an identifier used for indicating whether a large-page memory is started or not; creating a security container, and setting the virtual file system as a memory rear-end carrier of the security container through the virtual machine simulator; detecting mounting parameters through the virtual machine simulator to determine whether a large page memory is started or not; in response to determining to enable the large page memory, allocation of the large page memory is requested. According to the implementation mode, refined application of a transparent large page technology in a secure container environment is realized, and performance benefits close to static large pages are obtained while dynamic memory management advantages are kept.

The invention belongs to the technical field of computer software, and particularly relates to a collaborative review management system based on PDF (Portable Document Format) structured preview and precise annotation, which is characterized in that structural analysis is carried out on an uploaded PDF through a PDF JavaScript analysis engine, paragraph structure information is extracted, and hierarchical and paging quick preview is realized; providing a text-level precise annotation tool to firmly bind approval and original text character positions and perform reverse highlight positioning; a Web-based real-time cooperation platform is constructed, multi-user synchronous review is supported, and full-life-cycle management including states and persons in charge is carried out on review opinions; and a large PDF file is efficiently processed by adopting a paging loading and dynamic memory management strategy. According to the method, the problems of difficult cooperation, rough annotation, poor traceability, large file performance bottleneck and the like in a traditional review mode are effectively solved, and the efficiency and quality of scenes such as technical review and contract auditing are remarkably improved.

The invention relates to an agent memory compression method based on a large-model attention mechanism, and the method comprises the following steps: obtaining the attention weight of each memory in a historical memory sequence for a current conversation, and calculating the correlation degree of the attention weight; determining a demarcation point for distinguishing the early content from the recent content from the historical memory sequence based on the correlation degree and the system constraint condition; dividing the historical memory into an early memory section and a recent memory section by the demarcation point C; performing semantic extraction on the early memory segment to generate summary information, and replacing the original early memory segment with the summary information; and combining the abstracted early memory segment with the complete recent memory segment to form compressed memory input required by the next round of model reasoning. According to the dynamic memory management method based on early-stage content abstract extraction and recent content complete retention, the demarcation point of the early-stage content and the recent content is automatically determined through an attention weight calculation mode, and controllable compression and effective fidelity of the memory content are achieved.

The invention provides a dynamic hierarchical semantic memory method. The dynamic hierarchical semantic memory method comprises a hierarchical encoder, a dynamic memory management and access mechanism and context-based semantic resolution and fusion. A multi-level and structured dynamic semantic memory system can be constructed, surface sequence information of a text can be captured at the same time, the method is used for local coherence and deep semantic architecture and global understanding, and the most relevant part can be dynamically activated according to a current task.

The description relates to dynamic memory management. One example includes an assembly that entails processing elements and memory. A dynamic UMA / NUMA configuration module is configured to facilitate managing a first region of the memory based upon a Uniform Memory Access (UMA) architecture and a second region of the memory based upon a Non-Uniform Memory Access (NUMA) architecture. The dynamic UMA / NUMA configuration module is configured to dynamically adjust ratios of the memory in the first region and the second region based upon workload changes on the processing elements.

A memory controller is configured to receive segment pointers that identify physical memory devices and slices for associated segments. The memory controller is configured to identify a most-full first one of the physical memory devices, to identify a most-empty second one of the physical memory devices, and to identify one of the segments in the most-full first one of the physical memory devices. The memory controller is configured to move a slice from the identified one of the segments in the most-full first one of the physical memory devices to the most-empty second one of the physical memory devices. The memory controller is configured to update one of the segment pointers for the identified segment. The memory controller is configured to update one of the segment pointers for the identified segment by removing a previous segment pointer to the most-full first one of the physical memory devices.

A dynamic memory management circuit includes interfaces configured to receive data from a memory subsystem, the data describing memory allocation requests from accelerator processing units during a first time period; and one or more processors coupled to the interfaces and configured to: determine memory space to be allocated to a CPU and the accelerator processing units in accordance with a memory allocation policy. The policy includes: allocating a minimum memory space required for an operating system executed by the CPU; allocating a minimum memory space required by the CPU to fulfill a predefined quality of service requirement; one or more rules to amend memory space to be allocated to the CPU and the accelerator processing units based on the data. The processors are further configured to send an instruction to allocate memory for the CPU and the accelerator processing units in accordance with the determined memory space to be allocated.

The invention provides a multi-shot consistency video generation method based on dynamic memory management, and the method comprises the steps: carrying out the coding and feature extraction of a story script to obtain semantic information, and mapping the semantic information into a structured shot description; entity categories and attribute vectors are extracted according to the shot description; retrieving a matched target entity reference image in a dynamic memory library according to the entity category and the attribute vector; when the matched entity reference image does not exist in the dynamic memory bank, generating a new entity reference image according to the attribute vector and the historical image, and writing the new entity reference image into the dynamic memory bank; generating key frames according to the shot description and the target entity reference image, and synthesizing a video according to a plurality of key frames; the performance of role consistency, prop consistency and background consistency in long video generation can be remarkably improved, and the method has high robustness and expandability.

A computer system implements a unified framework integrating an adaptive elastic funnel (AEF) with a convergent intelligence fabric (CIF) for multi-agent AI collaboration. The system provides a multi-layer key-value subsystem for sharing computations across GPU partitions, applying hybrid placement strategies for dynamic memory management. It unifies physical GPU sub-allocation and virtual GPU time-slicing through a common abstraction layer while maintaining isolation through policy-based multi-tenancy. A predictive resource orchestration system forecasts computational needs, while speculative data scheduling proactively manages memory access patterns. The architecture includes risk-based scheduling for uncertain workloads, test-time compute scaling, and federated learning for edge collaboration. This framework delivers computational efficiency, security, and adaptive intelligence in high-dimensional environments while supporting incremental adoption through modular interfaces.

