CN117709128A

CN117709128A - Super-computing-oriented multi-dimensional parallel simulation method, device, equipment and storage medium

Info

Publication number: CN117709128A
Application number: CN202410160263.2A
Authority: CN
Inventors: 杨灿群; 仲彦旭; 王伟; 黄颖杰; 夏梓峻; 段莉莉; 郑伟龙; 卢海林
Original assignee: Haihe Laboratory Of Advanced Computing And Key Software Xinchuang; National Supercomputer Center In Tianjin; National University of Defense Technology
Current assignee: Haihe Laboratory Of Advanced Computing And Key Software Xinchuang; National Supercomputer Center In Tianjin; National University of Defense Technology
Priority date: 2024-02-05
Filing date: 2024-02-05
Publication date: 2024-03-15
Anticipated expiration: 2044-02-05

Abstract

The embodiment of the disclosure relates to a super-computing-oriented multi-dimensional parallel simulation method, device, equipment and storage medium, wherein the method comprises the following steps: performing simulation initialization aiming at a target engineering problem, wherein the simulation initialization comprises the steps of establishing a low-dimensional simulation model and a high-dimensional simulation model; encrypting the grid on the low-dimensional simulation model; calculating current low-dimensional data corresponding to the encrypted low-dimensional simulation model by adopting a first number of kernels until the high-dimensional reconstruction condition is monitored to be met; mapping the current low-dimensional data to a high-dimensional simulation model after high-dimensional reconstruction; calculating current high-dimensional data corresponding to the high-dimensional simulation model by adopting a second number of kernels until the simulation duration reaches a preset duration; wherein a ratio between the calculated amount of the grid of each core in the first number of cores and the calculated amount of the grid of each core in the second number of cores is within a preset range. Therefore, the problem of load mismatch between different dimensions and the problem of resource waste in the low-dimensional simulation process can be improved.

Description

Super-computing-oriented multi-dimensional parallel simulation method, device, equipment and storage medium

Technical Field

The embodiment of the disclosure relates to the technical field of computers, in particular to a multi-dimensional simulation method, a device, equipment and a storage medium.

Background

In the equipment research and development process of each field, the problems of long period, large risk coefficient, high cost and the like exist in research and development through a test method, and because the CAE simulation technology can effectively replace a test, the research and development period is greatly shortened, the research and development cost is reduced, the actual engineering problem is generally simulated based on the CAE simulation technology in the equipment research and development process.

At present, in order to reduce the simulation calculation amount, a multi-dimensional simulation method is generally adopted to simulate the actual engineering problem, but the multi-dimensional simulation method generally has the problem of load mismatch between different dimensions, so that resource waste can be generated in the low-dimensional simulation process, and the application of the multi-dimensional simulation method is restricted. For example, for very large scale explosion problems, the calculation scale can reach several kilometers, even tens kilometers, due to the wide range, large radiation area, so the full-scale model grid scale in the simulation process can reach hundreds of billions, or more. In order to reduce the simulation calculation amount and quickly respond in a short time, a traditional multi-dimensional (i.e. Remap) simulation method can be generally adopted for simulation, and the ultra-large scale explosion problem is reduced in dimension and simplified, but the multi-dimensional simulation method has the problem of load mismatch among different dimensions.

There is currently no effective solution to the above problems.

Disclosure of Invention

In order to solve the above technical problems or at least partially solve the above technical problems, embodiments of the present disclosure provide a multi-dimensional parallel simulation method, device, apparatus and storage medium for super-computing.

A first aspect of an embodiment of the present disclosure provides a super-computing-oriented multidimensional parallel simulation method, including:

performing simulation initialization for the target engineering problem, wherein the simulation initialization comprises the steps of establishing a low-dimensional simulation model and a high-dimensional simulation model corresponding to the target engineering problem;

encrypting the grid on the low-dimensional simulation model;

calculating current low-dimensional data corresponding to the encrypted low-dimensional simulation model by adopting a first number of kernels until the high-dimensional reconstruction condition is monitored to be met;

mapping the current low-dimensional data to a high-dimensional simulation model after high-dimensional reconstruction;

calculating current high-dimensional data corresponding to the high-dimensional simulation model by adopting a second number of kernels until the simulation duration reaches a preset duration;

wherein a ratio between the calculated amount of the grid of each core in the first number of cores and the calculated amount of the grid of each core in the second number of cores is within a preset range.

