Parallel simulation method and device based on GPU resources

Abstract

The invention discloses a parallel simulation method and device based on GPU resources. The method comprises the steps that hardware resource information and to-be-processed simulation task information are acquired; the hardware resource information comprises a plurality of pieces of node resource information; the node resource information comprises M pieces of first resource information and second resource information corresponding to the first resource information; the first resource information comprises first GPU resource information and second GPU resource information; the second resource information comprises first storage resource information and second storage resource information; analyzing and processing the hardware resource information and the to-be-processed simulation task information to obtain target resource analysis information and target task analysis information; and performing matching processing on the target resource analysis information and the target task analysis information to obtain target matching result information.

Description

Technical Field

[0001] The present invention relates to the field of simulation technology, and in particular, to a parallel simulation method and device based on GPU resources. Background Art

[0002] With the continuous development of computer technology, simulation technology has been widely applied in fields such as scientific research, engineering design, industrial manufacturing, and biomedicine. Traditional simulation methods mainly rely on the CPU for computing. However, as the complexity and scale of simulation models continue to increase, the computing power of the CPU gradually becomes difficult to meet the requirements of efficient simulation. Due to its powerful parallel processing ability, the GPU has gradually become an important tool for high-performance computing. The GPU has a large number of cores and can process multiple tasks simultaneously, thus significantly improving the computing efficiency. However, there are still some problems in the current GPU simulation technology, such as unbalanced resource allocation during multi-GPU parallelism, low data transmission efficiency, and unstable parallel performance. These problems limit the application of the GPU in complex simulation tasks. Therefore, a parallel simulation method and device based on GPU resources are provided to improve the simulation efficiency and the utilization rate of simulation resources. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a parallel simulation method and device based on GPU resources, which are beneficial to improving the simulation efficiency and the utilization rate of simulation resources.

[0004] To solve the above technical problem, in the first aspect of an embodiment of the present invention, a parallel simulation method based on GPU resources is disclosed, and the method includes:

[0005] Obtain hardware resource information and information of simulation tasks to be processed; the hardware resource information includes information of several node resources; the node resource information includes M pieces of first resource information and second resource information corresponding to the first resource information; the first resource information includes first GPU resource information and second GPU resource information; the second resource information includes first storage resource information and second storage resource information;

[0006] Analyze and process the hardware resource information and the information of simulation tasks to be processed to obtain target resource analysis information and target task analysis information; the target resource analysis information includes several sequentially arranged pieces of first target resource information; the first target resource information includes a first resource serial number and M pieces of target resource available information; the target task analysis information includes information of resource requirements of several simulation tasks;

[0007] Perform matching processing on the target resource analysis information and the target task analysis information to obtain target matching result information.

[0008] In the second aspect of the embodiments of the present invention, a parallel simulation device based on GPU resources is disclosed. The device includes:

[0009] An acquisition module, configured to acquire hardware resource information and information of simulation tasks to be processed; the hardware resource information includes information of a plurality of node resources; the node resource information includes M pieces of first resource information and second resource information corresponding to the first resource information; the first resource information includes first GPU resource information and second GPU resource information; the second resource information includes first storage resource information and second storage resource information;

[0010] A first processing module, configured to analyze and process the hardware resource information and the information of simulation tasks to be processed, so as to obtain target resource analysis information and target task analysis information; the target resource analysis information includes a plurality of sequentially arranged first target resource information; the first target resource information includes a first resource serial number and M pieces of target resource available information; the target task analysis information includes a plurality of simulation task resource requirement information;

[0011] A second processing module, configured to perform matching processing on the target resource analysis information and the target task analysis information, so as to obtain target matching result information; the target matching result information includes a plurality of target resource task matching result information.

[0012] In the third aspect of the present invention, another parallel simulation device based on GPU resources is disclosed. The device includes:

[0013] A memory storing executable program code;

[0014] A processor coupled to the memory;

[0015] The processor calls the executable program code stored in the memory and executes some or all of the steps in the parallel simulation method based on GPU resources disclosed in the first aspect of the embodiments of the present invention.

[0016] In the fourth aspect of the present invention, a computer-readable storage medium is disclosed. The computer-readable storage medium stores computer instructions, which are used to execute some or all of the steps in the parallel simulation method based on GPU resources disclosed in the first aspect of the embodiments of the present invention when called. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0018] Figure 1 It is a schematic diagram of the scenario of the parallel simulation system based on GPU resources provided by an embodiment of the present invention;

[0019] Figure 2 It is a schematic flowchart of a parallel simulation method based on GPU resources disclosed by an embodiment of the present invention;

[0020] Figure 3 It is a schematic structural diagram of a parallel simulation device based on GPU resources disclosed by an embodiment of the present invention;

[0021] Figure 4 It is a schematic structural diagram of another parallel simulation device based on GPU resources disclosed by an embodiment of the present invention. Detailed implementation manners

[0022] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts fall within the protection scope of the present invention.

[0023] The terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, device, product or equipment that includes a series of steps or units is not limited to the listed steps or units, but optionally further includes steps or units not listed, or optionally further includes other steps or units inherent to these processes, methods, products or equipment.

[0024] Referring to "embodiment" herein means that a specific feature, structure or characteristic described in connection with the embodiment can be included in at least one embodiment of the present invention. The phrase appears in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0025] In this application, the term "exemplary" is used to mean "serving as an example, illustration, or instance". Any embodiment described as "exemplary" in this application is not necessarily to be construed as more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use this application. In the following description, details are set forth for purposes of explanation. It should be understood that those of ordinary skill in the art can recognize that this application can be implemented without these specific details. In other instances, well-known structures and processes are not elaborated in detail to avoid obscuring the description of this application with unnecessary details. Therefore, this application is not intended to be limited to the embodiments shown, but is to be accorded the widest scope consistent with the principles and features disclosed in this application.

[0026] It should be noted that since the method of the embodiment of this application is executed in a computer device, the processing objects of each computer device exist in the form of data or information. For example, time, which is actually time information. It can be understood that in subsequent embodiments, if dimensions, quantities, positions, etc. are mentioned, they are all corresponding data existences for the computer device to process, and specific details are not elaborated here.

