185 results about "Toolchain" patented technology

The invention provides a dynamic tool integration method and system for an intention-driven AI dialogue system, and the method comprises the steps: S5, extracting resource occupation information from a transaction log according to a state persistence record, carrying out the resource release and temporary context cleaning of each heterogeneous component, and obtaining a released resource list; s6, searching a breakpoint state snapshot and a transaction log from the distributed database according to the released resource list, and reconstructing a tool chain execution flow historical state to obtain a fault context analysis result; and S8, according to the context recovery state and the semantic feature vector, updating a tool chain execution stream, and reallocating computing resources through a heterogeneous component coordination mechanism to obtain a task continuity confirmation result. According to the method, the recovery efficiency and the resource utilization rate of the tool chain in an abnormal scene are remarkably improved, and the task continuity and the system robustness in a high-concurrency environment are guaranteed.

A simulation system based on a large language model comprises a large language model module, a scenario generation tool module, a meta-scene database module, a simulation deduction engine module, a simulation model module, a simulation front-end module and a simulation tool set module. The large language model module analyzes an instruction and a situation description input by a user, splits semantic information and meta-scene information in the instruction and the situation description, sends the split semantic information and meta-scene information to the scenario generation tool module, and receives a scenario file output by the scenario generation tool module; further sending the received scenario file to a simulation deduction engine module, and receiving a simulation process and a simulation result generated by the simulation deduction engine module; and meanwhile, the large language model module can call the simulation tool set module by using a tool chain instruction through a simulation deduction engine, so that the simulation tool set module generates a situation processing result. According to the method, man-machine interaction based on the large language model is realized, and a tool set is called by utilizing the large language model to assist a user in completing simulation and analysis.

The invention discloses a large model calling system and method based on an MCP tool chain, and the method comprises the steps: receiving user input, and carrying out the preliminary matching of tools: querying whether the user input contains an entity in an entity database or not through a fuzzy matching technology, and carrying out the preliminary screening of a plurality of candidate tools based on a matching result, so as to obtain limited to-be-selected tools; user input is converted into semantic vectors, semantic matching is conducted in a vector database, and semantic similarities of all tools to be selected are ranked through a similarity calculation method; extracting key information in user input; abstracting the user input into a specified format; and matching the specified format with the historical workflow template, and if the historical workflow template is matched, replacing the key information to a corresponding position of the historical workflow template by adopting a structure-preserving parameter replacement algorithm to realize workflow reuse. According to the method and the device, a complete technical link from accurate tool screening to intelligent workflow multiplexing can be realized.

The invention discloses a building construction carbon emission dynamic optimization method and system based on Revit and AI linkage. The method comprises the following steps: step 1, data integration; step 2, model training; step 3, logistics simulation; step 4, collaborative optimization; 5, dynamic early warning is carried out; step 6, generating a thermodynamic diagram; step 7, integrating paths; 8, iterative optimization is carried out; and 9, multi-scene analysis is carried out. According to the method, real-time variables are dynamically captured through an Azure AutoML time sequence model, and the prediction error is reduced; automatic management is realized by integrating a code-free tool chain, and the cost is reduced; double objective functions are innovated to synchronously optimize transportation and environmental protection benefits; a full-life-cycle closed-loop system is constructed to improve the dynamic adjustment capability; and the generation of visual decision support efficiency is improved, and revolutionary technical support is provided for green construction.

The invention discloses a lightweight cross-platform PLC (Programmable Logic Controller) multi-language code generation method, which relates to the field of industrial process control, and firstly provides a unified intermediate data model, and on the basis, a module for converting each PLC language into the unified intermediate data model is realized; on the basis of a standardized JSON data structure, a modularized unified intermediate language conversion chain is further supported, an analyzer and a generator communicate through a JSON interface, a cross-platform conversion path is supported, and unified intermediate language conversion from PLC to JSON to XML to ST to C is completed; the method comprises the following steps of: integrating an open source compiler and runtime, constructing an open and pluggable automatic execution framework, compiling a generated C code into a platform executable file by using an adaptive GCC tool chain aiming at different target platforms, and deploying a program generated by compiling to a soft PLC runtime system supporting multiple platforms. According to the method, a JSON intermediate data structure which is light in weight, unrelated to languages and uniform in structure and a matched conversion framework of the JSON intermediate data structure are constructed.