The application relates to the technical field of memory management, in particular to a dynamic memory management method based on a one-way queue, electronic equipment and a medium. The application stores queue elements in the mode of setting a queue main memory and a queue auxiliary memory, first activates the queue main memory, then activates the queue auxiliary memory, and when the queue main memory and the queue auxiliary memory cannot meet the storage requirements of the queue elements, the size of the queue main memory and the queue auxiliary memory is updated to expand the capacity, the frequency of copying and releasing the memory of the queue elements in the dynamic change process of the one-way queue is reduced, and the queue memory utilization rate of the one-way queue is improved.

The invention relates to a database security migration method and device based on memory protection, and aims to solve the problems of memory overflow (OOM) and migration interruption caused by rate mismatching under a read-write thread separation architecture. The method comprises the steps that the memory occupation amount of data read by a read thread is counted in real time and accumulated to a cache accumulated value, and the value is deducted to dynamically update the memory occupation after write-in by a write thread; when the cache accumulated value exceeds a preset backlog threshold value, pausing a read thread and scheduling the read thread to assist a write task; and continuously checking the accumulated value in the assistance mode, and recovering the original task of the read thread if the accumulated value reaches the standard. According to the method, through dynamic memory management and control and thread scheduling, excessive accumulation of the memory is accurately inhibited, and OOM faults are avoided; the efficient migration advantage of read-write separation is reserved, migration stability is guaranteed through closed-loop regulation and control, collaborative improvement of efficiency, stability and data safety is achieved, and the method is suitable for high-concurrency and large-data-volume migration scenes of the database.

The invention relates to the technical field of memory management, in particular to a dynamic memory management method based on a one-way queue, electronic equipment and a medium. Queue elements are stored by setting a queue main memory and a queue auxiliary memory, the queue main memory is started firstly, and then the queue auxiliary memory is started; and when the queue main memory and the queue auxiliary memory are not enough to meet the queue element storage requirement, the sizes of the queue main memory and the queue auxiliary memory are updated to expand the capacity, so that the frequency of copying the queue elements and releasing the memories in the dynamic change process of the one-way queue is reduced, and the queue memory utilization rate of the one-way queue is improved.

A data platform is provided that collects memory usage data for a function during execution of the function and determines memory constraints of the function during execution using a configurable statistical method and the memory usage data. The data platform dynamically adjusts a memory limit for the function based on the memory constraints and schedules a job including the function using the memory limit without user intervention.

The application provides a multi-parameter synchronization hidden method and system in a memory address, comprising: creating a number-parameter mapping table and a parameter seed mapping table; in response to a parameter generation instruction, generating a split position based on a timestamp seed of a current parameter in the parameter seed mapping table, determining a split function; splitting a parameter original value, assigning a split number value to the split position, and updating the number-parameter mapping table; in response to a parameter modification instruction, finding all split positions of the current parameter, and preferentially selecting a split position corresponding to a parameter set with multiple parameters as an associated split position; extracting an associated parameter set from the number-parameter mapping table, triggering a parameter generation instruction of each parameter in the set, repeatedly executing the split position generation and number value assignment processes, and synchronously updating the number-parameter mapping table, so that the cracking difficulty is changed from linear to nonlinear by sharing the split position by multiple parameters, and the data anonymity and running efficiency are considered in a high-frequency scene by combining dynamic memory management.

The invention relates to an unmanned aerial vehicle route planning method and device, a medium and a program product. The method comprises the following steps: acquiring spatio-temporal data of multiple unmanned aerial vehicles; the spatio-temporal data are stored in blocks according to a time window and a space area at a GPU end, dynamic memory management is carried out based on the flight height and time, and the data are operated by adopting a parallel strategy combining multi-thread collaborative read-write and asynchronous read-write; and decomposing the route planning task into parallel subtasks at the GPU end, and generating a route planning result based on a flight height range matching planning algorithm. According to the invention, route planning corresponding to the plurality of unmanned aerial vehicles is carried out at the GPU end based on the spatio-temporal data of the unmanned aerial vehicles, so that collaborative operation of the plurality of unmanned aerial vehicles is realized, and the task execution efficiency and effect are improved; and a GPU parallel read-write strategy is adopted, so that the processing efficiency of the spatio-temporal data of the unmanned aerial vehicle is improved, and parallel read-write of the spatio-temporal data of the unmanned aerial vehicle is realized.

A neural network dataflow graph having a set of nodes and a set of edges is processed. An insertion point for a memory reduction or memory restoration operation is determined. The determination is based on computing tensor timing slack (TTS) for a set of input tensors; using input tensors having respective TTS values greater than a threshold (thTTS), editing a candidate list (SI) of input tensors in the set of input tensors; filtering the SI to retain input tensors whose size satisfies a threshold (thS); and using the filtered SI to determine the insertion point for the operation. A new dataflow graph is generated or an existing dataflow graph is modified using the process.

The invention belongs to the technical field of electronic information, and discloses an inter-core communication method and system based on hardware management and electronic equipment, and the method comprises the steps that a first processor core inquires a communication state flag bit of a second processor core; if the communication state is available, dynamically applying for a memory block from the shared memory; writing to-be-sent data into the memory block, and generating a data head node; the first processor core adds the data head node to a first data chain table maintained by the first processor core; the hardware management module monitors the first data chain table, and automatically sends an interrupt signal to the second processor core when monitoring that data head nodes are increased and the second processor core can communicate; and the second processor core responds to the interrupt signal, obtains the target data head node from the second data chain table, and reads the data in the shared memory according to the information contained in the target data head node. By introducing a special hardware management module and adopting a dynamic memory management mechanism, the utilization rate of the shared memory is greatly improved.