A second aspect of an embodiment of the present disclosure provides a super-computing-oriented multidimensional parallel simulation apparatus, the apparatus including:

the initialization module is used for carrying out simulation initialization aiming at the target engineering problem, wherein the simulation initialization comprises the steps of establishing a high-dimensional simulation model and a low-dimensional simulation model corresponding to the target engineering problem;

the encryption module is used for encrypting the grids on the low-dimensional simulation model;

the first calculation module is used for calculating the current low-dimensional data corresponding to the encrypted low-dimensional simulation model by adopting a first number of kernels until the high-dimensional reconstruction condition is monitored to be met;

the reconstruction mapping module is used for mapping the current low-dimensional data to a high-dimensional simulation model after high-dimensional reconstruction;

the second calculation module is used for calculating the current high-dimensional data corresponding to the high-dimensional simulation model by adopting a second number of kernels until the simulation duration reaches the preset duration;

the ratio of the grid quantity calculated by each core in the first number of cores to the grid quantity calculated by each core in the second number of cores is a preset scaling factor, and the preset scaling factor is in a preset range.

A third aspect of the disclosed embodiments provides an electronic device, the server comprising: a processor and a memory, wherein the memory has stored therein a computer program which, when executed by the processor, performs the method of the first aspect described above.

A fourth aspect of the disclosed embodiments provides a computer readable storage medium having stored therein a computer program which, when executed by a processor, can implement the method of the first aspect described above.

Compared with the prior art, the technical scheme provided by the embodiment of the disclosure has the following advantages:

according to the embodiment of the disclosure, simulation initialization can be performed aiming at the target engineering problem, wherein the simulation initialization comprises the steps of establishing a low-dimensional simulation model and a high-dimensional simulation model corresponding to the target engineering problem; encrypting the grid on the low-dimensional simulation model; calculating current low-dimensional data corresponding to the encrypted low-dimensional simulation model by adopting a first number of kernels until the high-dimensional reconstruction condition is monitored to be met; mapping the current low-dimensional data to a high-dimensional simulation model after high-dimensional reconstruction; calculating current high-dimensional data corresponding to the high-dimensional simulation model by adopting a second number of kernels until the simulation duration reaches a preset duration; wherein a ratio between the calculated amount of the grid of each core in the first number of cores and the calculated amount of the grid of each core in the second number of cores is within a preset range. Therefore, by adopting the technical scheme, the grid quantity of the encrypted low-dimensional simulation model can be increased by encrypting the grids on the low-dimensional simulation model, so that the grid calculation quantity of each core in the first number of cores is increased, the difference between the grid calculation quantity of each core in the first number of cores and the grid calculation quantity of each core in the second number of cores is reduced, the ratio between the two is within a preset range, and therefore, the problem of load mismatch between different dimensions can be solved, and the problem of resource waste in the low-dimensional simulation process is solved.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the disclosure and together with the description, serve to explain the principles of the disclosure.

In order to more clearly illustrate the embodiments of the present disclosure or the solutions in the prior art, the drawings that are required for the description of the embodiments or the prior art will be briefly described below, and it will be obvious to those skilled in the art that other drawings can be obtained from these drawings without inventive effort.

FIG. 1 is a flow chart of a super-computing oriented multi-dimensional parallel simulation method provided by an embodiment of the present disclosure;

FIG. 2 is a cut-away schematic view of a high-dimensional reconstruction map provided by an embodiment of the present disclosure;

fig. 3 is a schematic diagram of a setting situation of intermittent monitoring points according to an embodiment of the present disclosure;

FIG. 4 is a schematic diagram of a case where high-dimensional reconstruction conditions are satisfied provided by an embodiment of the present disclosure;

FIG. 5 is a schematic flow diagram of a multi-dimensional parallel simulation process for supercomputing provided by an embodiment of the present disclosure;

FIG. 6 is a schematic diagram of a high-dimensional simulation model corresponding to an explosive explosion problem provided by an embodiment of the present disclosure;

FIG. 7 is a graph of the results of a typical time-of-day encrypted low-dimensional simulation model calculation provided by an embodiment of the present disclosure;

FIG. 8 is a high-dimensional simulation model containing low-dimensional information provided by an embodiment of the present disclosure;

FIG. 9 is a graph of exemplary time-of-day high-dimensional simulation model calculations provided by an embodiment of the present disclosure;

FIG. 10 is a schematic structural diagram of a multi-dimensional parallel simulation device for supercomputing provided in an embodiment of the present disclosure;

fig. 11 is a schematic structural view of an electronic device in an embodiment of the present disclosure.

Detailed Description

In order that the above objects, features and advantages of the present disclosure may be more clearly understood, a further description of aspects of the present disclosure will be provided below. It should be noted that, without conflict, the embodiments of the present disclosure and features in the embodiments may be combined with each other.

