1071 results about "Symmetric multiprocessor system" patented technology

The invention relates to the technical field of computing power scheduling, and discloses a heterogeneous resource computing power intelligent scheduling method and system. According to the method, real-time state monitoring is conducted on heterogeneous computing resources, and resource state parameters such as the computing unit utilization rate and the memory occupancy rate are obtained; task attributes and user request parameters of the task queue are collected, historical task data are processed based on the genetic algorithm optimization model to execute task demand prediction, and predicted demand parameters are generated. A dependency graph containing resource unit nodes and communication link roadsides is constructed through a resource topology analysis tool, predicted demand parameters are input into a scheduling priority classifier trained by a graph neural network, and an actual scheduling priority is identified. And executing resource conflict prediction based on the priority, inputting task feature vectors into a conflict resolution module of a fuzzy logic decision maker, outputting actual conflict resolution parameters, and finally integrating to generate a scheduling scheme containing a resource allocation sequence and an execution time table.

The invention provides a complex scene-oriented AI large model lightweight deployment method, and relates to the technical field of edge computing, and the method comprises the steps: carrying out the structured pruning of a pre-trained Transform network based on the attention head importance score, carrying out the dynamic sparsification of the activation state of a feedforward network according to the input tensor entropy value, employing the dynamic mixing precision quantization, and carrying out the reconstruction of an AI large model. Obtaining network parameters after pruning quantization; deploying the pruned and quantized network parameters to an edge computing device, distributing a feature extraction operator to a neural network processor through a heterogeneous computing scheduler, and unloading a classification operator to a multi-core central processing unit; and managing an on-chip memory in combination with a virtual memory paging mechanism, realizing zero-copy data transmission by utilizing a direct memory access controller, and outputting a reasoning result tensor. According to the method, efficient and reliable operation of the large model at the resource-constrained edge node is realized.

The invention discloses an artificial intelligence large model training method in a heterogeneous multi-machine and multi-card environment, and belongs to the technical field of artificial intelligence large model training. Load balancing of heterogeneous equipment is realized by constructing a uniform interface, and the communication efficiency is optimized by adopting hierarchical pipeline aggregation and dynamic quantization compression; the node dynamic adjustment is realized in combination with the elastic topological structure, the problems of poor equipment compatibility, high communication delay and rigid topological structure in the prior art are effectively solved, and the method has the remarkable advantages of improving the utilization rate of heterogeneous computing resources, reducing the cross-node communication overhead and enhancing the fault-tolerant capability of the system.

The present application belongs to the technical field of parallel task scheduling. Specifically, disclosed are a heterogeneous computing resource scheduling method and apparatus based on multi-objective optimization. The method comprises: selecting at least two performance indicators from both a task dimension and a resource dimension as optimization objectives, and establishing a multi-objective optimization model for heterogeneous computing resource scheduling; converting a computation task request process into a computation task waiting model on the basis of a queuing theory; on the basis of an observed resource state, constructing a multi-task adaptive scheduling model based on reinforcement learning; on the basis of the multi-objective optimization model and the computation task waiting model, constructing a Markov decision process model from a multi-task-oriented heterogeneous computing resource scheduling problem and a resource mapping process; and on the basis of the Markov decision process model and the multi-task adaptive scheduling model, realizing adaptive multi-task heterogeneous computing resource scheduling. The embodiments of the present application solve the problem of it being difficult for a homogeneous computing resource scheduling method to adapt to heterogeneous computing resource scheduling.

The invention relates to a self-adaptive multi-algorithm scheduling method and system based on cloud edge collaboration, and the method comprises the steps: constructing a cloud algorithm knowledge base and a scheduling strategy model, which are used for storing a plurality of algorithms, and providing a unified scheduling rule and optimization criterion; designing an edge node real-time sensing and reporting module, dynamically monitoring the computing power state, task characteristics and operation environment of the node, and transmitting related information to the cloud in real time; a cloud intelligent scheduling decision module is constructed, a knowledge base and a scheduling strategy are combined, a received edge state is comprehensively analyzed, and an optimal algorithm selection and execution position decision is generated; an algorithm dynamic scheduling and heterogeneous execution module is deployed on the edge side, and a target algorithm is flexibly loaded and executed on a local cache or heterogeneous computing resources according to an instruction issued by the cloud; and the cloud edge collaborative closed-loop iteration and adaptive optimization module realizes adaptive iteration and continuous optimization of algorithm scheduling, so that high robustness and high efficiency of task execution in a complex and changeable scene are ensured.