[0027] A brief introduction to the artificial intelligence-related technologies that may be involved in this application is provided. Artificial Intelligence (AI) is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to obtain the best results. In other words, artificial intelligence is a comprehensive technology in computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can react in a way similar to human intelligence. Artificial intelligence also studies the design principles and implementation methods of various intelligent machines to enable the machines to have the functions of perception, reasoning, and decision-making.

[0028] Artificial intelligence technology is a comprehensive discipline that involves a wide range of fields, including both hardware-level technologies and software-level technologies. The basic technologies of artificial intelligence generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, and mechatronics. The software technologies of artificial intelligence mainly include several major directions such as computer vision technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0029] Computer Vision Technology (CV) Computer vision is a science that studies how to enable machines to "see". More specifically, it refers to using cameras and computers to replace human eyes for tasks such as object recognition and measurement in machine vision, and further performing image processing to make the computer-processed images more suitable for human eye observation or transmission to instrument detection. As a scientific discipline, computer vision researches related theories and technologies, and attempts to establish artificial intelligence systems that can obtain information from images or multi-dimensional data. Computer vision technology usually includes image processing, image recognition, image semantic understanding, image retrieval, OCR, video processing, video semantic understanding, video content / behavior recognition, three-dimensional object reconstruction, 3D technology, virtual reality, augmented reality, simultaneous localization and mapping, etc. technologies, and also includes common biometric recognition technologies such as face recognition and fingerprint recognition.

[0030] Single-modal information is data of only one type, such as one of the data information types of text, image, audio, video, electromagnetic signal, etc. Multi-modal information is data information that includes at least two single-modal information. Further, multi-modal information is applicable to complex tasks that require integrating multiple information sources, such as sentiment analysis, robot interaction, autonomous driving, etc. By integrating information of multiple modalities, higher performance and accuracy can usually be achieved in tasks.

[0031] A large model refers to an artificial neural network model with a very large number of parameters. In the field of artificial intelligence, a large model usually refers to a model with hundreds of millions to trillions of parameters. The model usually needs to be trained on a large-scale dataset and requires a large amount of computing resources for optimization and adjustment. Large models are usually used to solve complex natural language processing, computer vision, and speech recognition tasks. Generative AI is a type of AI that can create new content and ideas, including conversations, stories, images, videos, and music. In the embodiments of this application, the large model can be large language models such as ChatGPT, BERT, XLNet, Zhipu Model, Claude, Moonshot AI Model, ChatGLM Model, Tongwen Qianyi Model, MiniMax Model, Spark Model, Llama Model, 360GPT Model, Qwen Model, Baichuan Model, Lark Model, vivoLM Model, and Wenxin Yiyan, and the embodiments of this application do not make any limitations.

[0032] The embodiments of this application provide a parallel simulation method, device, computer device, and computer-readable storage medium based on GPU resources, which will be described in detail below.

[0033] Please refer to Figure 1 , Figure 1A schematic diagram of the scenario of the parallel simulation system based on GPU resources provided by the embodiments of the present application. The parallel simulation system based on GPU resources may include a computer device 100, and a parallel simulation device based on GPU resources is integrated in the computer device 100, such as Figure 1 the computer device in

[0034] In the embodiments of the present application, the computer device 100 is mainly used to obtain hardware resource information and information of simulation tasks to be processed; the hardware resource information includes information of several node resources; the node resource information includes M pieces of first resource information and second resource information corresponding to the first resource information; the first resource information includes first GPU resource information and second GPU resource information; the second resource information includes first storage resource information and second storage resource information;

[0035] Analyze and process the hardware resource information and the information of simulation tasks to be processed to obtain target resource analysis information and target task analysis information; the target resource analysis information includes several sequentially arranged pieces of first target resource information; the first target resource information includes a first resource serial number and M pieces of target resource available information; the target task analysis information includes several simulation task resource requirement information;

[0036] Perform matching processing on the target resource analysis information and the target task analysis information to obtain target matching result information.

[0037] It can improve the simulation efficiency and the utilization rate of simulation resources.

[0038] In the embodiments of the present application, the computer device 100 may be an independent server or a server network or server cluster composed of servers. For example, the computer device 100 described in the embodiments of the present application includes, but is not limited to, a computer, a network host, a single network server, a set of multiple network servers, or a cloud server composed of multiple servers. Among them, the cloud server is composed of a large number of computers or network servers based on cloud computing (Cloud Computing).

[0039] It can be understood that the computer device 100 used in the embodiments of the present application may be a device that includes both receiving and transmitting hardware, that is, a device having receiving and transmitting hardware capable of performing two-way communication on a two-way communication link. Such devices may include: cellular or other communication devices, which have a single-line display or a multi-line display or cellular or other communication devices without a multi-line display. Specifically, the computer device 100 may specifically be a desktop terminal or a mobile terminal, and the computer device 100 may specifically also be one of a mobile phone, a tablet computer, a laptop computer, etc.

[0040] Those skilled in the art can understand,Figure 1 The application environment shown is only one application scenario of the solution of this application, and does not constitute a limitation on the application scenarios of the solution of this application. Other application environments may also include more or fewer computer devices than those shown in Figure 1 For example, Figure 1 only 1 computer device is shown in . It can be understood that the parallel simulation system based on GPU resources may also include one or more other services, which are not specifically limited here.

[0041] In addition, as Figure 1 shown, the parallel simulation system based on GPU resources may also include a memory 200 for storing data, such as image data, location information, etc.

[0042] It should be noted that Figure 1 the schematic diagram of the scenario of the parallel simulation system based on GPU resources shown is only an example. The parallel simulation system and scenario described in the embodiments of this application are for more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. Those of ordinary skill in the art know that with the evolution of the parallel simulation system based on GPU resources and the emergence of new business scenarios, the technical solutions provided by the embodiments of this application are equally applicable to similar technical problems.