In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present disclosure, but the present disclosure may be practiced otherwise than as described herein; it will be apparent that the embodiments in the specification are only some, but not all, embodiments of the disclosure.

Fig. 1 is a flowchart of a super-computing oriented multi-dimensional parallel simulation method provided by an embodiment of the present disclosure, which may be performed by an electronic device. The electronic device may be exemplarily understood as a device having a page presentation function, such as a mobile phone, a tablet computer, a notebook computer, a desktop computer, a smart television, etc. As shown in fig. 1, the method provided in this embodiment includes the following steps:

S110, carrying out simulation initialization on the target engineering problem, wherein the simulation initialization comprises the steps of establishing a low-dimensional simulation model and a high-dimensional simulation model corresponding to the target engineering problem.

In the embodiment of the disclosure, the target engineering problem needs to be subjected to simulation initialization to obtain a corresponding low-dimensional simulation model and a high-dimensional simulation model, so that the physical state of the low-dimensional simulation model is calculated in a subsequent simulation manner in the low-dimensional simulation process, and the physical state of the high-dimensional simulation model is calculated in a simulation manner in the high-dimensional simulation process, and multi-dimensional simulation calculation is realized.

Specifically, the target engineering problem may include engineering problems in the ultra-large scale explosion field, engineering problems in the energy-gathering jet damage and protection field, engineering problems in the hypersonic advanced weapon research and development design field, and the like, wherein the engineering problems in the ultra-large scale explosion field may include urban safety protection, risk assessment and emergency pre-warning, hazardous chemical production/storage/disposal, underwater/aerial explosion damage and protection, bubble pulsation, and the like, but are not limited thereto.

In some embodiments, the simulation initialization may include: establishing an Euler coordinate system, establishing a calculation region and grid division (corresponding to S111), establishing a model (corresponding to S112), defining a target discontinuity (corresponding to S113), setting initial conditions (corresponding to S114), selecting a control equation (corresponding to S115), and setting solving time (corresponding to S116). Accordingly, S110 may include S111-S116 as follows:

S111, establishing an Euler coordinate system according to the target engineering problem, defining the size of a calculation region, and carrying out grid division on the calculation region according to the preset Euler grid cell size.

The specific value of the preset euler mesh cell size may be set according to practical situations, and is not limited herein.

S112, establishing a high-dimensional simulation model and a low-dimensional simulation model corresponding to the target engineering problem.

Specifically, the high-dimensional simulation model includes a high-dimensional material model corresponding to a plurality of materials in the target engineering problem, and the high-dimensional material model may include a three-dimensional model. The low-dimensional simulation model comprises a low-dimensional material model corresponding to a plurality of materials in the target engineering problem, and the low-dimensional material model can comprise a one-dimensional model and/or a two-dimensional model. It should be noted that, the low-dimensional simulation model includes a low-dimensional material model corresponding to which materials in the target engineering problem, which can be determined according to an actual physical scene, and is not limited herein.

For example, S112 may include: establishing a corresponding high-dimensional material model aiming at a material in a target engineering problem so as to obtain a high-dimensional simulation model, wherein the high-dimensional material model corresponding to the material with a fluid phase is a Gao Weiou Law model, and the high-dimensional material model corresponding to the material with a solid phase is a Gao Weiou Law model or a high-dimensional Lagrange model; and performing dimension reduction modeling on the high-dimensional material model with the spherical symmetry and/or the axial symmetry to obtain a corresponding low-dimensional material model, thereby obtaining a low-dimensional simulation model. Therefore, the method for establishing the high-dimensional simulation model and the low-dimensional simulation model is simple and easy to realize, the one-dimensional low-dimensional material model is established according to the spherical symmetry characteristic, the two-dimensional low-dimensional material model is established according to the axisymmetric characteristic, the dimension of the low-dimensional material model is as low as possible, and the calculation amount is reduced.

S113, determining the initial position of the target discontinuity.

Specifically, when the target discontinuity is a contact discontinuity (i.e., a multi-media interface), the initial position of the multi-media interface corresponding to the fluid is defined by using a multi-media interface tracking method. For example, an interface tracking method based on a level set method may be used to determine the initial position of the multi-media interface based on a symbolic distance function.

In particular, when the target discontinuity is a shockwave discontinuity, a shockwave discontinuity indicator may be employed to determine an initial location of the shockwave discontinuity.

S114, setting the initial physical state of the high-dimensional simulation model material and the initial physical state of the material in the low-dimensional simulation model.