The invention belongs to the technical field of computing resource scheduling, and particularly relates to an intelligent scheduling method and system for Huaan Atlas heterogeneous computing resources based on dynamic load awareness. The method comprises the steps that the real-time state of multi-dimensional hardware data is collected, and a basic data source is provided for subsequent steps; dynamically adapting tasks and hardware characteristics through a matching degree matrix, modeling aiming at various basic data, and constructing a state vector required by reinforcement learning; predicting a fault risk score through a lightweight prediction model deployed at each computing node; and a deep Q network is adopted as a model architecture, a state vector and a fault risk score are input, reinforcement learning training is performed through a reward function in a multi-target vector form, a final scheduling model is obtained, and a task allocation decision is output. The problems that in the prior art, the hardware state cannot be sensed in real time, hardware characteristic matching is ignored, consequently, the computing resource utilization rate is insufficient, and fault recovery is passive are solved.

The invention relates to the field of scene rendering, and particularly discloses a three-dimensional scene optimization method and system for collaborative rendering of dynamic LOD and view cone rejection based on space-time prediction, and the method comprises the steps: dynamically calculating the visibility frequency of an object, and dynamically adjusting the loading and rendering modes of objects with different priorities; predicting a view cone range of multiple frames in the future by using an LSTM network architecture; dividing the scene into uniform grid blocks, and then performing grading elimination; optimizing a heterogeneous computing pipeline; rendering the execution process; the system comprises a dynamic LOD and visual cone rejection collaborative optimization module, an LSTM visual cone prediction module, a block-level mixed rejection module, a heterogeneous calculation pipeline module, an optimization CPU-GPU task allocation and data transmission module and a rendering execution control module. According to the method, the technologies of collaborative optimization of dynamic LOD and view cone removal, block-level mixed removal, heterogeneous calculation assembly line optimization and the like are adopted, so that unnecessary rendering calculation and resource loading are reduced, and the rendering efficiency is improved.

The invention discloses a heterogeneous computing power resource dynamic cooperative scheduling method and system and a computer device. The method comprises the steps of obtaining multi-dimensional feature information of a to-be-scheduled task, current state information of each node in a heterogeneous computing power resource pool and scheduling environment state information; training a time sequence model by adopting historical associated data, inputting the multi-dimensional feature information, the current state information and the scheduling environment state information into the trained model to obtain associated trend information in a future preset time window, and then combining the multi-dimensional feature information and the current state information of all the nodes to obtain the scheduling environment state information. A multi-objective optimization function fusing task delay, cost consumption and energy consumption is constructed, the function is optimized based on a preset strategy, a target decision strategy is obtained, a scheduling instruction set is generated after analysis, allocation information and target nodes are determined according to the scheduling instruction set, and tasks are issued. According to the method, dynamic collaborative scheduling of heterogeneous computing power resources can be realized, and the overall utilization efficiency of the resources is accurately, efficiently and effectively improved.

The invention relates to the technical field of edge computing, and discloses an intelligent edge computing co-processing system based on an integrated circuit, which comprises a heterogeneous computing cluster module, a hardware computing unit set, a multi-core control processor based on RISC-V, a programmable pulse tensor computing array and a reconfigurable engine oriented to streaming processing, according to the intelligent edge computing cooperative processing system based on the integrated circuit, zero-delay data exchange is realized through silicon intermediate layer integration of the heterogeneous computing cluster module and a snakelike data channel, and traditional bus arbitration delay is eliminated through real-time operation code analysis and optimal optical communication path mapping of the hardware task routing matrix (TRF) module; the priority path distribution of the task scheduling subsystem is synchronously coordinated with the time-sensitive task, so that the collaborative efficiency of the computing unit is improved in multiple dimensions, the non-blocking transmission of the high-priority task is ensured, and the effect of enhancing the overall processing capability and response speed of the intelligent edge computing is achieved.