[0043] The present invention discloses a parallel simulation method and device based on GPU resources, which are beneficial to improving the simulation efficiency and the utilization rate of simulation resources. The following will be described in detail respectively.

[0044] Embodiment 1

[0045] Please refer to Figure 2 , Figure 2 which is a schematic flowchart of a parallel simulation method based on GPU resources disclosed in an embodiment of the present invention. Among them, Figure 2 the described parallel simulation method based on GPU resources is applied to a management system, such as a local server or a cloud server for management, etc., which is not limited in the embodiments of the present invention.

[0046] As Figure 2 shown, the parallel simulation method based on GPU resources may include the following operations:

[0047] 101. Obtain hardware resource information and.

[0048] In the embodiments of the present invention, the hardware resource information includes several node resource information; the node resource information includes M pieces of first resource information and second resource information corresponding to the first resource information; the first resource information includes first GPU resource information and second GPU resource information; the second resource information includes first storage resource information and second storage resource information.

[0049] 102. Analyze and process the hardware resource information and the simulation task information to be processed, and obtain target resource analysis information and target task analysis information.

[0050] In an embodiment of the present invention, the target resource analysis information includes a plurality of first target resource information arranged in sequence; the first target resource information includes a first resource serial number and M pieces of target resource availability information; the target task analysis information includes a plurality of simulation task resource requirement information.

[0051] 103. Perform matching processing on the target resource analysis information and the target task analysis information to obtain target matching result information.

[0052] It should be noted that the above M is an integer not less than 0, and the embodiments of the present invention do not make any limitations.

[0053] It should be noted that the above first resource information and the second resource information corresponding to the first resource information characterize the GPU resource situation and the video memory resource situation corresponding to the server node, and the embodiments of the present invention do not make any limitations.

[0054] It should be noted that the above first GPU resource information and the second GPU resource information respectively characterize the total amount of GPU resources and the available amount of GPU resources corresponding to a single hardware device, and the embodiments of the present invention do not make any limitations.

[0055] It should be noted that the above first storage resource information and the second storage resource information respectively characterize the total amount of video memory resources and the available amount of video memory resources corresponding to a single hardware device, and the embodiments of the present invention do not make any limitations.

[0056] It should be noted that the above target matching result information is used to indicate that the simulation tasks corresponding to the target matching result information are deployed in parallel on different GPUs for parallel simulation, so as to improve the utilization efficiency of hardware resources and the simulation efficiency, and the embodiments of the present invention do not make any limitations.

[0057] It can be seen that implementing the parallel simulation method based on GPU resources described in the embodiments of the present invention is beneficial to improving the simulation efficiency and the utilization rate of simulation resources.

[0058] In an optional embodiment, the above analyzing and processing the hardware resource information and the simulation task information to be processed to obtain target resource analysis information and target task analysis information includes:

[0059] Perform resource availability analysis and processing on the hardware resource information to obtain target resource analysis information;

[0060] Perform parsing processing on the simulation task information to be processed to obtain target task analysis information.

[0061] It should be noted that the above parsing and processing of the simulation task information to be processed can be based on a large model to decompose the parsing and processing of the simulation task information input by the user, so as to obtain the simulation task resource requirement information (task number, GPU resource requirement, video memory requirement, etc.) corresponding to multiple different simulation tasks, or it can be based on the parsing rules formulated by the user. The embodiments of the present invention do not make any limitations.

[0062] It can be seen that implementing the parallel simulation method based on GPU resources described in the embodiments of the present invention is beneficial to improving the simulation efficiency and the utilization rate of simulation resources.

[0063] In another optional embodiment, resource availability analysis and processing are performed on the hardware resource information to obtain target resource analysis information, including:

[0064] For any node resource information in the hardware resource information, the first resource calculation model is used to perform calculation processing on the node resource information to obtain the node resource analysis result value corresponding to the node resource information;

[0065] Among them, the first resource calculation model is:

[0066]

[0067] In the formula, JDFX represents the node resource analysis result value; A j1 and a i1 respectively represent the first GPU resource information corresponding to the j1-th first resource information and the second GPU resource information corresponding to the i1-th first resource information in the node resource information; B j2 and b i2 respectively represent the second storage resource information corresponding to the j2-th first resource information and the second storage resource information corresponding to the i2-th second resource information in the node resource information; x1 and x2 respectively represent the first calculation coefficient and the second calculation coefficient;

[0068] Sort all the node resource analysis result values from large to small to obtain the node resource analysis result information;

[0069] For any node resource analysis result value in the node resource analysis result information, based on the position serial number of the node resource analysis result value in the node resource analysis result information and the node resource information corresponding to the node resource analysis result value, the first target resource information corresponding to the node resource analysis result value is determined.

[0070] It should be noted that the above first calculation coefficient and second calculation coefficient are values between 0 and 1, and the sum of the two is 1. The embodiments of the present invention do not make any limitations.

[0071] It should be noted that the above calculation and processing of the node resource information using the first resource calculation model to obtain the node resource analysis result value corresponding to the node resource information is an analysis of the available situation of the overall hardware resources of the server node. Therefore, the GPU resources and video memory resources of the entire server node are used as the denominator, and then the sum of all available GPU resources and video memory resources is used as the numerator for calculation, so as to evaluate the available situation of the resources of the entire node, rather than summing after evaluating a single GPU hardware. This can avoid the problem of inaccurate overall resource evaluation caused by the differences in the GPU resources themselves, thereby improving the accuracy and efficiency of the evaluation of the overall node resources, and further improving the simulation efficiency and simulation resource utilization rate. The embodiments of the present invention do not make any limitations in this regard.

[0072] It should be noted that the above sorting of all node resource analysis result values from large to small to obtain the node resource analysis result information is to sort the server nodes according to resource availability, so that the servers with better overall resource availability can give priority to ensuring simulation tasks. Then, in the case where the guarantee ability of the previous nodes is insufficient, the nodes sorted later are used for supplementary guarantee to improve the efficiency of ensuring simulation tasks. The embodiments of the present invention do not make any limitations in this regard.

