Chip simulation method and device for molecular dynamics calculation, equipment and storage medium

By decomposing the task into subtasks in molecular dynamics calculations and matching it with the chip, the problems of high-performance chips are solved, efficient development and testing are achieved, and hardware costs are reduced.

CN119004825APending Publication Date: 2024-11-22SHANGHAI SMARTLOGIC TECHNOLOGY LTD
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411097486.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-09
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

In large-scale molecular dynamics computing, the expensive cost of high-performance chips and long development and verification cycles make early large-scale chip deployments impractical, resulting in high hardware costs and maintenance costs and inefficient development.

Method used

By determining the simulation parameters, initializing the simulation environment, loading the molecular dynamics simulation program onto the chip, decomposing the calculation task into subtasks and matching the chip, and performing the subtasks, obtaining the execution results of the subtasks, and using parallel computing power to improve efficiency.

Benefits of technology

It shortens the software development cycle, reduces test time, reduces hardware cost requirements, improves development efficiency, realizes the full-scale chip function of some chips, and reduces the actual chip demand.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119004825A_ABST
    Figure CN119004825A_ABST
Patent Text Reader

Abstract

The invention provides a chip simulation method and device for molecular dynamics calculation, equipment and a storage medium, and the method comprises the steps: determining simulation parameters, and initializing a simulation environment according to the simulation parameters; loading a molecular dynamics simulation program on the chip; decomposing a molecular dynamics calculation task into subtasks, matching the subtasks with chips, and loading the subtasks to the matched chips; executing the subtask; and obtaining an execution result of the subtask. According to the method provided by the invention, the simulation parameters are determined, the simulation environment is initialized according to the simulation parameters, the molecular dynamics simulation program is loaded to the chip, the molecular dynamics calculation task is decomposed into the sub-tasks, the sub-tasks are matched with the chip, the sub-tasks are loaded to the matched chip, and the sub-tasks are executed. According to the method, the software development cycle is shortened, the test time is shortened, and the development efficiency is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of molecular dynamics technology, and in particular to a chip simulation method, device, equipment, and storage medium for molecular dynamics calculations. Background Art

[0002] In large-scale molecular dynamics calculations, a large amount of computing resources are usually required to process complex molecular trajectory calculations. These computing resources are often composed of high-performance chips that can process large amounts of data at extremely high speeds.

[0003] However, chips are expensive to manufacture, and the more chips involved in a calculation, the higher the hardware cost and maintenance fee required.

[0004] In addition, developing and verifying the functions of chips in actual applications also requires a long period of time and a large amount of resource investment, which makes large-scale chip deployment in the early stages impractical. Summary of the invention

[0005] In order to solve one of the above-mentioned technical defects, the present application provides a chip simulation method, device, equipment, and storage medium for molecular dynamics calculations.

[0006] In a first aspect, the present application provides a chip simulation method for molecular dynamics calculation, the method comprising:

[0007] Determine simulation parameters, and initialize the simulation environment according to the simulation parameters;

[0008] Loading a molecular dynamics simulation program onto the chip;

[0009] Decompose the molecular dynamics calculation task into subtasks, match the subtasks with chips, and load the subtasks onto the matching chips;

[0010] Execute subtasks;

[0011] Get the execution result of the subtask.

[0012] Optionally, the simulation parameters include: force field parameters, time step, number of participating chips, spatial grid size, and generation of random tasks.

[0013] Optionally, the time step is between 1-2 femtoseconds.

[0014] Optionally, the molecular dynamics calculation task is decomposed into subtasks and the subtasks are matched to the chip, including:

[0015] Decompose the molecular dynamics calculation task into multiple subtasks according to the number of participating chips and chip characteristics;

[0016] Subtasks are matched with chips according to the number of participating chips and chip characteristics.

[0017] Optionally, the chip is physically present;

[0018] Load the molecular dynamics simulation program onto the chip, including:

[0019] Transferring molecular dynamics simulation programs to chip memory via a high-bandwidth interface;

[0020] Load subtasks onto matching chips, including:

[0021] The subtasks are transferred to matching on-chip memory via a high-bandwidth interface.

[0022] Optionally, the chip is physically present;

[0023] Perform subtasks, including:

[0024] According to the subtask, start the molecular dynamics simulation program;

[0025] The chip's parallel computing capabilities are used to simulate the execution of subtasks through a molecular dynamics simulation program.

