222results about "Compiler construction" patented technology

The invention discloses a structured text code generation and review method, which comprises the following steps of: receiving a control task demand input by a user, supporting natural language, structured form or graphical process input, performing deep semantic analysis by utilizing a demand analysis module, identifying control conditions, control actions and constraint factors, and generating and reviewing a structured text code. The task is converted into a plurality of control units through the SCoT chain; the SCoT chain reader generates a semantic vector according to a task demand, the AST reader generates AST representation through a grammar rule structure and a hierarchical relationship, and the decoder predicts a next grammar rule and generates an ST code in combination with the SCoT chain and AST features; and verifying grammar and logic of the ST code through an interactive automatic review and repair mechanism, if compiling errors occur, analyzing errors by the teacher model and providing repair suggestions, and performing code repair by the student model according to guidance until the code conforms to an expected function and passes compiling.

The invention relates to the technical field of artificial intelligence compilers, in particular to an AI compiler and compiling method based on a multi-level intermediate representation framework, and the compiler comprises a high-level semantic retention layer which converts models of different AI frameworks into Lalg-on-Tensor IR intermediate representations; the hardware perception optimization layer comprises a tensor packaging and propagation module which is used for performing block packaging, layout propagation and folding of redundant packaging / unpackaging operation on the input tensor; the dynamic partitioning module is used for automatically selecting the partitioning size based on the cache capacity and the core number of the target hardware; the microkernel fusion module is used for fusing matrix multiplication, bias addition and an activation function into a single composite operator; and the microkernel collaboration layer is in butt joint with the hardware acceleration library through the XSMM dialect to generate a target hardware code. The hardware perception optimization layer can perform optimization according to different hardware characteristics, so that codes generated by the compiler can better adapt to target hardware, and the hardware utilization rate is improved.

The invention relates to the technical field of artificial intelligence and software development, in particular to a context-aware code generation and real-time optimization method and system based on a large language model.The method comprises the following steps that a compiler front-end technology is utilized, input codes are analyzed, and an abstract syntax tree is constructed; function definition, variable declaration and statement block structure information are extracted through an abstract syntax tree, so that syntax and semantic structures of codes are obtained, and basic data support is provided for subsequent code generation and optimization; the method has the beneficial effects that the context graph covering grammar, semantics and runtime behaviors is constructed by fusing static grammar analysis (AST construction) and dynamic execution tracking (such as variable value change and function call time sequence), and the problem of context perception deficiency in a traditional code generation technology is solved. And the compatibility, the maintainability and the expandability of the generated code and the project architecture are improved.

The invention provides a business processing method and device for an online transaction scene and electronic equipment. Comprising the following steps: acquiring business data, defining a logic arrangement keyword based on an online transaction execution language, arranging the business data according to the logic arrangement keyword to obtain DSL arrangement data, constructing a grammar verification analyzer, and performing grammar verification processing on the DSL arrangement data by adopting the grammar verification analyzer to obtain the DSL arrangement data. An abstract syntax tree is constructed based on the verified DSL arrangement data, the abstract syntax tree comprises root nodes and leaf nodes, the root nodes represent arrangement rules and scheduling logic, and the leaf nodes represent actually called atomic transactions in the service data; and traversing each node in the abstract syntax tree by adopting an execution engine, and calling a node processor corresponding to each node to process the service data. The problem that an existing rule arrangement engine is poor in business reliability when being applied to banking businesses is solved.

The invention provides a low-code enterprise-level intelligent assistant development platform, and relates to the field of intelligent assistant development. The low-code enterprise-level intelligent assistant development platform comprises an application layer, a platform layer, a capability layer and an underlying service. According to the low-code enterprise-level intelligent assistant development platform, a low-code enterprise-level intelligent assistant development platform is realized, and by means of the platform, a user can develop enterprise-level intelligent assistants suitable for various industries, thereby helping enterprises to realize cost reduction and benefit increase; a user can complete the creation of a knowledge base, a workflow and an intelligent agent only through simple configuration or arrangement, a low use threshold is achieved, enterprise internal services or third-party MCP services can be accessed by means of a plug-in tool and a platform, uploading and publishing of user-defined tools are supported at the same time, high expandability is achieved, and the method can be widely applied to the field of enterprise management. The platform supports high-order functions such as multi-person collaborative development, multi-modal knowledge base construction, multi-modal data interpretation and processing and construction of multiple types of agents, and can meet complex enterprise-level application scenes.