The invention provides a script language-oriented seismic batch processing job editing method and device, and the method comprises the steps: obtaining a to-be-processed seismic job demand, and determining a seismic job project and database information; initializing each module object in a pre-established seismic batch processing operation framework, wherein the framework is a scripting language framework comprising a preset core library, a module library and a tool chain; determining a standard type parameter corresponding to each module object; initializing an operation topological structure of the seismic batch processing operation framework; selecting a target module object and adding the target module object to the target operation based on the earthquake operation project and the database information; and outputting the topological graph of the module according to the debugging requirement, and generating a target operation corresponding to the seismic operation requirement by calling the XML format file. According to the method, seamless connection of job editing and code editing is realized through a script language-oriented seismic batch processing job framework, so that the method is more flexible and efficient; and diversified requirements of a user for job editing by using a script programming language are met.

A method and a system for compiling a neural network, a computer storage medium, and a compilation device are provided. The method for compiling the neural network comprises: translating a network file into an intermediate representation file; optimizing the intermediate representation file to obtain an optimized intermediate representation file, based on a performance analysis, single-node optimization, and collaborated optimization; generating a network template file based on hardware interfaces through the optimized intermediate representation file; compiling the network template file into an executable inference application. The present disclosure aims to design and implement an automated compilation toolchain framework. This framework adjusts parameters, generates code, creates intermediate representations (IRs), and applies optimization algorithms based on software and hardware information. When this compilation toolchain operates on a target chip, it ensures consistent network output results, achieves higher computation rates within shorter optimization times, reduces computation delays, and facilitates user debugging and tuning.

The invention provides a method for searching structured data based on natural language driven by a large model, which comprises the following steps of: preprocessing interface parameters through a data interface of an openAPI (Application Program Interface) to generate an interface file, and adding the preprocessed interface parameters into an agent tool chain; the large model extracts keyword parameters from a question input by a user through a natural language and performs semantic analysis on the extracted keyword parameters, so that the keyword parameters are converted into interface fields or table fields in an interface file; the large model adapts an analyzed result to a tool name in the agent tool chain; calling an adaptive interface name by the large model, and calling a corresponding tool to query a data result; the large model outputs the result according to the form of the user demand, standardization processing is performed on the interface of the traditional data, the agent tool is added, the large model is used, different data search requirements are met, and the personalized data query requirement is met.

The invention discloses a smart home control method and system based on a multi-modal large model, and the method comprises the steps: carrying out the cross-modal feature extraction and semantic alignment processing of voice, image and sensor data through an improved multi-modal large model, and achieving the semantic space mapping of multi-source data based on a context sharing mechanism and reasoning logic; the method comprises the following steps: abstracting local smart home equipment into a standardized resource entity by using an MCP technology, defining an equipment function feature set through a protocol interface, and packaging the equipment function feature set into a callable tool chain; constructing a real-time data fusion model, and generating a control instruction sequence through the multi-modal large model based on the real-time data fusion model; and calling a corresponding tool in the tool chain according to the control instruction sequence to execute equipment control. The problems that an existing system depends on single-mode data and is easily interfered by the environment, the access cost of new equipment is high due to the fact that a unified data fusion framework does not exist, a control strategy is difficult to adapt to user behavior changes, equipment communication protocols are different, and the user requirement matching degree is low are solved.

The invention relates to a cross-platform software development method and related equipment. The method is applied to the electronic equipment providing a low-code tool chain, and the low-code tool chain comprises a low-code development platform and a compiler. Based on the configuration of the meta model of the vehicle-mounted resource, the low-code development platform generates an automobile domain specific language VDSL file for describing the vehicle-mounted software function; the compiler obtains the VDSL file, compiles the VDSL file into running program codes running in different controllers and sends the running program codes to a cloud server, and the cloud server is used for sending the running program codes to the different controllers. According to the method and the device, the running program codes corresponding to different controllers are generated through the unified VDSL, so that cross-platform running software programs can be conveniently developed in the same platform, the software development efficiency is improved, and the development cost is reduced.