[0026] Optionally, the chip is physically present;

[0027] Get the execution results of the subtask, including:

[0028] When the chip executes the subtask, the calculation data files continuously generated by the chip are transmitted to the shared storage device; wherein the file name of the calculation data file carries the chip identification and the calculation round.

[0029] In a second aspect of the present application, a chip simulation device for molecular dynamics calculation is provided, the device comprising:

[0030] A simulation module is used to determine simulation parameters and initialize the simulation environment according to the simulation parameters;

[0031] A first loading module, used for loading a molecular dynamics simulation program onto the chip;

[0032] The second loading module is used to decompose the molecular dynamics calculation task into subtasks, match the subtasks with chips, and load the subtasks onto the matching chips;

[0033] An execution module, used for executing the subtask loaded by the second loading module;

[0034] The acquisition module is used to obtain the execution results of the subtasks obtained by the execution module.

[0035] In a third aspect of the present application, an electronic device is provided, including:

[0036] Memory;

[0037] Processor; and

[0038] Computer programs;

[0039] The computer program is stored in the memory and is configured to be executed by the processor to implement the method as described in the first aspect above.

[0040] In a fourth aspect of the present application, a computer-readable storage medium is provided, on which a computer program is stored; the computer program is executed by a processor to implement the method described in the first aspect above.

[0041] The present application provides a chip simulation method, device, equipment, and storage medium for molecular dynamics calculations, the method comprising: determining simulation parameters, initializing a simulation environment according to the simulation parameters; loading a molecular dynamics simulation program onto a chip; decomposing a molecular dynamics calculation task into subtasks, matching the subtasks with chips, and loading the subtasks onto matching chips; executing subtasks; and obtaining the execution results of subtasks. The method provided in the present application can shorten the software development cycle, reduce testing time, and improve development efficiency by determining simulation parameters, initializing a simulation environment according to the simulation parameters, loading a molecular dynamics simulation program onto a chip, decomposing a molecular dynamics calculation task into subtasks, matching the subtasks with chips, and loading the subtasks onto matching chips, executing subtasks, and obtaining the execution results of subtasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0043] Figure 1 A schematic diagram of a flow chart of a chip simulation method for molecular dynamics calculation provided in an embodiment of the present application;

[0044] Figure 2 A schematic diagram of an implementation of determining simulation parameters and initializing a simulation environment according to the simulation parameters provided in an embodiment of the present application;

[0045] Figure 3 A schematic diagram of a grid structure provided in an embodiment of the present application;

[0046] Figure 4 A schematic diagram of an implementation of loading a molecular dynamics simulation program onto a chip provided in an embodiment of the present application;

[0047] Figure 5A schematic diagram of an implementation of obtaining the execution result of a subtask provided in an embodiment of the present application;

[0048] Figure 6 A schematic diagram of the structure of a chip simulation device for molecular dynamics calculation provided in an embodiment of the present application;

[0049] Figure 7 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0050] In order to make the technical solutions and advantages in the embodiments of the present application more clearly understood, the exemplary embodiments of the present application are further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than an exhaustive list of all the embodiments. It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict.

[0051] In the process of implementing the present application, the inventors found that in large-scale molecular dynamics calculations, a large amount of computing resources are usually required to process complex molecular trajectory calculations. These computing resources are often composed of high-performance chips that can process large amounts of data at extremely high speeds. However, chips are expensive to manufacture, and the more chips that need to be involved in the calculation, the higher the hardware cost and maintenance costs required. In addition, the development and verification of chip functions in actual applications also requires a long period of time and a large amount of resource investment. This makes large-scale chip deployment impractical in the early stages.

[0052] In response to the above problems, the present application provides a chip simulation method, device, equipment, and storage medium for molecular dynamics calculations, the method comprising: determining simulation parameters, initializing the simulation environment according to the simulation parameters; loading the molecular dynamics simulation program onto the chip; decomposing the molecular dynamics calculation task into subtasks, matching the subtasks with the chip, and loading the subtasks onto the matching chip; executing the subtasks; and obtaining the execution results of the subtasks. The method provided in the present application can shorten the software development cycle, reduce testing time, and improve development efficiency by determining simulation parameters, initializing the simulation environment according to the simulation parameters, loading the molecular dynamics simulation program onto the chip, decomposing the molecular dynamics calculation task into subtasks, matching the subtasks with the chip, and loading the subtasks onto the matching chip, executing the subtasks, and obtaining the execution results of the subtasks.