[0073] It can be seen that implementing the parallel simulation method based on GPU resources described in the embodiments of the present invention is beneficial to improving the simulation efficiency and simulation resource utilization rate.

[0074] In another optional embodiment, based on the position serial number of the node resource analysis result value in the node resource analysis result information and the node resource information corresponding to the node resource analysis result value, determining the first target resource information corresponding to the node resource analysis result value includes:

[0075] Determining the position serial number of the node resource analysis result value in the node resource analysis result information as the first resource serial number of the first target resource information corresponding to the node resource analysis result value;

[0076] For any first resource information in the node resource information corresponding to the node resource analysis result value, using a second resource calculation model to calculate and process the first resource information and the second resource information corresponding to the first resource information to obtain the target resource availability information corresponding to the first resource information;

[0077] Among them, the second resource calculation model is:

[0078]

[0079] Wherein, ZYKY represents the available information of the target resource; AA and aa respectively represent the first GPU resource information and the second GPU resource information in the first resource information; bb and BB respectively represent the first storage resource information and the second storage resource information in the second resource information; x3 and x4 respectively represent the third calculation coefficient and the fourth calculation coefficient.

[0080] It should be noted that the above third calculation coefficient and fourth calculation coefficient are values between 0 and 1, and the sum of the two is 1. The embodiments of the present invention do not make any limitations.

[0081] It should be noted that the above calculation and processing of the first resource information and the second resource information corresponding to the first resource information by using the second resource calculation model analyze the resource available status of the hardware device corresponding to a single GPU resource separately. Therefore, the single GPU resource amount and the video memory resource amount are used as the denominators for separate calculations, without considering the situation of the entire server, thereby improving the efficiency and accuracy of the availability analysis and evaluation of GPU resources. The embodiments of the present invention do not make any limitations.

[0082] It can be seen that implementing the parallel simulation method based on GPU resources described in the embodiments of the present invention is beneficial to improving the simulation efficiency and the utilization rate of simulation resources.

[0083] In another optional embodiment, the target resource analysis information and the target task analysis information are matched to obtain the target matching result information, including:

[0084] The simulation task resource requirement information in the target task analysis information is classified to obtain the target resource requirement information; the target resource requirement information includes the first target requirement information and the second target requirement information; the first target requirement information includes N sequentially arranged first sub-target requirement information; the second target requirement information includes L sequentially arranged second sub-target requirement information;

[0085] Judge whether both N and L are greater than 0 to obtain the zero value judgment result;

[0086] When the zero value judgment result is yes, based on the order of the first resource serial numbers from small to large, one target resource available information in the first target resource information is sequentially selected as the resource available information to be matched;

[0087] Based on the resource available information to be matched and the target resource requirement information, the target matching result information is determined;

[0088] When the zero value judgment result is no, end the processing flow corresponding to the zero value judgment result.

[0089] It should be noted that both N and L above are integers not less than 0, and their values can change dynamically. The embodiments of the present invention do not make any limitations.

[0090] It should be noted that the above first sub-goal requirement information indicates that the simulation task requires exclusive use of GPU resources. Further, the sorting of the above first sub-goal requirement information is based on the parsed simulation task numbers, which are not limited in the embodiments of the present invention.

[0091] It should be noted that the above second sub-goal requirement information indicates the need for multiple simulation tasks to be performed on the same GPU without the need to exclusively occupy GPU resources for simulation. Further, the sorting of the above second sub-goal requirement information is based on the parsed simulation task numbers, which are not limited in the embodiments of the present invention.

[0092] It can be seen that implementing the parallel simulation method based on GPU resources described in the embodiments of the present invention is beneficial to improving the simulation efficiency and the utilization rate of simulation resources.

[0093] In an alternative embodiment, based on the available information of the resources to be matched and the target resource requirement information, the target matching result information is determined, including:

[0094] Based on the magnitude relationship between the target resource available value corresponding to the target resource available information in the available information of the resources to be matched and 1, all the target resource available information in the available information of the resources to be matched is classified and sorted to obtain the first available information of the resources to be matched and the second available information of the resources to be matched; the first available information of the resources to be matched includes P first sub-resource available information distributed in sequence; the second available information of the resources to be matched includes Q second sub-resource available information distributed in sequence;

[0095] Based on the first available information of the resources to be matched, the second available information of the resources to be matched, and the target resource requirement information, the target matching result information is determined.

[0096] It should be noted that the above P and Q are integers not less than 0. Further, their values are dynamically changing, which are not limited in the embodiments of the present invention.

[0097] It should be noted that the above target resource available value represents the available state of the current hardware resources of the hardware device, and its value is a positive number greater than 0 and not greater than 1, which is not limited in the embodiments of the present invention. Further, when the target resource available value is 1, it indicates that the hardware device has not been used for the simulation task, and thus it is classified as the first sub-resource available information. If it is less than 1, it is classified as the second sub-resource available information to achieve a secondary classification of the hardware resources. So that when performing a simulation task later, if the simulation task requires exclusive use of GPU resources, it is selected from the first available information of the sub-resources, without the need to analyze and judge the second available information of the sub-resources, improving the efficiency and accuracy of resource allocation, which is not limited in the embodiments of the present invention.

[0098] It should be noted that the above first sub-resource available information is sorted according to the computing power of the corresponding GPU, and the embodiments of the present invention do not make any limitations in this regard.

[0099] It should be noted that the above second sub-resource available information is sorted from largest to smallest according to the available value of the target resource, and the embodiments of the present invention do not make any limitations in this regard. Further, the available value of the target resource represents the comprehensive available situation of the GPU and the video memory. Therefore, sorting from largest to smallest can give priority to using better hardware resources with available resources for the simulation task to ensure the efficient execution of the simulation task, and the embodiments of the present invention do not make any limitations in this regard.

[0100] It can be seen that implementing the parallel simulation method based on GPU resources described in the embodiments of the present invention is beneficial to improving the simulation efficiency and the utilization rate of simulation resources.