The invention discloses an agent and tool chain collaborative data processing method and device and a medium, and relates to the technical field of agent application. The method comprises the steps of receiving a data processing task target and related configuration parameters input by a user, and initializing task context data based on the task target and the configuration parameters; executing a circular processing process by an agent based on a ReAct framework, and dynamically adjusting a task strategy by iteratively executing a thinking stage, an action stage and an observation stage until a task completion condition is met; in the circular processing process, based on standardized tool calling instruction data generated by the intelligent agent, cooperatively calling the MCP tool chain and a plurality of tools in the external tool adaptation layer to execute a data processing task; performing aggregation processing and data verification on the multi-source result data returned by the tool execution to generate integrated analysis result data; and displaying task execution process data and a final analysis result in real time through a front-end interaction interface, and dynamically adjusting a task execution process based on user intervention instruction data.

The invention relates to a container application migration method and system and a storage medium, and the method comprises the steps: extracting a dependency list and call information through a script, compiling a source code of a current architecture of a container application into an executable file under a target architecture, replacing a dependency library file under the current architecture of the container application with the dependency library file under the target architecture, and carrying out the migration of the container application. Based on the dependency list and the calling information, using the basic mirror image of the target architecture to reconstruct the container mirror image so as to ensure that the container application can run on the target architecture, intercepting and redirecting the system calling through an LDPRELOAD mechanism so as to enable the system calling to adapt to the target architecture, and using a cross compilation tool chain so as to improve the system calling efficiency. And replacing the instruction set of the current architecture with the instruction set of the target architecture, and carrying out container application migration operation. The dependency list and the calling information are extracted through the script, so that the possibility of omitting the dependency list and the calling information required by operation of the container application is reduced, and efficient and stable operation of the container application on the target architecture is ensured.

The invention discloses an AI dialogue flow control method, system and device supporting real-time interruption of a user, a medium and a program product. The method comprises the steps that a user request is received, and a flow session is established; based on the stream session, requesting LLM increment generation, receiving fragments generated by increment, returning the fragments, and pushing the fragments to the client; when a click stop request of a user is received, an STOP signal is sent; when the STOP signal is monitored, sending a mark termination request; after a mark termination request is received, a subsequent generation termination request is sent to the LLM, a STREAMTERMINATED semantic event is published and broadcasted, and generated content and a termination mark are returned to the client; terminating subsequent generation comprises stopping untriggered subsequent reasoning and tool chain calling; discarding the tail cache which is not sent; and recording the output token count, and terminating the offset. The AI dialogue flow control method supports real-time interruption of the user, is low in delay, clear in semantics and recoverable, and further has the advantages of saving resources, being high in stability, capable of being universally expanded, gaining behavior analysis and the like.

The invention relates to the technical field of artificial intelligence, in particular to a brain-equipped robot end-side cloud collaborative management system, which comprises a brain-equipped robot unit, a man-machine interaction client, an APP / applet, a resource management platform and an application market platform, the brain-equipped robot unit comprises a multi-mode sensing module which is used for understanding natural language instructions, screen contents and voice interaction; a large model decision module; the execution control module is used for simulating a keyboard and a mouse to operate computer software; the man-machine interaction client and the APP / applet serve as an interface for interaction; the resource management platform comprises equipment and a user management module; the resource matching algorithm module is used for evaluating and distributing optimal computing resources; the block chain registration module is used for recording ownership and use right information of the task; the sandbox execution environment module is used for creating a running environment; the application market platform is used for providing a developer tool chain and a revenue distribution mechanism. The application efficiency of the agent system can be improved.

The invention provides a macro-cell layout evaluation optimization method based on a large language model, and belongs to the field of electronic design automation. According to the method, a large language model is introduced to assist layout generation, a current initial layout is improved in combination with connection weight analysis, on the other hand, an improved result is judged and continuously iterated according to natural language feedback of an experience manual or an expert, and closed-loop optimization is formed; carrying out weight quantization on direct and indirect signal paths among macro units, constructing a logic association strength matrix, and realizing accurate identification and priority ranking of key time sequence paths; the availability score of each candidate layout scheme is calculated through a layout availability scoring module, dynamic measurement is carried out on surrounding standard unit resources, the resource distribution state is evaluated in real time, the position of a macro unit is actively adjusted, and resource congestion or waste is avoided; the method is deeply integrated with an open source EDA tool, the dependence on a commercial tool is reduced, the tool chain deployment and iteration period is shortened, and the transportability and usability of the method are guaranteed.