[0053] See also Figure 1 This embodiment provides a chip simulation method for molecular dynamics calculation. The main body of the chip simulation party for molecular dynamics calculation provided by this embodiment can be a central control system, and its implementation process is as follows:

[0054] 101, determine simulation parameters, and initialize the simulation environment according to the simulation parameters.

[0055] Among them, the simulation parameters include: force field parameters, time step, number of participating chips, spatial grid size, and generation of random tasks.

[0056] For example, the time step is between 1-2 femtoseconds (fs).

[0057] Step 101 implements the configuration simulation process. Figure 2 As shown, the configuration simulation process is the basic process of the entire molecular dynamics calculation, which mainly involves setting simulation parameters and initializing the simulation environment.

[0058] 1. Set simulation parameters

[0059] (1) Force field parameters

[0060] The force field parameters define the force field model used for intermolecular interactions, including potential functions (such as Lennard-Jones potential, Coulomb potential) and related parameters (such as atomic charges, bond constants).

[0061] (2) Time step

[0062] The time step is the length of the time step. Determining the length of each time step during the simulation can ensure the stability and accuracy of the calculation.

[0063] (3) Number of participating chips

[0064] The number of participating chips is the actual number of chips used to perform molecular dynamics calculations in order to reasonably distribute the computing tasks.

[0065] (4) Spatial grid size

[0066] Setting the spatial grid size can define the grid size (length, width, and height) of the simulation space, which is used to discretize the computational domain.

[0067] The spatial grid size can also be adjusted according to the size of the simulation system, the particle density, and the desired spatial resolution.

[0068] (5) Generate random tasks

[0069] The process of generating a random task is as follows: the central control system generates a random task through a random number, and combines the above parameters as input to generate a task ID, which matches an executable program capable of performing molecular dynamics as a simulation calculation program.

[0070] 2. Initialize the simulation environment

[0071] (1) Setting force field parameters and step size

[0072] For example, the force field parameters and time step in the simulation parameters are designed accordingly so that the force field parameters and time step meet the simulation requirements.

[0073] (2) Spatial grid construction

[0074] Construct the grid structure of the computational space based on parameter information (such as spatial grid size), such as Figure 3 As shown, the chip labeled 1 is the actual chip, and the chip labeled 0 is the simulated chip. The grid is used to discretize the simulation space so that the calculation can be performed in the discrete space, which is convenient for parallel calculation.

[0075] (3) Chip identification

[0076] Distinguish between actual chips and simulated chips, such as using is_present to identify the chips that actually participate in the calculation, with the actual chip identified as true and the simulated chip identified as false. This makes it clear which calculations are performed by real chips and which are performed by simulated chips, thereby reasonably allocating computing resources and ensuring the accuracy and efficiency of the simulation.

[0077] For example, users can manually fill in parameter configurations on the configuration page. The page combines the parameters into a json format message and sends it to the central control system. The central control system initializes and generates a spatial grid for subsequent calculations by inputting the number of participating chips (including actual chips and simulated chips), length, width and height. A task ID is generated by generating a random number and the force field parameters and time step of the input parameters to match the corresponding molecular dynamics program. For example, if there are 1,000 chips participating in the calculation, a 10*10*10 spatial grid is generated, and the random number ID generated is 89. The matching molecular dynamics program is compressed together with the file written in the configuration to generate 1,000 compressed packages, named 0.tar.gz->999.tar.gz, matching the ID of the spatial grid coordinates, for example, the ID of the spatial grid coordinates [2,3,4] = 2*10^0+3*10^1+4*10^2 is 432.

[0078] 102, loading a molecular dynamics simulation program onto the chip.

[0079] Step 102 can load the molecular dynamics simulation program onto the chip so that subsequent tasks can be executed, thereby realizing the simulation process of sending the calculation program.

[0080] If the chip exists physically, the process of loading the molecular dynamics simulation program onto the chip is: transferring the molecular dynamics simulation program to the chip memory through a high-bandwidth interface.

[0081] After completing step 101, the central control module switches the state to the computing program issuing, and uses the communication interface and the transmission protocol ftp / scp to transmit the tar.gz compressed package generated in step 101 to the default designated location of the chip.

[0082] like Figure 4 As shown, the program is issued: Figure 4 The program is issued as shown, and the molecular dynamics simulation program is loaded onto the chip.