[0101] In another optional embodiment, based on the first to-be-matched resource available information, the second to-be-matched resource available information, and the target resource requirement information, the target matching result information is determined, including:

[0102] Judge whether P is greater than 0 to obtain the first numerical judgment result;

[0103] When the first numerical judgment result is yes, judge whether N is greater than 0 to obtain the second numerical judgment result;

[0104] When the second numerical judgment result is yes, sequentially select a first sub-target requirement information from the first target requirement information as the first to-be-processed requirement information, and delete the first sub-target requirement information corresponding to the first to-be-processed requirement information from the first target requirement information;

[0105] Judge whether the priority corresponding to the first to-be-processed requirement information is the first priority to obtain the priority judgment result;

[0106] When the priority judgment result is yes, sequentially select a first sub-resource available information from the first to-be-matched resource available information as the target resource task matching result information corresponding to the first to-be-processed requirement information;

[0107] Delete the first sub-resource available information corresponding to the target resource task matching result information from the first to-be-matched resource available information, and trigger the execution of judging whether P is greater than 0 to obtain the first numerical judgment result;

[0108] When the priority judgment result is no, inversely select a first sub-resource available information from the first to-be-matched resource available information as the target resource task matching result information corresponding to the first to-be-processed requirement information;

[0109] Delete the first sub-resource availability information corresponding to the target resource task matching result information from the first to-be-matched resource availability information, and trigger the execution of judging whether P is greater than 0 to obtain a first numerical judgment result;

[0110] When the second numerical judgment result is negative, judge whether L is greater than 0 to obtain a third numerical judgment result;

[0111] When the third numerical judgment result is positive, sequentially select a second sub-target requirement information from the second target requirement information as the second to-be-processed requirement information;

[0112] Judge whether there is a second sub-resource availability information that meets the resource matching condition in the second to-be-matched resource availability information to obtain an existence judgment result;

[0113] When the existence judgment result is positive, use the second sub-resource availability information with the earliest sorting in the resource matching condition as the target resource task matching result information corresponding to the second to-be-processed requirement information;

[0114] Delete the second sub-target requirement information corresponding to the second to-be-processed requirement information from the second target requirement information, and trigger the execution of judging whether L is greater than 0 to obtain a third numerical judgment result;

[0115] When the existence judgment result is negative, judge whether the second sub-target requirement information corresponding to the second to-be-processed requirement information is the last second sub-target requirement information in the second target requirement information to obtain a first sequence judgment result;

[0116] When the first sequence judgment result is negative, trigger the execution of sequentially selecting a second sub-target requirement information from the second target requirement information as the second to-be-processed requirement information;

[0117] When the first sequence judgment result is positive, trigger the execution of judging whether both N and L are greater than 0 to obtain a zero-value judgment result;

[0118] When the third numerical judgment result is negative, trigger the execution of judging whether both N and L are greater than 0 to obtain a zero-value judgment result;

[0119] When the first numerical judgment result is negative, trigger the execution of judging whether L is greater than 0 to obtain a third numerical judgment result.

[0120] It should be noted that judging whether P is greater than 0 first and then judging whether N is greater than 0 ensures that when there is a simulation task that requires exclusive GPU resources, the hardware node with unused GPUs is selected first. Otherwise, the next hardware node will be selected without further matching analysis between the first sub-target requirement information and the first sub-resource availability information, improving the efficiency and accuracy of resource allocation. The embodiments of the present invention are not limited.

[0121] It should be noted that determining whether to sequentially select GPUs with high resources or inversely select GPUs with relatively fewer resources to ensure simulation tasks according to the above priorities (the first priority indicates that the simulation task is an important simulation task that needs to be key guaranteed, and if it is not the first priority, hardware with relatively fewer GPUs can be used to guarantee, which improves the reliability of the parallel simulation of the simulation task by hardware resources and also improves the effective utilization rate of resources) can improve the efficiency and accuracy of parallel simulation resource allocation. The embodiments of the present invention do not make any limitations in this regard.

[0122] It should be noted that the above resource matching conditions indicate that the GPU resource amount corresponding to the second sub-resource available information is greater than the GPU resource amount required by the simulation task, and the video memory resource amount corresponding to the second sub-resource available information is greater than the video memory resource amount required by the simulation task. The embodiments of the present invention do not make any limitations in this regard.

[0123] It can be seen that implementing the parallel simulation method based on GPU resources described in the embodiments of the present invention is beneficial to improving the simulation efficiency and the utilization rate of simulation resources.

[0124] Embodiment 2

[0125] Please refer to Figure 3 , Figure 3 which is a schematic structural diagram of a parallel simulation device based on GPU resources disclosed in the embodiments of the present invention. Among them, Figure 3 the described device can be applied to a management system, such as a local server or a cloud server for management, etc. The embodiments of the present invention do not make any limitations in this regard. As Figure 3 shown, the device may include:

[0126] An acquisition module 201, configured to acquire hardware resource information and information of simulation tasks to be processed; the hardware resource information includes several node resource information; the node resource information includes M pieces of first resource information and second resource information corresponding to the first resource information; the first resource information includes first GPU resource information and second GPU resource information; the second resource information includes first storage resource information and second storage resource information;

[0127] A first processing module 202, configured to analyze and process the hardware resource information and the information of simulation tasks to be processed to obtain target resource analysis information and target task analysis information; the target resource analysis information includes several sequentially arranged first target resource information; the first target resource information includes a first resource serial number and M pieces of target resource available information; the target task analysis information includes several simulation task resource requirement information;

[0128] The second processing module 203 is configured to perform matching processing on the target resource analysis information and the target task analysis information to obtain target matching result information; the target matching result information includes a plurality of target resource task matching result information.

[0129] It can be seen that implementing Figure 3 the described parallel simulation device based on GPU resources is beneficial to improving simulation efficiency and simulation resource utilization rate.

[0130] In another optional embodiment, as Figure 3 shown, analyzing and processing the hardware resource information and the simulation task information to be processed to obtain target resource analysis information and target task analysis information, including:

[0131] Performing resource availability analysis processing on the hardware resource information to obtain target resource analysis information;

[0132] Performing parsing processing on the simulation task information to be processed to obtain target task analysis information.