Specifically, the physical state may include at least one of a material property, a speed, a pressure, a material model, a boundary condition, a detonation point position, a detonation type, and the like, but is not limited thereto.

S115, selecting a proper control equation for the target engineering problem.

Specifically, the control equation may include an N-S equation, a non-stick euler equation, an elastoplastic euler equation considering the strength of a solid, etc., but is not limited thereto.

S116, setting a preset duration.

S120, encrypting the grid on the low-dimensional simulation model.

In the disclosed embodiments, the reason for load mismatch between multiple dimensions was found to be: the dimension of the low-dimensional simulation model is lower than that of the high-dimensional simulation model, so that the grid quantity of the low-dimensional simulation model is far smaller than that of the high-dimensional simulation model, but the quantity of cores used for simulation calculation in the low-dimensional simulation calculation process and the high-dimensional simulation calculation process is the same, and therefore the ratio between the calculated quantity of each core grid in the low-dimensional simulation calculation process and the calculated quantity of each core grid in the high-dimensional simulation calculation process is too small, namely the load is not matched. Therefore, the grid on the low-dimensional simulation model can be encrypted, so that the grid quantity of the low-dimensional simulation model is increased, the difference between the grid quantity of the low-dimensional simulation model and the grid quantity of the high-dimensional simulation model is reduced, the ratio (called as load ratio) between the calculated quantity of each core grid in the low-dimensional simulation calculation process and the calculated quantity of each core grid in the high-dimensional simulation calculation process is increased, and the load ratio is within a preset range, namely the problem of load mismatch is solved.

In some embodiments, S120 may include: and encrypting the grid on the low-dimensional simulation model according to a preset fixed encryption multiple.

Specifically, the specific value of the preset fixed encryption multiple may be set by those skilled in the art according to the actual situation, and is not limited herein.

In other embodiments, S120 may include: s121, determining the encryption multiple according to the grid quantity of the low-dimensional simulation model, the grid quantity of the high-dimensional simulation model and a preset scaling factor.

Specifically, the preset scaling factor is an expected value for the load ratio, and is within a preset range. It should be noted that, the specific range of the preset range may be set by a person skilled in the art according to the actual load matching requirement, and is not limited herein. For example, the preset range is [0.05,0.15], and the preset scaling factor is 0.1.

Specifically, the encryption multiple can be calculated by the following formula (1):

formula (1)

Wherein,for encryption multiple->Grid quantity for high-dimensional simulation model, +.>Grid quantity for low-dimensional simulation model, +.>Is a preset scaling factor.

S122, under the condition that the encryption multiple is smaller than or equal to the limit encryption multiple and the minimum grid size corresponding to the encryption multiple meets the continuous medium assumption, the grids on the low-dimensional simulation model are encrypted according to the encryption multiple, wherein the minimum grid size corresponding to the encryption multiple is the size of the grid with the minimum size after the low-dimensional simulation model is encrypted according to the encryption multiple.

It can be understood that the grid quantity of the low-dimensional simulation model can be increased by encrypting the low-dimensional simulation model, so that the grid calculation quantity of each kernel in the low-dimensional simulation process is increased, the problem of high-low dimensional load mismatch is further improved or even avoided, the grid quantity of the low-dimensional simulation model is increased, and the calculation precision in the low-dimensional simulation process can be improved.

Optionally, S120 may further include: s123, correcting the encryption multiple to be the limit encryption multiple under the condition that the encryption multiple is larger than the limit encryption multiple; and encrypting the grid on the low-dimensional simulation model according to the corrected encryption multiple.

Specifically, the specific value of the limiting encryption multiple can be set by those skilled in the art according to factors such as actual solving time steps, and the like, and is not limited herein. For example, for a one-dimensional low-dimensional simulationThe true model is used to determine the true model,for a two-dimensional low-dimensional simulation model +.>。

Specifically, at the encryption multipleEncryption multiple greater than limit->In the case of (2), the ++can be expressed by the following formula (2)>And (3) correcting:

formula (2)

Wherein,for encryption multiple according to limit>Modified encryption multiplier->Is a limit encryption multiple.

Illustratively, in the case where the encryption multiple is greater than the limit encryption multiple and the minimum mesh size corresponding to the encryption multiple satisfies the continuous medium assumption: correcting the encryption multiple to be a limit encryption multiple; and encrypting the grid on the low-dimensional simulation model according to the corrected encryption multiple.

It can be understood that the simulation calculation has a certain requirement on the grid size of the grid on the model, and the undersize of the grid not only can cause the undersize of the solving time step, so that the encryption multiple of the low-dimensional simulation model can be limited by the limit encryption multiple, and the problem that the solving time step is greatly shortened due to the undersize of the grid of the low-dimensional simulation model after encryption is avoided.