[0083] For an actual chip (i.e. a chip that physically exists), the molecular dynamics simulation program can be transferred to the chip memory via a high-bandwidth interface (such as PCIe or Ethernet).

[0084] For the simulated chip, the actual transmission operation is not performed, but the program is directly marked as successfully transmitted to simulate the transmission process. The transmission simulation of the simulated chip can ignore the actual network delay and bandwidth limitation.

[0085] 103, decompose the molecular dynamics calculation task into subtasks, match the subtasks with chips, and load the subtasks onto the matching chips.

[0086] When implementing step 103, the molecular dynamics calculation task can be first decomposed into multiple subtasks according to the number of participating chips and chip characteristics. Then, the subtasks are matched with the chips according to the number of participating chips and chip characteristics. Finally, the subtasks are loaded onto the matching chips.

[0087] If the chip exists physically, the process of loading the subtask onto the matching chip is as follows: the subtask is transmitted to the matching chip memory via a high-bandwidth interface.

[0088] Step 103 implements the simulation of issuing molecular dynamics calculation tasks.

[0089] The simulation of computing task dispatching includes three parts: task decomposition, task scheduling and task transmission, to ensure that molecular dynamics computing tasks can be efficiently allocated and executed.

[0090] 1. Task decomposition (i.e. decomposing the molecular dynamics calculation task into subtasks)

[0091] Task decomposition: Based on the number and characteristics of the chips involved in the calculation, as well as the overall scale and computing requirements of the task (e.g., a 100w system), the algorithm program is called to decompose the molecular dynamics calculation task into multiple subtasks. The final product of each task is a compressed package. The subtask is named 1_task.tar.gz. The naming rule is consistent with step 101. For example, if there are 1000 chips involved in the calculation, 1000 subtask compressed packages are generated, named 0_task.tar.gz->999_task.tar.gz, so that it can be processed in parallel on the chip. The goal of task decomposition is to improve computing efficiency and resource utilization.

[0092] 2. Task scheduling (matching subtasks with chips)

[0093] Task scheduling can match the decomposed subtasks with the actual chips according to the number and characteristics of the chips involved in the calculation. The matching rule is that the ID of the subtask is consistent with the spatial grid coordinate ID of the chip.

[0094] In addition, the scheduling rules are dynamically adjusted according to the number and computing power of chips to ensure balanced distribution of tasks and maximize the use of computing resources.

[0095] 3. Task transfer (i.e. loading subtasks onto matching chips)

[0096] For an actual chip (i.e., a chip that physically exists), the subtasks can be transferred to the matching chip memory via a high-bandwidth interface (such as PCIe or Ethernet).

[0097] For the simulated chip, the actual transmission operation is not performed, but the task is directly marked as successfully transmitted to simulate the task transmission process. The task transmission simulation of the simulated chip can ignore the actual data transmission overhead and potential network failures.

[0098] 104, execute subtask.

[0099] If the chip exists physically, the process of executing the subtask is as follows: according to the subtask, start the molecular dynamics simulation program. Utilize the parallel computing capability of the chip to simulate the execution of the subtask through the molecular dynamics simulation program.

[0100] Step 104 implements the startup simulation and the running simulation of the subtask.

[0101] 1. Subtask startup simulation

[0102] For an actual chip (ie, a physically existing chip), the loaded molecular dynamics simulation program can be started, and the input parameters are the subtasks loaded in step 103 to ensure that the program starts running as expected.

[0103] For the simulated chip, the actual startup operation is not performed, but the task is directly marked as successfully started to simulate the program startup process. The startup simulation of the simulated chip can ignore the initialization time and resource allocation process.

[0104] 2. Subtask operation simulation

[0105] For actual chips (i.e., physically existing chips), the chip's parallel computing capabilities can be used to simulate the execution of subtasks through molecular dynamics simulation programs, such as calculating the interaction forces between particles, updating positions and velocities, etc.

[0106] For analog chips, actual computing operations are not performed, but tasks are directly marked as running to simulate the computing process. The computing simulation of analog chips can ignore the actual computing process and possible computing errors.

[0107] 105. Obtain the execution result of the subtask.

[0108] If the chip exists physically, the process of obtaining the execution result of the subtask is as follows: when the chip executes the subtask, the calculation data file continuously generated by the chip is transferred to the shared storage device. The file name of the calculation data file carries the chip identifier and the calculation round.