[0133] It can be seen that implementing Figure 3 the described parallel simulation device based on GPU resources is beneficial to improving simulation efficiency and simulation resource utilization rate.

[0134] In yet another optional embodiment, as Figure 3 shown, performing resource availability analysis processing on the hardware resource information to obtain target resource analysis information, including:

[0135] For any node resource information in the hardware resource information, using the first resource calculation model to perform calculation processing on the node resource information to obtain a node resource analysis result value corresponding to the node resource information;

[0136] Among them, the first resource calculation model is:

[0137]

[0138] In the formula, JDFX represents the node resource analysis result value; A j1 and a i1 respectively represent the first GPU resource information corresponding to the j1-th first resource information and the second GPU resource information corresponding to the i1-th first resource information in the node resource information; B j2 and b i2 respectively represent the second storage resource information corresponding to the j2-th first resource information and the second storage resource information corresponding to the i2-th second resource information in the node resource information; x1 and x2 respectively represent the first calculation coefficient and the second calculation coefficient;

[0139] Sort the analysis result values of all node resources from largest to smallest to obtain node resource analysis result information;

[0140] For any node resource analysis result value in the node resource analysis result information, based on the position serial number of the node resource analysis result value in the node resource analysis result information and the node resource information corresponding to the node resource analysis result value, determine the first target resource information corresponding to the node resource analysis result value.

[0141] It can be seen that implementing Figure 3 the described parallel simulation device based on GPU resources is beneficial to improving simulation efficiency and simulation resource utilization rate.

[0142] In another optional embodiment, as Figure 3 shown, based on the position serial number of the node resource analysis result value in the node resource analysis result information and the node resource information corresponding to the node resource analysis result value, determining the first target resource information corresponding to the node resource analysis result value includes:

[0143] Determine the position serial number of the node resource analysis result value in the node resource analysis result information as the first resource serial number of the first target resource information corresponding to the node resource analysis result value;

[0144] For any first resource information in the node resource information corresponding to the node resource analysis result value, use the second resource calculation model to perform calculation processing on the first resource information and the second resource information corresponding to the first resource information to obtain the target resource available information corresponding to the first resource information;

[0145] Wherein, the second resource calculation model is:

[0146]

[0147] In the formula, ZYKY represents the target resource available information; AA and aa respectively represent the first GPU resource information and the second GPU resource information in the first resource information; bb and BB respectively represent the first storage resource information and the second storage resource information in the second resource information; x3 and x4 respectively represent the third calculation coefficient and the fourth calculation coefficient.

[0148] It can be seen that implementing Figure 3 the described parallel simulation device based on GPU resources is beneficial to improving simulation efficiency and simulation resource utilization rate.

[0149] In another optional embodiment, as Figure 3 shown, perform matching processing on the target resource analysis information and the target task analysis information to obtain target matching result information, including:

[0150] Classify the simulation task resource requirement information in the target task analysis information to obtain the target resource requirement information; the target resource requirement information includes the first target requirement information and the second target requirement information; the first target requirement information includes N sequentially arranged first sub-target requirement information; the second target requirement information includes L sequentially arranged second sub-target requirement information;

[0151] Judge whether both N and L are greater than 0 to obtain a zero value judgment result;

[0152] When the zero value judgment result is yes, based on the order of the first resource serial numbers from small to large, sequentially select one target resource available information in the first target resource information as the to-be-matched resource available information;

[0153] Based on the to-be-matched resource available information and the target resource requirement information, determine the target matching result information;

[0154] When the zero value judgment result is no, end the processing flow corresponding to the zero value judgment result.

[0155] It can be seen that implementing Figure 3 the described parallel simulation device based on GPU resources is beneficial to improving simulation efficiency and simulation resource utilization rate.

[0156] In another optional embodiment, as Figure 3 shown, based on the to-be-matched resource available information and the target resource requirement information, determine the target matching result information, including:

[0157] Based on the size relationship between the target resource available value corresponding to the target resource available information in the to-be-matched resource available information and 1, classify and sort all the target resource available information in the to-be-matched resource available information to obtain the first to-be-matched resource available information and the second to-be-matched resource available information; the first to-be-matched resource available information includes P sequentially distributed first sub-resource available information; the second to-be-matched resource available information includes Q sequentially distributed second sub-resource available information;

[0158] Based on the first to-be-matched resource available information, the second to-be-matched resource available information and the target resource requirement information, determine the target matching result information.

[0159] It can be seen that implementing Figure 3 the described parallel simulation device based on GPU resources is beneficial to improving simulation efficiency and simulation resource utilization rate.

[0160] In another optional embodiment, as Figure 3 shown, based on the first to-be-matched resource available information, the second to-be-matched resource available information and the target resource requirement information, determine the target matching result information, including:

[0161] Judge whether P is greater than 0 to obtain the first numerical judgment result;

[0162] When the first numerical judgment result is yes, judge whether N is greater than 0 to obtain the second numerical judgment result;

[0163] When the second numerical judgment result is yes, sequentially select a first sub-goal requirement information from the first target requirement information as the first requirement information to be processed, and delete the corresponding first sub-goal requirement information from the first target requirement information;

[0164] Judge whether the priority corresponding to the first requirement information to be processed is the first priority to obtain the priority judgment result;

[0165] When the priority judgment result is yes, sequentially select a first sub-resource availability information from the first available resource information to be matched as the target resource task matching result information corresponding to the first requirement information to be processed;

[0166] Delete the corresponding first sub-resource availability information of the target resource task matching result information from the first available resource information to be matched, and trigger the execution of judging whether P is greater than 0 to obtain the first numerical judgment result;

[0167] When the priority judgment result is no, select a first sub-resource availability information from the first available resource information to be matched in reverse order as the target resource task matching result information corresponding to the first requirement information to be processed;