The invention relates to the technical field of new energy storage regulation and control, and discloses an electricity-hydrogen coupling energy storage system and method for a new energy field station. The system comprises a feature analysis unit, a topology construction unit, a regulation and control domain positioning unit and a decision optimization unit. The feature analysis unit receives operation condition information of the new energy station, and generates a condition feature vector according to real-time load fluctuation, a historical output curve and an equipment operation database by using a deep feature extraction network; the topology construction unit constructs an energy topological graph for distributed energy storage equipment and hydrogen production equipment in the electricity-hydrogen coupling system, dynamically updates a graph structure and a node state based on a real-time monitoring mechanism, and deploys the graph structure and the node state in a heterogeneous computing framework; a regulation and control domain positioning unit positions a matched regulation and control domain in a dynamic representation learning mode according to the working condition feature vector; and the decision optimization unit utilizes a multi-source decision fusion algorithm to screen regulation and control instructions in a regulation and control domain and generate a scheduling sequence, so that the adaptability and the operation efficiency of the system to working conditions are improved.

The invention discloses a heterogeneous GPU resource management scheduling method, a computer device, a medium and a product. The method comprises the following steps: acquiring computing power resources of each node, including a GPU model, a GPU video memory, a GPU number, a computing power segmentation scheme and a GPU use condition; automatically segmenting the heterogeneous GPU of each node according to the computing power segmentation scheme of each node to obtain resource segmentation information; obtaining an expected computing power and an expected video memory of a to-be-executed task; screening out target nodes meeting the expected computing power and the expected video memory according to a node analysis strategy; and according to the sub-resource analysis strategy, screening out sub-resources meeting the expected computing power and the expected video memory, recording the sub-resources as target sub-resources, and allocating the to-be-executed task to the target sub-resources. According to the method, GPU resource fragmentation is effectively reduced through global and node resource conjoint analysis scheduling, meanwhile, pooling management, dynamic configuration and intelligent scheduling of heterogeneous computing equipment can be achieved, and the resource utilization rate is effectively increased.

The invention relates to the field of computer systems, in particular to an operation stability testing method and system oriented to a heterogeneous computing system, and the method comprises the steps: constructing a system overall operation logic model comprising a resource topological structure, a task scheduling rule and a communication link matrix; identifying a key communication path according to the communication link matrix, analyzing a resource competition hotspot in combination with a task scheduling rule, and establishing an interference model library; calling a specific model instance in the interference model library according to a test target, configuring injection parameters, and dynamically applying interference in a system operation process; key performance indexes in the system operation process are collected in real time; based on the real-time monitoring data, a multi-index weighted scoring algorithm is adopted to calculate an overall stability coefficient, a robustness grade and a self-healing capability index of the system, and a stability evaluation result is generated; and generating a visual report according to the topological structure of the operation logic model, and supporting comparison and analysis of historical versions. Unified centralized management is realized, and test efficiency and evaluation accuracy are improved.

The invention discloses a distributed computing power dynamic scheduling method and device and a medium, and the method comprises the steps: packaging heterogeneous computing power resources of a home terminal and an edge cloud node through a lightweight containerization technology, and collecting the hardware resource state data of the home terminal and the load data of the edge cloud node in real time; generating a dynamic scheduling strategy according to the task calculation type and the real-time network state of the to-be-calculated task, and splitting the to-be-calculated task into a plurality of sub-tasks according to the dynamic scheduling strategy; distributing the sub-tasks to home terminals or edge cloud nodes according with a dynamic scheduling strategy, and aggregating calculation results of the sub-tasks; and performing end-to-end encryption and fragmentation verification on the cross-domain transmitted data stream, and dynamically adjusting the task allocation permission of the home terminal based on the equipment security score.

The invention discloses a distributed training scheduling and communication optimization method and system of a multi-modal large model on a domestic computing power platform. The method comprises the following steps: virtualizing a heterogeneous computing unit of a preset platform into a virtual device pool, and fusing first-order gradient of a multi-modal sample and Hessian matrix information based on quantitative perception training to generate a sample sensitivity grading atlas; virtual device pool attributes and the sensitivity grading atlas are used as input, an optimal hybrid parallel configuration scheme is automatically generated through a configuration search algorithm, and a parallel combination mode, resource mapping and a high-sensitivity sample scheduling strategy are defined; a distributed training code of an integrated communication optimization strategy is automatically generated according to a configuration scheme, pipeline parallel communication and data parallel gradient synchronization constraint are executed in a topology adjacent equipment subset, and a hierarchical aggregation mechanism is adopted; and dynamically screening a core training set and scheduling a calculation task to complete distributed training. According to the method, efficient cooperative training of the multi-modal large model on the domestic computing power platform is realized.