[0109] Step 105 implements the simulation of the subtask calculation task running results.

[0110] like Figure 5 As shown, the calculation subtask runs the simulation result. The chip will continuously generate calculation data files when performing molecular dynamics. The calculation data files are named with the chip ID and calculation round.

[0111] For actual chips (i.e. physically existing chips), the chips continuously generate computational data files while executing subtasks. These files contain information such as chip ID and computation rounds, and are transmitted to shared storage via PCIe or Ethernet.

[0112] The naming of data files can explicitly display information such as the calculation round and calculation chip ID to facilitate subsequent data analysis and processing.

[0113] For the simulation chip, the calculation data file generation process is marked as successful, and no actual transmission is performed. The data generation simulation of the simulation chip can ignore the actual data storage and transmission process.

[0114] After the central control system receives the first data file of a calculation round, it traverses the spatial grid to generate a soft link for each simulation chip. The name of the soft link is adjusted according to the simulation chip ID and the calculation round of this data task, and the soft link points to the first data file. When the calculation is completed, the calculation files of all actual chips are uploaded to the shared storage, and the calculation files of the simulation chip are softly connected to the calculation data. At this point, the molecular dynamics data calculation has been completed.

[0115] In molecular dynamics calculations, in order to reduce chip costs, the test environment or the R&D environment uses a partial chip to implement the functions of all chips instead of using the full chip. The chip simulation method for molecular dynamics calculations provided in this embodiment is aimed at this characteristic. When using a partial chip to simulate the full chip, it involves configuration simulation (i.e., step 101), calculation program delivery simulation (i.e., step 102), calculation task delivery simulation (i.e., step 103), calculation task startup simulation and calculation task operation simulation (i.e., step 104), and calculation task operation result simulation (i.e., step 105).

[0116] The chip simulation method for molecular dynamics calculation provided in this embodiment can reduce the demand for actual high-performance chips, thereby significantly reducing hardware costs. In addition, the chip simulation method for molecular dynamics calculation provided in this embodiment performs software function testing and optimization through chip simulation, reduces the time of actual software testing, shortens the software development cycle, and improves development efficiency.

[0117] The present embodiment provides a chip simulation method for molecular dynamics calculation, which determines simulation parameters, initializes the simulation environment according to the simulation parameters, loads the molecular dynamics simulation program onto the chip, decomposes the molecular dynamics calculation task into subtasks, matches the subtasks with the chip, and loads the subtasks onto the matching chip, executes the subtasks, and obtains the execution results of the subtasks. The method provided in the present embodiment can shorten the software development cycle, reduce testing time, and improve development efficiency by determining simulation parameters, initializing the simulation environment according to the simulation parameters, loading the molecular dynamics simulation program onto the chip, decomposing the molecular dynamics calculation task into subtasks, matching the subtasks with the chip, and loading the subtasks onto the matching chip, executing the subtasks, and obtaining the execution results of the subtasks.

[0118] Based on the same inventive concept of the chip simulation method of molecular dynamics calculation, this embodiment provides a chip simulation device of molecular dynamics calculation, such as Figure 6 As shown, the device comprises:

[0119] The simulation module 601 is used to determine simulation parameters and initialize the simulation environment according to the simulation parameters.

[0120] The first loading module 602 is used to load the molecular dynamics simulation program onto the chip.

[0121] The second loading module 603 is used to decompose the molecular dynamics calculation task into subtasks, match the subtasks with the chips, and load the subtasks onto the matching chips.

[0122] The execution module 604 is used to execute the subtask loaded by the second loading module 603 .

[0123] The acquisition module 605 is used to acquire the execution result of the subtask obtained by the execution module 604 .

[0124] Among them, the simulation parameters include: force field parameters, time step, number of participating chips, spatial grid size, and generation of random tasks.

[0125] The time step is between 1 and 2 femtoseconds.

[0126] The second loading module 603 is used to decompose the molecular dynamics calculation task into multiple subtasks according to the number of participating chips and chip characteristics, and to match the subtasks with the chips according to the number of participating chips and chip characteristics.

[0127] Among them, the chip exists physically.

[0128] The first loading module 602 is used to transfer the molecular dynamics simulation program to the chip memory through a high bandwidth interface.

[0129] The second loading module 603 is used to transfer the subtask to the matching chip memory through the high bandwidth interface.