A system and method for a protocol parser that has native support for parsing TLVs and flag-fields and allow users to code in a language convenient to them and leverage standard tool chains and tools. The same parser source can be compiled to arbitrary software and hardware targets without code change and provide the highest performance possible given its capabilities.

A system and method for designing and optimizing a hardware accelerator of a neural network is provided. In a pre-design phase, rules are extracted from compilation features that describe a transformational relationship between a neural network operator, a coarse-grained operator, and a fine-grained data stream. A fast mapper and a data flow mapper are generated for converting the neural network model into coarse-grained operator descriptions. In the coarse-grained design stage, a plurality of preliminary hardware accelerator designs are generated using an architecture optimizer. The coarseness operator descriptions are simulated using a coarseness simulator to obtain performance metrics for each preliminary accelerator design. In the fine-grained design stage, a data stream mapper and a fine-grained simulator are adopted to complete preliminary hardware accelerator design. A hardware accelerator is generated from the finally determined hardware accelerator design and a corresponding software tool chain is created, including a compiler and a software development kit for programming, debugging and deploying the hardware accelerator design.

The invention relates to the technical field of embedded system software development and continuous integration, in particular to a docker-based AGV (automatic guided vehicle) cross compiling method, which comprises the following steps of: constructing an ARM (advanced RISC machines) root file system on an x86 development host by using a virtualization tool, configuring a header file and a dependency library, and packaging and compressing to obtain an ARM root file system mirror image; pulling the basic mirror image from the public warehouse to create an x86 Docker container, copying the ARM mirror image to a specific path of the container and mounting the ARM mirror image to prepare a Docker mirror image special for AGV cross compiling; writing a CMake cross tool chain file in the container, introducing an ARM mounting path, setting a dependency path and appointing a tool chain; then integrating the commands, automatically updating the container through mirror image ID comparison, and constructing operation commands such as Makefile packaging and compiling and the like at the same time; and finally, calling an integration command to start cross compiling, and generating a program capable of running on an ARM board card of the AGV. According to the method, unified and efficient multiplexing of the AGV cross compiling environment can be realized, the development efficiency is improved, the maintenance cost is reduced, and the method is adaptive to an AGV project needing cross compiling.

The invention discloses a resource optimization method and system based on demand driving, belongs to the technical field of software product development, and aims to solve the technical problems that resource allocation in software product life cycle management depends on artificial experience, is low in efficiency, subjective in allocation standard and difficult to dynamically adapt to project changes. Comprising the following steps: acquiring demand characteristics, and endowing the demand characteristics with weights; the method comprises the following steps: modeling available resources of a project to form a resource constraint model, defining attributes and constraint conditions of the resources through the resource constraint model, and collecting state information of the resources in real time; on the basis of the weights of the demand features, the state information of the resources and the attributes and constraint conditions of the resources, the resources are allocated through a multi-objective optimization algorithm, and an optimal resource allocation scheme is formed; issuing the resource allocation scheme to an execution end through tool chain integration; and optimizing the weight of the demand feature and the optimal resource allocation scheme based on a machine learning method.

The invention provides a collaborative management system and method for an EDA all-in-one machine, the collaborative management system comprises a business layer, an application layer, a platform layer and a hardware layer, the hardware layer is used for providing hardware resources and deploying container instances; the platform layer is used for performing instance management, resource quota allocation and instance life cycle control of container instances on the hardware resources provided by the hardware layer; the application layer is used for running the EDA tool of the corresponding version in the container instance; and the business layer is used for carrying out resource planning, template configuration, task scheduling and result management on the chip design whole process of the EDA all-in-one machine. According to the scheme, the whole process of chip design is integrated through the business layer, scattered project information, resource requirements and result management are incorporated into a unified framework, the tool chain information barrier is broken, and the manual coordination cost is reduced. And meanwhile, the predictability and traceability of task execution are improved, and an efficient and ordered collaborative management scheme is provided for the EDA all-in-one machine.