[0168] Delete the corresponding first sub-resource availability information of the target resource task matching result information from the first available resource information to be matched, and trigger the execution of judging whether P is greater than 0 to obtain the first numerical judgment result;

[0169] When the second numerical judgment result is no, judge whether L is greater than 0 to obtain the third numerical judgment result;

[0170] When the third numerical judgment result is yes, sequentially select a second sub-goal requirement information from the second target requirement information as the second requirement information to be processed;

[0171] Judge whether there is a second sub-resource availability information that meets the resource matching condition in the second available resource information to be matched to obtain the existence judgment result;

[0172] When the existence judgment result is yes, use the second sub-resource availability information with the earliest sorting in the resource matching condition as the target resource task matching result information corresponding to the second requirement information to be processed;

[0173] Delete the second sub-goal requirement information corresponding to the second to-be-processed requirement information from the second target requirement information, and trigger the execution of determining whether L is greater than 0 to obtain a third numerical judgment result;

[0174] When there is a judgment result of no, determine whether the second sub-goal requirement information corresponding to the second to-be-processed requirement information is the last second sub-goal requirement information in the second target requirement information to obtain a first sequence judgment result;

[0175] When the first sequence judgment result is no, trigger the execution of sequentially selecting a second sub-goal requirement information from the second target requirement information as the second to-be-processed requirement information;

[0176] When the first sequence judgment result is yes, trigger the execution of determining whether both N and L are greater than 0 to obtain a zero-value judgment result;

[0177] When the third numerical judgment result is no, trigger the execution of determining whether both N and L are greater than 0 to obtain a zero-value judgment result;

[0178] When the first numerical judgment result is no, trigger the execution of determining whether L is greater than 0 to obtain a third numerical judgment result.

[0179] It can be seen that implementing Figure 3 the parallel simulation device based on GPU resources described is beneficial to improving simulation efficiency and simulation resource utilization rate.

[0180] Embodiment III

[0181] Please refer to Figure 4 , Figure 4 which is a schematic structural diagram of another parallel simulation device based on GPU resources disclosed in an embodiment of the present invention. Among them, Figure 4 the described device can be applied to a management system, such as a local server or a cloud server for management, etc., which is not limited in the embodiments of the present invention. As Figure 4 shown, the device may include:

[0182] A memory 301 storing executable program code;

[0183] A processor 302 coupled to the memory 301;

[0184] The processor 302 calls the executable program code stored in the memory 301 to execute the steps in the parallel simulation method based on GPU resources described in Embodiment I.

[0185] Embodiment IV

[0186] An embodiment of the present invention discloses a computer-readable storage medium that stores a computer program for electronic data exchange. Among them, the computer program enables a computer to execute the steps in the parallel simulation method based on GPU resources described in Embodiment 1.

[0187] Embodiment 5

[0188] An embodiment of the present invention discloses a computer program product. The computer program product includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to enable a computer to execute the steps in the parallel simulation method based on GPU resources described in Embodiment 1.

[0189] The device embodiments described above are merely illustrative. The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed to multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. A person of ordinary skill in the art can understand and implement it without creative labor.

[0190] Through the specific descriptions of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the above technical solution, in essence, or the part that contributes to the prior art can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, and the storage medium includes a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electrically-erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc memories, a magnetic disk memory, a tape memory, or any other computer-readable medium capable of carrying or storing data.

[0191] Finally, it should be noted that: The parallel simulation method and device based on GPU resources disclosed in the embodiments of the present invention only disclose the preferred embodiments of the present invention, and are only used to illustrate the technical solutions of the present invention, rather than limiting them; Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; And these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A parallel simulation method based on GPU resources, characterized in that: The method comprises: Acquire hardware resource information and to-be-processed simulation task information; the hardware resource information includes a plurality of node resource information; the node resource information includes M first resource information and second resource information corresponding to the first resource information; the first resource information includes first GPU resource information and second GPU resource information; the second resource information includes first storage resource information and second storage resource information; Analyze and process the hardware resource information and the to-be-processed simulation task information to obtain target resource analysis information and target task analysis information; the target resource analysis information includes a plurality of first target resource information arranged in sequence; the first target resource information includes a first resource sequence number and M target resource available information; the target task analysis information includes a plurality of simulation task resource requirement information; The target resource analysis information and the target task analysis information are matched to obtain target matching result information.

2. The parallel simulation method based on GPU resources according to claim 1, characterized in that: The analyzing and processing the hardware resource information and the to-be-processed simulation task information to obtain target resource analysis information and target task analysis information includes: Performing resource availability analysis on the hardware resource information to obtain target resource analysis information; The simulation task information to be processed is analyzed to obtain target task analysis information.

3. The parallel simulation method based on GPU resources according to claim 2, characterized in that: The performing resource availability analysis processing on the hardware resource information to obtain target resource analysis information includes: For any of the node resource information in the hardware resource information, use the first resource calculation model to calculate and process the node resource information to obtain a node resource analysis result value corresponding to the node resource information; Wherein, the first resource calculation model is: Wherein, JDFX represents the value of the node resource analysis result; A j1 and a i1 respectively represent the first GPU resource information corresponding to the j1th first resource information in the node resource information and the second GPU resource information corresponding to the i1th first resource information; B j2 and b i2 Respectively represent the second storage resource information corresponding to the j2th first resource information in the node resource information and the second storage resource information corresponding to the i2th second resource information; x1 and x2 represent the first calculation coefficient and the second calculation coefficient respectively; Sort all the node resource analysis result values ​​from large to small to obtain node resource analysis result information; For any of the node resource analysis result values ​​in the node resource analysis result information, the first target resource information corresponding to the node resource analysis result value is determined based on the position serial number of the node resource analysis result value in the node resource analysis result information and the node resource information corresponding to the node resource analysis result value.