The invention discloses a task scheduling method based on predictable resource state atlas modeling, a platform is oriented to a heterogeneous computing environment, a unified resource state atlas is constructed by collecting multi-dimensional resource state parameters of computing nodes, and performance characteristics and communication topological relations among the nodes are comprehensively described. On the basis, a bidirectional time sequence model and an attention mechanism are fused, and the load trend of each node in a future short time is predicted. The platform constructs a multi-factor scheduling scoring function based on a task feature vector and resource state prediction map, integrates parameters such as resource matching degree, prediction load, communication delay and energy consumption cost, dynamically evaluates the adaptability of tasks and resources, and realizes adaptive scheduling and optimal resource allocation of the tasks. Compared with the prior art, the method has the advantages of being high in resource state predictability, high in task allocation intelligence degree, outstanding in platform evolution capability and the like, and is suitable for intelligent task scheduling application in a large-scale heterogeneous resource environment.

The invention belongs to the technical field of defect identification, and particularly relates to a gas turbine blade defect identification method based on an improved YOLOV8 network. Comprising the following steps: capturing a defect image; effective information is extracted from the image or subsequent target detection, classification and segmentation task requirements are met; performing data enhancement through geometric transformation and color perturbation; adding structured labels or annotations to the images, and endowing the images with semantic information; the scheme is improved on the basis of YOLOv8 so as to be realized in gas turbine blade defect detection, and training of the model is completed in a distributed heterogeneous computing framework by using an acquired training data set and an acquired verification data set and is used for a prediction task. The method can achieve the pixel-level detection and positioning of the defects of the blade, can effectively meet the daily detection demands, and gives consideration to the detection precision and real-time performance.

The invention relates to the technical field of cloud edge collaborative computing power scheduling, and discloses a heterogeneous computing power scheduling optimization method based on a cloud edge collaborative architecture. The method comprises the following steps: acquiring real-time computing power state data of all available computing nodes in the cloud edge collaborative architecture; performing heterogeneous type division on the computing nodes according to the real-time computing power state data to generate a three-layer computing power resource pool containing cloud computing nodes, edge computing nodes and terminal computing nodes; extracting task calculation features for the current to-be-scheduled task set, wherein the features comprise calculation intensity, data dependence and real-time requirements; constructing an initial task allocation scheme based on the matching relationship between the task calculation features and the three-layer computing power resource pool; iteratively optimizing the initial scheme by adopting a dynamic load balancing strategy to generate a final task scheduling instruction; the instructions are distributed to the corresponding computing nodes to be executed, and computing power state changes in the task execution process are continuously monitored.

The invention discloses a heterogeneous computing power cooperative scheduling system and method for mixed precision training, and belongs to the technical field of artificial intelligence computing. The system comprises a computational graph analysis and operator portrait module which is used for analyzing and dividing a model computational graph and extracting operator features; the heterogeneous hardware capability sensing and matching module is used for managing performance files and real-time states of heterogeneous hardware in the cluster and matching optimal execution hardware for each calculation partition; and the data flow coordination and pipeline parallel controller is used for generating a global execution plan, managing cross-device data dependence and communication and calculating overlapping optimization execution efficiency through communication. According to the method, the problem of low scheduling efficiency of mixed precision training in a heterogeneous environment is solved, automatic and accurate mapping from a calculation task to heterogeneous hardware is realized, the training speed is remarkably improved, the training cost is reduced, and the overall resource utilization rate of a cluster is improved.

The invention provides a heterogeneous computing power resource pooling scheduling platform which comprises a task and resource tensor construction module, a scheduling decision module, a breakpoint state generation module and a rescheduling and recovery module. According to the platform, task requirements and node resources are modeled into a tensor structure in a unified mode, and accurate initial scheduling is achieved through a structured scoring function containing conditional penalty terms. In the task running process, interruption judgment is carried out based on the node dynamic state and the scheduling score, and a breakpoint state tensor containing an execution state is generated. When interruption occurs, a proper node is selected through compatibility screening, and task seamless migration and execution recovery are realized by utilizing the stored state. According to the method, the technical problems of non-uniform resource expression, lack of scene adaptability and incapability of realizing uninterrupted migration in a heterogeneous computing power environment are solved, and the resource utilization rate and the service continuity are remarkably improved.