A system and method is provided for designing and optimizing hardware accelerators for neural networks. During a pre-design phase, rules are extracted from compilation patterns that describe conversion between neural network operators, coarse-grained operators, and fine-grained dataflow. A fast mapper for converting neural network models to coarse-grained operator descriptions and a dataflow mapper are generated. A coarse-grained design phase employs an architecture optimizer to generate plural provisional hardware accelerator designs The coarse-grained operator descriptions are simulated using a coarse-grained simulator to obtain performance metrics of each provisional accelerator design. A fine-grained design phase employs a dataflow mapper and fine-grained simulator to finalize provisional hardware accelerator designs. A hardware accelerator is generated from a finalized hardware accelerator design and a corresponding software toolchain is created including a compiler and software development kit (SDK) for programming, debugging, and deploying the hardware accelerator design.

This invention discloses a method and system for constructing an AI-native operating system based on ecosystem collaboration, belonging to the field of operating system technology. The invention first acquires and decomposes target requirements to form a set of functional and non-functional requirements, constraints, and priority lists. Based on this, it generates an operating system architecture description including kernel configuration and subsystem division. Then, it generates and completes multi-dimensional checks on module code and configuration files, and compiles, links, and packages them through a toolchain to generate an operating system image. This invention employs a three-layer AI model to realize requirement reasoning, system construction, and test optimization, forming a complete automated closed loop from requirement input to runnable image output. Simultaneously, it establishes a unified ecosystem collaboration mechanism to support multi-vendor collaborative development. This invention effectively reduces errors and repetitive work caused by manual intervention, shortens the development cycle, improves system stability and adaptability, and promotes the collaborative evolution and improvement of the AI-native operating system ecosystem.

The invention relates to the technical field of national cryptographic algorithm acceleration, and discloses a national cryptographic algorithm acceleration system based on a Numba just-in-time compiling technology. The numerical calculation mode reconstruction module is used for performing numerical calculation mode reconstruction on SM2 elliptic curve scalar multiplication, SM3 message extension round function, SM4 nonlinear transformation and ZUC flow generation logic contained in the national cryptographic algorithm core calculation module by utilizing an LLVM compiling chain of Numba, and converting an interpretively executed Python code into an optimized machine code adaptive to hardware; key function compiling acceleration is achieved through a (at) njit decorator, hot spot operation is dynamically recognized through compiling scheduling, and a differential instruction optimization strategy is loaded; the platform shows multiple technical advantages: on the development efficiency level, by presetting a standardized acceleration module library and an automatic compiling tool chain, the integration complexity of a national cryptographic algorithm is greatly reduced; in the aspect of operation performance, the compilation optimization depth is better than that of a general interpreter acceleration scheme; in the aspect of safety controllability, the mathematical theory basis of the national cryptographic algorithm is completely reserved, and hidden risks possibly introduced by black box type hardware acceleration are avoided.

The application discloses a circuit input sensitivity analysis method based on Undef SAT, which comprises five steps of netlist analysis and input division, two groups of CNF clause construction, S set assignment and initial solving, first group of solution increment SAT processing and termination and result output. The method uses the incremental SAT processing mode to gradually determine the final assignment of the input variables of the L set, effectively avoids full space search, significantly reduces the search space, and improves the analysis and determination efficiency. The method constructs two groups of CNF clauses supporting irrelevant item propagation, directly processes the irrelevant input variables of the S set by using the Undef SAT solver, reduces the number of logic propagation, and thus reduces the operation time. The integrated process of the circuit input sensitivity analysis formed by the method is suitable for mainstream EDA tool chains, has good engineering applicability, and is suitable for different scenes such as logic synthesis stage, formal verification, functional equivalence checking and loop oscillation detection.

The application provides a script language-oriented seismic batch processing job editing method and device, including obtaining seismic job requirements to be processed, determining seismic job items and database information; initializing each module object in a previously established seismic batch processing job framework, the framework being a script language framework including a preset core library, a module library and a tool chain; determining standard type parameters corresponding to each module object; initializing a job topology structure of the seismic batch processing job framework; based on the seismic job items and the database information, selecting target module objects and adding them to target jobs; outputting a topology graph of a module according to debugging needs, and generating target jobs corresponding to seismic job requirements by calling an XML format file. The application edits a seismic batch processing job framework by facing a script language, realizes seamless connection between job editing and code editing, is more flexible and efficient, and meets diversified needs of users in editing jobs by using a script programming language.