Embodiments of the present disclosure provide a code generation method, a device, and a storage medium. In the embodiments of the present disclosure, a custom retrieval library can be constructed, wherein the custom retrieval library comprises sub-snippets segmented from reference code snippets; on this basis, a target sub-snippet meeting a preset similarity requirement can be retrieved for an uncompleted code from the custom retrieval library; and then a code context of the target sub-snippet can be acquired to generate a retrieval result. Since the custom retrieval library is at the sub-snippet level, the code diversity in the custom retrieval library can be effectively enhanced, and various code variants can be covered, so that the retrieval precision can be effectively improved, effectively improving the quality of retrieval results. Moreover, prompt information can be constructed on the basis of the generated retrieval result and input into a code generation model.

The invention relates to a modeling and formalized verification method for cache consistency protocol verification, and the method comprises the steps: building a protocol model which comprises a plurality of processor nodes, directory nodes and a communication mechanism through a structured modeling mode, and constructing an asynchronous message mechanism to simulate disordered concurrent communication; protocol behavior rules such as processor requests, directory responses and network receiving are defined, protocol property invariants are set, semantic constraints such as data consistency, confirmation before writing and sharing state legality are covered, model verification is carried out in a state space traversal mode, breadth-first search and a symmetry recognition mechanism are supported, and the method is suitable for the protocol behavior rules. A protocol error or deadlock state is effectively found, a traceable error path is generated, a verifier can automatically generate source codes through a compiler and operate on a host platform, and rapid verification and result output of a protocol model are achieved. The method has good protocol adaptability and model reusability, and is suitable for formalized verification of various cache consistency protocols.

The invention provides a two-hop incremental interface consistency verification method and system based on a knowledge graph, and the method comprises the steps: 1, carrying out the language analysis of a modified source file, and extracting an interface entity; 2, mapping the interface entity into a language-independent meta-model, and writing the language-independent meta-model into a knowledge graph; 3, executing a two-hop breadth-first traversal algorithm by taking a modified node in the knowledge graph as a starting point to obtain an affected sub-graph; 4, a hot-pluggable rule set is operated on the affected subgraph for verification, the rule set comprises a parameter number matching rule, a type compatibility rule, a version number increasing rule and a field deletion forbidding rule, and if verification is illegal, an error report is generated, and a blocking instruction is output; and if no violation exists in the verification, merging the affected sub-graphs back to the knowledge graph cache. According to the method and the device, online faults caused by field missing during operation are avoided, and the influence of back-end change on the front end is automatically detected and blocked.

The invention provides an inference method and device, equipment and a storage medium, and relates to the technical field of artificial intelligence, in particular to the technical fields of deep learning, large language models and the like. The method comprises the steps of determining target metadata of a target operator through a predefined function interface in response to an identified extended dependency package corresponding to the reference target operator in an inference framework; a calculation graph is generated according to the target metadata and native metadata, and the native metadata is metadata of other operators except the target operator in the reasoning framework; and reasoning on the target hardware according to the input data and the calculation graph to obtain a reasoning result. According to the method, the operator and the frame are decoupled, meanwhile, the fracture condition of the computational graph is avoided, and the compatibility and reasoning efficiency of the reasoning frame are improved.

The invention relates to a deep learning compiler automatic tuning method based on reinforcement learning. The method comprises the steps of comparing a scheduling task with a historical scheduling task in a multi-level offline database through a multi-level feature similarity matching strategy from top to bottom to determine a search starting point scheduling configuration; generating a feature matrix according to the feature vector corresponding to each type of scheduling task, and screening out a significant dimension reduction search space through a variance threshold selection method; based on a reinforcement learning method, selecting a DQN model, generating a random action through a Monte Carlo sampling strategy, and changing scheduling configuration according to the random action; and evaluating the changed scheduling configuration performance by using a cost model, and constructing a reward function based on an evaluation result to complete automatic tuning. By constructing a multi-level off-line database and utilizing feature similarity matching from top to bottom, starting point scheduling configuration can be quickly retrieved and determined; and a variance threshold selection method is adopted to carry out dimension reduction on feature dimension vectors, so that the retrieval efficiency is improved.