Provided in the embodiments of the present application are a server system, a resource scheduling method for a server system, and a chip and a chiplet. The server system comprises: a multivariate computing resource pool, a data storage resource pool, a switching module and a control module, wherein the multivariate computing resource pool comprises a general computing resource pool and a heterogeneous computing resource pool, the general computing resource pool comprising a group of general computing units, and the heterogeneous computing resource pool comprising a group of heterogeneous computing units, and the multivariate computing resource pool is connected to the data storage resource pool via the switching module by means of a cache consistency bus inside the server system; the data storage resource pool comprises data storage resources allowed to be shared by multivariate computing resources in the multivariate computing resource pool; and the control module is configured to perform computing power scheduling on the multivariate computing resources in the multivariate computing resource pool, and dynamically allocate the data storage resources in the data storage resource pool to the multivariate computing resources in the multivariate computing resource pool.

The invention relates to a heterogeneous computing resource capability radiation scheduling method and device. The method comprises the following steps: determining computing capabilities of a plurality of receiving nodes in a radiation center; splitting the original encryption calculation task into a plurality of privacy protection calculation task packages; determining a node matching result according to the computing capabilities of the plurality of receiving nodes and the Pareto optimal solutions of the plurality of privacy protection computing task packages; distributing the plurality of privacy protection computing task packets to corresponding receiving nodes to obtain a distribution voucher of each receiving node; performing distributed verification on the distribution voucher to obtain a distributed verification result; performing multi-modal aggregation on all the distributed verification results to obtain an encrypted calculation radiation scheduling result; the problems of data privacy leakage, low resource utilization rate and insufficient task collaboration efficiency in a traditional scheduling mechanism are solved through encryption task disassembly, dynamic node matching capability, distributed verification and multi-modal aggregation, and the method has the advantages of improving data privacy, resource utilization rate and task collaboration efficiency.

The invention relates to a heterogeneous computing-based task scheduling method and device, a storage medium and equipment. The method comprises the following steps of: obtaining an inference task of a large language model; in the execution process of the reasoning task, at least based on the memory access intensity and the current reasoning stage, memory access intensive operators and calculation intensive operators in the reasoning task are recognized; allocating the memory access intensive operator to a first computing unit for executing a memory intensive task, and allocating a computing intensive operator to a second computing unit for executing a computing intensive task; obtaining estimated execution time of the two calculation units for executing the corresponding tasks and data transmission time between the two calculation units; according to the pre-estimated execution time and the data transmission time, the task starting moments of the first calculation unit and the second calculation unit are determined with the purpose of minimizing the overall execution delay of the reasoning task; and controlling the first calculation unit and the second calculation unit to asynchronously execute the corresponding tasks in parallel according to the task starting time.

The invention relates to the technical field of computing power intelligent scheduling, and discloses a computing power intelligent scheduling method and system for a heterogeneous computing power cluster, and the method comprises the steps: firstly analyzing task features from task description submitted by a user, and generating an internal task object; and then, constructing a resource portrait by collecting real-time resource monitoring data and static configuration information of each node in the heterogeneous computing power cluster. Based on the information, the execution performance of different tasks on each node is predicted by using a machine learning model, and a performance prediction mapping table is formed. Then, candidate resources are screened and sorted according to the mapping table, and it is ensured that the optimal node is selected to be bound with the task; therefore, not only is the matching degree of the task characteristics and the hardware attributes considered, but also the dynamic state of the hardware resources is combined, so that a more accurate task scheduling strategy is realized, and the overall operation efficiency and the resource utilization rate are effectively improved.

The invention provides a large model AI reasoning acceleration and deployment method based on heterogeneous calculation, and relates to the technical field of AI.The method comprises the steps that a neural network model and a reasoning request are obtained, and a calculation graph is analyzed; decomposing a calculation task into sub-tasks and marking calculation characteristics; establishing an adaptive mapping relationship between the subtasks and the heterogeneous computing units; constructing a cross-computing-unit data flow path and executing a scheduling strategy; and executing reasoning calculation and returning a result. According to the method, the heterogeneous computing resource characteristics can be fully utilized, the reasoning delay is reduced, the computing resource utilization efficiency is improved, and efficient deployment of the large model AI is realized.

The invention relates to the technical field of power system automation, and particularly discloses a cross-domain heterogeneous computing power real-time calling and unified scheduling system for the power industry, which comprises a computing power resource global sensing unit, a business demand dynamic modeling unit, a cross-domain unified scheduling decision unit and a task execution and feedback control unit, according to the method, heterogeneous computing power states are collected in real time, task requirements are dynamically modeled, a scheduling scheme is generated based on multi-objective optimization, and task execution monitoring and load balancing are realized by means of closed-loop feedback, so that the utilization efficiency of global computing power resources is improved, and the real-time performance and reliability of power business are guaranteed.