[0130] Among them, the chip exists physically.

[0131] The execution module 604 is used to start the molecular dynamics simulation program according to the subtask and utilize the parallel computing capability of the chip to simulate and execute the subtask through the molecular dynamics simulation program.

[0132] Among them, the chip exists physically.

[0133] The acquisition module 605 is used to transfer the calculation data files continuously generated by the chip to the shared storage device during the chip executing the subtask, wherein the file name of the calculation data file carries the chip identification and the calculation round.

[0134] The device provided in this embodiment can determine simulation parameters, initialize the simulation environment according to the simulation parameters, load the molecular dynamics simulation program onto the chip, decompose the molecular dynamics calculation task into subtasks, match the subtasks with the chip, load the subtasks onto the matching chip, execute the subtasks, obtain the execution results of the subtasks, shorten the software development cycle, reduce testing time, and improve development efficiency.

[0135] Based on the same inventive concept of the chip simulation method calculated by molecular dynamics, this embodiment provides an electronic device, such as Figure 7 As shown, it includes: a memory 701, a processor 702, and a computer program.

[0136] The computer program is stored in the memory 701 and is configured to be executed by the processor 702 to implement the chip simulation method of the molecular dynamics calculation.

[0137] Specifically,

[0138] Determine simulation parameters and initialize the simulation environment according to the simulation parameters.

[0139] Load the molecular dynamics simulation program onto the chip.

[0140] Decompose the molecular dynamics calculation task into subtasks, match the subtasks with chips, and load the subtasks onto the matching chips.

[0141] Execute subtasks.

[0142] Get the execution result of the subtask.

[0143] Optionally, the simulation parameters include: force field parameters, time step, number of participating chips, spatial grid size, and generation of random tasks.

[0144] Optionally, the time step is between 1-2 femtoseconds.

[0145] Optionally, the molecular dynamics calculation task is decomposed into subtasks and the subtasks are matched to the chip, including:

[0146] The molecular dynamics calculation task is decomposed into multiple subtasks according to the number of participating chips and chip characteristics.

[0147] Subtasks are matched with chips according to the number of participating chips and chip characteristics.

[0148] Optionally, the chip is physically present.

[0149] Load the molecular dynamics simulation program onto the chip, including:

[0150] The molecular dynamics simulation program is transferred to the chip memory via a high-bandwidth interface.

[0151] Load subtasks onto matching chips, including:

[0152] The subtasks are transferred to matching on-chip memory via a high-bandwidth interface.

[0153] Optionally, the chip is physically present.

[0154] Perform subtasks, including:

[0155] According to the subtask, start the molecular dynamics simulation program.

[0156] The chip's parallel computing capabilities are used to simulate the execution of subtasks through a molecular dynamics simulation program.

[0157] Optionally, the chip is physically present.

[0158] Get the execution results of the subtask, including:

[0159] When the chip executes the subtask, the calculation data files continuously generated by the chip are transmitted to the shared storage device. The file name of the calculation data file carries the chip identification and calculation round.

[0160] The electronic device provided by this embodiment has a computer program executed by a processor to determine simulation parameters, initialize a simulation environment according to the simulation parameters, load a molecular dynamics simulation program onto a chip, decompose the molecular dynamics calculation task into subtasks, match the subtasks with the chip, load the subtasks onto the matching chip, execute the subtasks, obtain the execution results of the subtasks, shorten the software development cycle, reduce testing time, and improve development efficiency.

[0161] Based on the same inventive concept of the chip simulation method for molecular dynamics calculation, this embodiment provides a computer-readable storage medium, and a computer program is stored thereon. The computer program is executed by a processor to implement the chip simulation method for molecular dynamics calculation.

[0162] Specifically,

[0163] Determine simulation parameters and initialize the simulation environment according to the simulation parameters.

[0164] Load the molecular dynamics simulation program onto the chip.

[0165] Decompose the molecular dynamics calculation task into subtasks, match the subtasks with chips, and load the subtasks onto the matching chips.

[0166] Execute subtasks.

[0167] Get the execution result of the subtask.

[0168] Optionally, the simulation parameters include: force field parameters, time step, number of participating chips, spatial grid size, and generation of random tasks.

[0169] Optionally, the time step is between 1-2 femtoseconds.