4. The parallel simulation method based on GPU resources according to claim 3, characterized in that: The determining, based on the position sequence number of the node resource analysis result value in the node resource analysis result information and the node resource information corresponding to the node resource analysis result value, the first target resource information corresponding to the node resource analysis result value comprises: Determine the position sequence number of the node resource analysis result value in the node resource analysis result information as the first resource sequence number of the first target resource information corresponding to the node resource analysis result value; For any of the first resource information in the node resource information corresponding to the node resource analysis result value, use the second resource calculation model to calculate and process the first resource information and the second resource information corresponding to the first resource information to obtain the target resource availability information corresponding to the first resource information; Among them, the second resource calculation model is: In the formula, ZYKY represents the target resource available information; AA and aa represent the first GPU resource information and the second GPU resource information in the first resource information respectively; bb and BB represent the first storage resource information and the second storage resource information in the second resource information respectively; x3 and x4 represent the third calculation coefficient and the fourth calculation coefficient respectively.

5. The parallel simulation method based on GPU resources according to claim 1, characterized in that: The matching process of the target resource analysis information and the target task analysis information to obtain target matching result information includes: Classify the simulation task resource requirement information in the target task analysis information to obtain target resource requirement information; the target resource requirement information includes first target requirement information and second target requirement information; the first target requirement information includes N first sub-target requirement information arranged in sequence; the second target requirement information includes L second sub-target requirement information arranged in sequence; Determine whether both N and L are greater than 0, and obtain a zero value determination result; When the zero value judgment result is yes, based on the ascending order of the first resource sequence number, one of the target resource available information in the first target resource information is selected in sequence as the resource available information to be matched; Determining target matching result information based on the available information of the to-be-matched resources and the target resource demand information; When the zero value judgment result is no, the processing flow corresponding to the zero value judgment result is terminated.

6. The parallel simulation method based on GPU resources according to claim 5, characterized in that: The determining target matching result information based on the available information of the to-be-matched resources and the target resource demand information includes: Based on the size relationship between the target resource available value corresponding to the target resource available information in the available resource information to be matched and 1, all the target resource available information in the available resource information to be matched are classified and sorted to obtain first available resource information to be matched and second available resource information to be matched; the first available resource information to be matched includes P first sub-resource available information distributed in sequence; the second available resource information to be matched includes Q second sub-resource available information distributed in sequence; Target matching result information is determined based on the first available information of resources to be matched, the second available information of resources to be matched, and the target resource demand information.

7. The parallel simulation method based on GPU resources according to claim 6, characterized in that: The determining target matching result information based on the first available information of resources to be matched, the second available information of resources to be matched and the target resource requirement information includes: Determine whether P is greater than 0, and obtain a first numerical determination result; When the first numerical judgment result is yes, determine whether N is greater than 0 to obtain a second numerical judgment result; When the second value judgment result is yes, one of the first sub-target demand information is selected from the first target demand information in sequence as the first demand information to be processed, and the first sub-target demand information corresponding to the first demand information to be processed is deleted from the first target demand information; Determine whether the priority corresponding to the first to-be-processed demand information is the first priority, and obtain a priority determination result; When the priority judgment result is yes, sequentially selecting one of the first sub-resource available information from the first available resource information to be matched as the target resource task matching result information corresponding to the first demand information to be processed; The first sub-resource available information corresponding to the target resource task matching result information is deleted from the first available information of resources to be matched, and the judgment of whether P is greater than 0 is triggered to obtain a first numerical judgment result; When the priority judgment result is no, selecting one of the first sub-resource available information from the first available resource information to be matched in reverse order as the target resource task matching result information corresponding to the first demand information to be processed; The first sub-resource available information corresponding to the target resource task matching result information is deleted from the first available information of resources to be matched, and the judgment of whether P is greater than 0 is triggered to obtain a first numerical judgment result; When the second numerical judgment result is no, determining whether L is greater than 0, and obtaining a third numerical judgment result; When the third value judgment result is yes, selecting one of the second sub-target demand information from the second target demand information in sequence as the second demand information to be processed; Determine whether there is the second sub-resource available information satisfying the resource matching condition in the second available resource information to be matched, and obtain an existence determination result; When the existence judgment result is yes, the available information of the second sub-resource that satisfies the resource matching condition and ranks the highest is used as the target resource task matching result information corresponding to the second to-be-processed demand information; The second sub-target demand information corresponding to the second demand information to be processed is deleted from the second target demand information, and the judgment of whether L is greater than 0 is triggered to obtain a third numerical judgment result; When the existence judgment result is no, determining whether the second sub-target demand information corresponding to the second demand information to be processed is the last second sub-target demand information in the second target demand information, and obtaining a first sequence judgment result; When the first sequence judgment result is no, triggering the execution of selecting one of the second sub-target demand information from the second target demand information in sequence as the second demand information to be processed; When the first sequence judgment result is yes, triggering the execution of the judgment of whether both N and L are greater than 0, and obtaining a zero value judgment result; When the third value judgment result is no, triggering the execution of the judgment of whether both N and L are greater than 0, and obtaining a zero value judgment result; When the first numerical judgment result is no, the judgment of whether L is greater than 0 is triggered to obtain a third numerical judgment result.

8. A parallel simulation device based on GPU resources, characterized in that: The device comprises: An acquisition module is used to acquire hardware resource information and information of simulation tasks to be processed; the hardware resource information includes information of several nodes; the node resource information includes M first resource information and second resource information corresponding to the first resource information; the first resource information includes first GPU resource information and second GPU resource information; the second resource information includes first storage resource information and second storage resource information; The first processing module is used to analyze and process the hardware resource information and the simulation task information to be processed to obtain target resource analysis information and target task analysis information; the target resource analysis information includes a plurality of first target resource information arranged in sequence; the first target resource information includes a first resource sequence number and M target resource available information; the target task analysis information includes a plurality of simulation task resource demand information; The second processing module is used to match the target resource analysis information and the target task analysis information to obtain target matching result information; the target matching result information includes a plurality of target resource task matching result information.

9. A parallel simulation device based on GPU resources, characterized in that: The device comprises: A memory storing executable program code; a processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the GPU resource-based parallel simulation method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and when the computer instructions are called, they are used to execute the parallel simulation method based on GPU resources as described in any one of claims 1-7.

Images