The invention provides a heterogeneous computing resource scheduling method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining a to-be-scheduled computing task, and extracting operator features of the to-be-scheduled computing task; splitting the to-be-scheduled calculation task into a first sub-task suitable for being executed by a graphics processor and a second sub-task suitable for being executed by a quantum calculation unit according to the operator characteristics; a first resource group is matched for the first subtask in a preset resource group set, a second resource group is matched for the second subtask, and the preset resource group set comprises the first resource group and the second resource group which are obtained by dividing a heterogeneous computing resource pool; the first resource group comprises at least one graphics processor computing unit, and the second resource group comprises at least one quantum computing unit; and scheduling the first sub-task in the first resource group, and scheduling the second sub-task in the second resource group.

The invention discloses a heterogeneous computing power resource dynamic scheduling system and method based on multi-objective optimization, and the system is deployed in a digital ecological platform complex based on multivariate consensus and embedded intelligent management. Comprising a platform access module, a data acquisition and perception module, a multi-target modeling and optimization module, a dynamic scheduling and execution module and a fault-tolerant mechanism module. The system registers as a computing power scheduling service node through an intelligent contract interface, obtains and verifies the compliance of a computing task, collects heterogeneous computing power resource node state data, and generates a scheduling decision by dynamically adjusting the weight of a target function; evaluating task suitability by using an AMCU suitability scoring model, and executing a scheduling AMCU strategy; according to the method, the technical problems of low resource scheduling efficiency and poor fault-tolerant capability in a heterogeneous computing power environment are solved, and the system throughput and the resource utilization rate are improved.

The invention discloses a computing power collaborative management method and system based on a cross-architecture state sensing engine, and relates to the technical field of GPU computing power pooling scheduling. The method comprises the following steps: establishing a communication link set between a plurality of GPU instances and a plurality of CPU controllers; collecting and analyzing indexes such as data throughput rate, queue length and communication time delay of each link in real time, and dynamically identifying a system bottleneck link; aiming at the detected bottleneck link, the fine granularity of the bottleneck link is split into a plurality of sub-links, and each sub-link is allocated with an independent cache queue and a flow control strategy, so that fine management and isolation of link resources are realized; furthermore, the scheduling priority is calculated according to the real-time state index of each sub-link, and the optimal sub-link is preferentially selected for task scheduling during task allocation. Therefore, the resource utilization rate and the task scheduling efficiency in the multi-tenant heterogeneous computing power pool environment are effectively improved, and the conditions of local link congestion and false idle / false busy of the computing power pool are remarkably relieved.

The invention provides a large-model-oriented heterogeneous computing power adaptive compiling method, which comprises the following steps that: a user inputs a trained large model through a system interface, and a system front-end conversion module analyzes a computational graph of the model and converts the computational graph into an intermediate representation based on a unified operator description language (UDL); the system hardware sensing module automatically detects and extracts hardware feature fingerprints of at least one piece of target hardware; based on the unified operator description language UDL intermediate representation and the hardware feature fingerprint, automatically generating an optimization adaptation rule oriented to at least one piece of target hardware; wherein the basis of parameterized filling comprises specific parameters of hardware feature fingerprints and optimized attribute tags carried in an intermediate representation of a unified operator description language (UDL); and generating and deploying multiple back-end codes. The method has the beneficial effects that intelligent compiling based on hardware features can be realized, and the deployment efficiency and the operation performance of a large model in a complex heterogeneous computing power cluster are remarkably improved.

The invention relates to the technical field of computers, discloses a heterogeneous computing low-delay communication method and system, and aims to solve the problem of high delay caused by high communication protocol overhead, lack of dynamic scheduling collaboration, memory migration redundancy and non-uniform cross-node communication abstraction in existing heterogeneous computing. The method comprises the following steps: receiving a task scheduling request and analyzing a task dependency graph; tasks are dynamically allocated based on node loads and link states; rDMA, NVLink or PCIe straight-through protocols are adaptively selected according to node types to establish communication channels; zero-copy data exchange is realized through a shared memory mapping buffer area; hardware timestamps are utilized to synchronize feedback delays with PTP to optimize scheduling. The system comprises a heterogeneous computing node cluster, a unified communication scheduling controller, a low-delay communication protocol stack, a shared memory mapping buffer area and a communication delay sensing task distributor. According to the scheme, the communication delay is remarkably reduced, and the throughput and the task execution efficiency are improved.