[0170] Optionally, the molecular dynamics calculation task is decomposed into subtasks and the subtasks are matched to the chip, including:

[0171] The molecular dynamics calculation task is decomposed into multiple subtasks according to the number of participating chips and chip characteristics.

[0172] Subtasks are matched with chips according to the number of participating chips and chip characteristics.

[0173] Optionally, the chip is physically present.

[0174] Load the molecular dynamics simulation program onto the chip, including:

[0175] The molecular dynamics simulation program is transferred to the chip memory via a high-bandwidth interface.

[0176] Load subtasks onto matching chips, including:

[0177] The subtasks are transferred to matching on-chip memory via a high-bandwidth interface.

[0178] Optionally, the chip is physically present.

[0179] Perform subtasks, including:

[0180] According to the subtask, start the molecular dynamics simulation program.

[0181] The chip's parallel computing capabilities are used to simulate the execution of subtasks through a molecular dynamics simulation program.

[0182] Optionally, the chip is physically present.

[0183] Get the execution results of the subtask, including:

[0184] When the chip executes the subtask, the calculation data files continuously generated by the chip are transmitted to the shared storage device. The file name of the calculation data file carries the chip identification and calculation round.

[0185] The computer-readable storage medium provided in this embodiment has a computer program thereon which is executed by a processor to determine simulation parameters, initialize a simulation environment according to the simulation parameters, load a molecular dynamics simulation program onto a chip, decompose a molecular dynamics calculation task into subtasks, match the subtasks with the chip, load the subtasks onto a matching chip, execute the subtasks, obtain the execution results of the subtasks, shorten the software development cycle, reduce testing time, and improve development efficiency.

[0186] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of complete hardware embodiments, complete software embodiments, or embodiments in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code. The scheme in the embodiments of the present application can be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0187] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0188] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0189] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0190] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0191] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A chip simulation method for molecular dynamics calculation, characterized in that: The method comprises: Determining simulation parameters, and initializing a simulation environment according to the simulation parameters; Loading a molecular dynamics simulation program onto the chip; Decompose the molecular dynamics calculation task into subtasks, match the subtasks with chips, and load the subtasks onto the matching chips; executing the subtask; Get the execution result of the subtask.

2. The method according to claim 1, characterized in that The simulation parameters include: force field parameters, time step, number of participating chips, spatial grid size, and generation of random tasks.

3. The method according to claim 2, characterized in that The time step is between 1-2 femtoseconds.

4. The method according to claim 1, characterized in that: The molecular dynamics calculation task is decomposed into subtasks, and the subtasks are matched with the chip, including: Decompose the molecular dynamics calculation task into multiple subtasks according to the number of participating chips and chip characteristics; Subtasks are matched with chips according to the number of participating chips and chip characteristics.

5. The method according to claim 1, characterized in that The chip exists physically; The step of loading the molecular dynamics simulation program onto the chip comprises: Transferring molecular dynamics simulation programs to chip memory via a high-bandwidth interface; The step of loading the subtask onto the matching chip includes: The subtasks are transferred to matching on-chip memory via a high-bandwidth interface.

6. The method according to claim 1, characterized in that The chip exists physically; The performing of the subtask comprises: According to the subtask, start the molecular dynamics simulation program; The parallel computing capability of the chip is utilized to simulate and execute the subtasks through a molecular dynamics simulation program.

7. The method according to claim 1, characterized in that The chip exists physically; The obtaining the execution result of the subtask includes: During the process of the chip executing the subtask, the calculation data file continuously generated by the chip is transmitted to the shared storage device; wherein the file name of the calculation data file carries the chip identification and the calculation round.

8. A chip simulation device for molecular dynamics calculation, characterized in that: The device comprises: A simulation module, used to determine simulation parameters and initialize a simulation environment according to the simulation parameters; A first loading module, used for loading a molecular dynamics simulation program onto the chip; The second loading module is used to decompose the molecular dynamics calculation task into subtasks, match the subtasks with the chips, and load the subtasks onto the matching chips; An execution module, used for executing the subtask loaded by the second loading module; The acquisition module is used to acquire the execution result of the subtask obtained by the execution module.

9. An electronic device, characterized in that: include: Memory; processor; as well as Computer programs; The computer program is stored in the memory and is configured to be executed by the processor to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: A computer program is stored thereon; the computer program is executed by a processor to implement the method according to any one of claims 1 to 7.

Citation Information

Cited By

  • Chip architecture simulation system for molecular dynamics

    CN120032731A