The invention provides an automatic testing method and system for tool chain database file analysis and mapping, and belongs to the technical field of automatic testing. The method comprises the steps that a source file is obtained, and first configuration data is obtained through analysis; obtaining a vehicle control operation system configuration file generated after the tool chain performs analysis and mapping processing on the source file, and performing analysis to obtain second configuration data; calling a mapping rule according to a to-be-tested configuration parameter object, and performing mapping processing on the first configuration data to obtain target configuration data; and performing automatic data comparison on the second configuration data and the target configuration data to generate a test report. The method has the beneficial effects that special verification is performed on the correctness of the configuration file of the vehicle control operation system mapped by the source file through tool chain analysis in the design and configuration stage, so that the configuration error of the tool chain caused by the file analysis and mapping error is found in time, the error is prevented from being left in the downstream code generation, integration and function test stage, and the operation efficiency is improved. And the later-stage troubleshooting, debugging and repairing cost is reduced.

The invention provides a real-time kernel-oriented program real-time compiling method and device and a storage medium. The method comprises the following steps: constructing and running an operating system containing real-time kernel capability on a target hardware platform, and determining a real-time interface library and a real-time equipment driving model of the operating system as dependence of compiling and linking; real-time transformation is carried out on the high-level language compiler and the runtime key library to generate a transformed tool chain, and the real-time transformation comprises the steps that a real-time interface library is introduced into construction and link configuration of the runtime key library, an interface and a data structure of a real-time equipment driving model are included in a visible range of the runtime key library, and the real-time interface library and the real-time equipment driving model are constructed. Real-time constraints are set for a concurrent scheduling and timer mechanism; and compiling an application program source code by utilizing the transformed tool chain, and introducing a real-time interface library in a linking stage to obtain a real-time application program running on an operating system. According to the method, real-time compilation without change of the source code, uniform interface adaptation, cross-architecture multiplexing and deterministic operation can be realized.

The invention discloses a visual system batch point binding tool chain and method based on coordinate transformation. The tool chain comprises a coordinate transformation tool and a point binding tool. The coordinate conversion tool is used for acquiring coordinate data and equipment information of each piece of equipment, converting the coordinate data into space coordinates and then generating a JSON configuration file; and the binding tool is used for binding the equipment information and the space coordinates of each piece of equipment in the three-dimensional model, and previewing the binding result of each piece of equipment in a three-dimensional visual form. According to the visual system batch point binding tool chain and method, the working efficiency can be greatly improved, the manual operation time and labor intensity are reduced, the situations of manual wrong binding, manual missing binding and the like are eliminated, the correctness of the binding result depends on an original drawing, and the consistency can be effectively guaranteed.

This invention relates to the field of vehicle communication technology, providing a method and system for converting Excel to ARXML format. The method pre-defines an Excel conversion template, specifying that different worksheets correspond to different top-level containers in the ARXML document, and pre-defined mapping rules between Excel column names and ARXML element attributes or child nodes. Based on these mapping rules, the method automatically converts data rows in the Excel worksheet into corresponding ARXML element nodes, and performs reference relationship parsing on the generated nodes, replacing reference path placeholders with reference nodes conforming to the AUTOSAR standard. Finally, it integrates all ARXML element nodes that have undergone reference parsing and updates, generating a compliant ARXML file that can be directly used in the AUTOSAR toolchain. This invention transforms complex manual coding into parameterized table configuration, effectively improving the efficiency, consistency, and compliance of vehicle communication configuration file creation through end-to-end rule mapping and automated processing.

The application discloses a kind of UVM verification platform automatic generation system and method based on generative artificial intelligence, the system includes generative AI dynamic anchoring module, LLM-VCS self-feedback verification module and conflict detection correction module;The generative AI dynamic anchoring module is used to solve the context window limit problem on large language model, by the key component in UVM platform is marked as anchor file and generates subsequent file based on dependency relationship priority;LLM-VCS self-feedback verification module cooperates with large language model and EDA tool chain, forms closed loop system from specification input to UVM platform generation, simulation and debugging;The conflict detection correction module is used to detect and correct the interface inconsistency and logic error between generated files.The application can adapt to the verification needs of different RTL designs, effectively solve the interface consistency problem of high coupling between UVM modules, flexibly generate UVM platform of different structures, significantly improve the verification efficiency and quality.