The embodiment of the invention discloses an operator compiling method, an operator compiling device, electronic equipment and a computer readable storage medium. The method comprises the steps of obtaining an intermediate representation of a to-be-compiled operator, wherein the intermediate representation of the to-be-compiled operator at least comprises an intermediate representation of a shared memory and an intermediate representation of an interactive operator; based on the size of the shared memory and the intermediate representation of the interactive operator, configuring operator rearrangement granularity in the intermediate representation of the shared memory to generate the configured intermediate representation of the shared memory; and based on the configured intermediate representation of the shared memory, updating the intermediate representation of the operator to be compiled, and compiling the intermediate representation of the operator to be compiled to generate an assembly code corresponding to the operator to be compiled. According to the embodiment of the invention, the operator development efficiency is effectively improved, the error probability is reduced, and the performance of the operator is optimized, so that the operator is more adaptive to an artificial intelligence chip.

The invention discloses an extranet-driven intranet and extranet security isolation compiling system, and belongs to the technical field of software development. Comprising an SFTP server, and the SFTP server is connected with an intranet and an extranet. The SFTP server is used for internal and external network upload management and control, internal and external network physical isolation and file synchronization. The intranet comprises an automatic script, a configuration file detection module, an automatic compiling module and an intranet log module; the external network comprises a front end and a rear end, the front end is used for providing a unified operation entrance for a user, and the rear end is used for driving the front end to work through a data processing algorithm; when a user initiates a request through a webpage, a system creates task groups, and each task group comprises a plurality of sub-tasks; at the moment, the extranet executes an extranet process to perform extranet legality verification, and performs task release; and meanwhile, the intranet executes an intranet process and periodically performs intranet legality detection. The method can ensure that each link masters the edition issuing key information consistently, and effectively improves the information transmission efficiency and accuracy.

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.

The invention provides a method, a system, equipment and a medium for optimizing page layout based on a large model, and belongs to the technical field of webpage design, the method comprises the following steps: creating an intelligent agent for page optimization, and installing server deployment, Web application construction and Web test tools; responding to a page generation demand, calling a Web application construction tool through a large model to create an initial page code with a burying point, and performing modification confirmation according to user opinions; responding to the implementation request, calling a server deployment tool through a large model to deploy simulation and formal environments, and running page codes in the simulation and formal environments in sequence; and in response to an optimization demand, regularly collecting user operation data through burying points, inputting the user operation data into the large model to generate an optimization suggestion, calling a Web application construction tool to modify page codes, and sequentially running in a simulation environment and a formal environment until the demand is met. According to the method, the page optimization efficiency and accuracy are improved, the labor cost is reduced, the user demand is responded quickly, and the user experience and the page performance are improved.

An embodiment of the present invention discloses a vehicle electronic automation test method, device, system and medium. The method includes: if it is monitored that a new version is released in the version database, downloading the new version project from the version database and determining the version number information of the new version project; copying the lower computer test code from the test database and editing the lower computer test code according to the version number information of the new version project to obtain a lower computer editing result; compiling according to the new version project and the lower computer editing result to obtain a target file; burning the target file into the electronic control unit in the vehicle and testing the electronic control unit with the target file. The embodiment of the present invention reduces the degree of manual participation and improves the generality, sustainability and test efficiency of vehicle electronic automation testing.

The invention discloses an analysis method and device for a Simulink model and electronic equipment, and the method comprises the steps: extracting an intermediate file ModelName.rtw from a target Simulink model, traversing the intermediate file, determining a hierarchical structure relation of the model based on a preset hierarchical division symbol "sum", and retrieving the intermediate file by using a plurality of keywords, basic information, parameter information and signal information of the target Simulink model are extracted; generating an analysis result based on the basic information, the parameter information and the signal information, wherein the analysis result is used for communication mapping between the upper computer and the controller; according to the method, the problems that in the prior art, the structure and content of a Simulink model are variable, and data loss and conversion errors are prone to occurring in the conversion process are solved, keyword retrieval and extraction are conducted through a hierarchical structure, then analysis of the Simulink model is achieved, key information in any Simulink model is obtained, and when a rapid prototype control system is adopted, the conversion efficiency of the Simulink model is improved. And the universality of the function applied to the software development of the upper computer is enhanced.

The invention provides a Wasserstein distance and Bayesian Thompson sampling-based compiler collaborative optimization method and system, and the method comprises the steps: determining a decision order based on a preset optimization stage mapping list, and constructing an optimization space into a hierarchical decision structure; traversing the hierarchical decision structure, selecting optimal child nodes layer by layer according to Thompson sampling, and generating candidate configuration; selecting a target configuration parameter based on the candidate configuration, injecting the target configuration parameter into the to-be-optimized program, and obtaining performance data; establishing a binning histogram model, extracting the distribution data of the optimal configuration and the worst configuration in the same optimization stage, and quantifying the Wasserstein distance value of the two distributions; and judging a Wasserstein distance value, triggering an optimization stage reordering mechanism, and reconstructing a hierarchical decision structure. According to the method, the integration of the robust distribution measurement and the adaptive decision in the automatic tuning of the compiler is realized.

Adaptive on board estimation of exogenous pharmacon responsive to transient (i.e., impermanent) physiological effects is provided. Dynamically estimating an equivalent amount of an exogenous pharmacon on board (XOB), such as insulin and / or carbohydrates, left in the subject, is based on predictions of glucose time-series data. These estimated values, such as insulin on board (IOB), are useful for diabetes management software, including decision support and / or artificial pancreas (AP) algorithms, for example.

The invention discloses a data encoding and decoding method and device for power regulation and control application service information interaction, and belongs to the technical field of power system regulation and control, and the method comprises the steps: obtaining structured data from a power regulation and control system through a client, the structured data comprising power system operation parameters; calling a coding interface through a client, coding the structured data into a binary data stream according to a predefined message description language specification, and transmitting the binary data stream to a server through a power communication network through the client; receiving the binary data stream through a server, and calling a decoding interface to decode the binary data stream into structured data; the decoded structured data is transmitted to a power regulation and control application service module for use through the server side; according to the method, the network overhead in the interaction process of the power regulation and control system is greatly optimized, and the overall performance and safety of the system are improved.

The invention discloses an actuarial calculation engine code compiler, an operation method and a related product. The actuarial calculation engine code compiler comprises an editing layer, an analysis layer and a mapping layer. The editing layer is configured to compile an actuarial process about the target actuarial business by adopting actuarial business terms and actuarial logic in response to the operation of the client; the analysis layer is configured to analyze the actuarial process and identify variables, formulas and time sequences in the actuarial process; generating a main syntax tree comprising target actuarial business semantics and a sub syntax tree comprising operation symbols and functions; and a mapping layer configured to parse the main syntax tree and the sub-syntax tree into machine executable code. In the embodiment of the invention, the actuarial process is edited through the actuarial business terms and the actuarial logic, and the actuarial process is converted into the machine executable code, so that the actuarial process adapts to high professionality and high suitability of professional evaluation of actuarial professionals.

The invention discloses a software module compiling method, electronic equipment, a medium and a product, and the method comprises the following steps: after a modification operation on a first module is monitored, determining changed first content in the first module; obtaining a pre-generated module dependency graph and a binding mapping table; determining at least one second module having a dependency relationship with the first module according to the module dependency graph; according to the binding mapping table, screening out at least one recompiling module of which the imported content comprises the first content from the at least one second module; according to the method, at least one screened recompilation module is recompiled, so that the consistency and correctness of codes are ensured, only the affected part is compiled, the time and computing resources for software module compiling and hot updating are saved, the development efficiency is improved, the resource utilization is optimized, and the development cost is reduced. And the maintainability, the expandability and the project management efficiency of the software project are improved.

The invention provides a program compiling method and device, electronic equipment, a storage medium and a program product. The method comprises the steps of obtaining a to-be-processed program; from the to-be-processed program, determining a task interface used for indicating to distribute the coprocessor task to the coprocessor for execution; generating a first execution code of the coprocessor task on the coprocessor based on the logic relationship of the task interface in the to-be-processed program; original logic used for calling a task interface in the to-be-processed program is updated into control logic executed by the central processing unit, an updated program is obtained, and the control logic is used for the central processing unit to drive the coprocessor to execute the first execution code; and based on the update program, generating a second execution code of the central processing unit task on the central processing unit, and packaging the first execution code and the second execution code to obtain a compiling result of the to-be-processed program. Through the method, the compiling efficiency can be effectively improved.

A source code translation system translates a source code snippet in an original source code programming language into distilled code that preserves the core semantics, logic flow, and data flow of the source code snippet in a language-independent format. The distilled code is then translated into a semantically-equivalent source code snippet in a different programming language by a decompiler. The decompiler is configured as a neural transformer model with attention pre-trained for translation using masked language modeling, denoising auto-encoding and multilingual program generation. The decompiler translates the distilled code into one of several different high-level programming languages.

This invention discloses a method and system for flattening a multi-level hybrid model in the X language, belonging to the field of model compilation technology. The method comprises: obtaining an X language coupling model; and recursively flattening the coupling models within the X language coupling model. By recursively flattening the coupling models within the X language coupling model, the present invention can integrate and flatten continuous models at different levels of a multi-level continuous-discrete hybrid model established in the X language, connected by equations, into a unified system of equations, thereby facilitating the underlying solution and simulation of the X language hybrid model.

The invention discloses a RESTFUL API rapid generation and release device and method, and the device comprises an API service management module which is used for providing an API interface development and management interactive interface for a data server user, achieving the configuration, testing and release of a newly-added API, and achieving the management function of the existing API, project and environment; the API service opening module is used for providing an API reading and subscribing interaction interface for a user at the service system end; and a network management and control module, a data basic service module and the like. The device provides an interface mode or an SQL configuration mode to generate an API; dynamic online and offline management can be carried out on the API; multiple sets of environments can be managed, the APIs can be dynamically published to the specified environment, and the registration publishing and management efficiency of the APIs is effectively improved.

The invention discloses a code processing method and device, electronic equipment, a storage medium and a computer program product. The method comprises the steps that an original code is obtained, the original code is used for declaring that one or more quantum bits are written into a pointer array, the pointer array comprises one or more pointers, and the one or more pointers are used for pointing to the one or more quantum bits; determining the type of the pointer array and the name of the one or more quantum bits by using the original code; constructing an object array by using the type of the pointer array and the name of the one or more quantum bits, the object array being used for pointing to the one or more quantum bits; and replacing the pointer array with the object array to obtain a compiled intermediate code. According to the technical scheme, in the code compiling process, the pointers of the pointer array are replaced with the specific objects pointed by the pointers, and the processing efficiency when the objects associated with the pointer array are called subsequently can be improved.

The embodiment of the invention provides a method for collecting data, a storage medium, a processor and an engineering vehicle, the method is applied to a first terminal provided with a nifi application program, and the method comprises the steps of obtaining a first operation instruction which is triggered by a user through an operation interface of the nifi application program and aims at a first icon, the first icon is a display icon of a preset nifi component in a nifi application program; according to the first operation instruction, obtaining a configuration parameter which is input by a user and aims at each preset nifi component; generating a first template file corresponding to the target device according to the configuration parameters of each preset nifi component; and sending the first template file to a second terminal corresponding to the target equipment so as to carry out a corresponding data acquisition operation through the second terminal according to the first template file. According to the technical scheme, a large number of components do not need to be repeatedly developed, the data acquisition time period is shortened, and the data acquisition efficiency is greatly improved.

The invention relates to the technical field of function engine configuration, and discloses a development method and system based on a low-code configuration function engine, and the method comprises the steps: decoupling bottom configuration information and specific business configuration according to the core capability demand of the function engine, and configuring a basic support module for the operation of the function engine; configuring a core function of the function engine based on the basic support module, and converting a core function configuration process of the function engine into an executable logic or instruction; configuration of executable logic or instructions is achieved through a visual user interface, an operation and maintenance mechanism is adjusted in real time in combination with the operation state in the operation process of the function engine, and core functions of the function engine are continuously iterated in combination with user feedback and service changes. According to the method, the function engine can be quickly configured while the code quantity is reduced, so that business changes are quickly responded, the efficiency is improved, and the cost is reduced.