Optionally, S120 may further include: s124, under the condition that the minimum grid size corresponding to the encryption multiple does not meet the assumption of the continuous medium, determining a correction coefficient according to the minimum grid size corresponding to the encryption multiple and the grid size under the assumption limit of the continuous medium; correcting the encryption multiple according to the correction coefficient; and encrypting the grid on the low-dimensional simulation model according to the corrected encryption multiple.

Specifically, in the case where the minimum mesh size corresponding to the encryption multiple does not satisfy the continuous medium assumption, that is, the mesh size at the continuous medium assumption limitLess than encryption multiple->Corresponding minimum mesh size->In the case of (2), the encryption multiple can be ++>And (3) correcting:

formula (3)

Wherein,for encryption multiple->Assuming a grid size under limit for a continuous medium, +. >Is encryption multiple->Corresponding minimum mesh size, < >>For correction factor +.>For rounding up the function, i.e. taking the smallest integer value not smaller than this value, ++>The encryption multiple is corrected according to the correction coefficient.

Illustratively, in the case where the encryption multiple is equal to or less than the limit encryption multiple and the minimum mesh size corresponding to the encryption multiple does not satisfy the continuous medium assumption: determining a correction coefficient according to the minimum grid size corresponding to the encryption multiple and the grid size under the assumption limit of the continuous medium; correcting the encryption multiple according to the correction coefficient; and encrypting the grid on the low-dimensional simulation model according to the corrected encryption multiple.

It can be appreciated that the calculation accuracy is reduced after the minimum mesh size exceeds the continuous medium assumption limit, so that the encryption multiple of the low-dimensional simulation model can be limited according to the continuous medium assumption limit, so that the problem that the mesh size of the low-dimensional simulation model exceeds the continuous medium assumption limit after encryption is avoided.

Optionally, S120 may further include: under the condition that the encryption multiple is larger than the limit encryption multiple and the minimum grid size corresponding to the encryption multiple does not meet the continuous medium assumption, correcting the encryption multiple to be the limit encryption multiple; under the condition that the minimum grid size corresponding to the corrected encryption multiple (namely the limit encryption multiple) still does not meet the continuous medium assumption, determining a correction coefficient according to the minimum grid size corresponding to the corrected encryption multiple and the grid size under the limit of the continuous medium assumption; correcting the corrected encryption multiple again according to the correction coefficient; encrypting the grids on the low-dimensional simulation model according to the encryption multiple after the re-correction; or alternatively;

Determining a correction coefficient according to the minimum grid size corresponding to the encryption multiple and the grid size under the assumption limit of the continuous medium; correcting the encryption multiple according to the correction coefficient; and taking the minimum value in the corrected encryption multiple and the limit encryption multiple as a target encryption multiple, and encrypting the grid on the low-dimensional simulation model according to the target encryption multiple.

S130, calculating current low-dimensional data corresponding to the encrypted low-dimensional simulation model by adopting a first number of kernels until the high-dimensional reconstruction condition is monitored.

In the embodiment of the disclosure, in the process of simulating the target engineering problem, different parts of the model are distinguished according to whether the dimension reduction simplification can be performed, wherein the parts which cannot be performed with the dimension reduction, namely the target substance, can interact with the target intermittently. In order to reduce the amount of calculation, when the high-dimensional reconstruction condition is not satisfied, indicating that the target substance and the target discontinuity distance are far, the calculation may be performed in a low dimension, that is, a low-dimensional simulation process may be performed, specifically, the following steps may be periodically performed until it is monitored that the high-dimensional reconstruction condition is satisfied: and calculating the current low-dimensional data corresponding to the encrypted low-dimensional simulation model by adopting a first number of kernels. When the high-dimensional reconstruction condition is met, the target substance and the target interruption are indicated to be relatively close, interaction between the target interruption and the target substance is about to occur, and at the moment, the encrypted low-dimensional simulation model can be subjected to high-dimensional reconstruction and mapping so as to perform calculation in high dimensions, namely, a high-dimensional simulation process.

In some embodiments, the high-dimensional reconstruction condition is determined to be satisfied in response to receiving Gao Weichong construct operations.

Specifically, the high-dimensional reconstruction operation may include a triggering operation of the control by a mouse, a keyboard, and a touch pad pair Gao Weichong, but is not limited thereto.

In other embodiments, intermittent monitoring points are arranged in the encrypted low-dimensional simulation model, and when the distance between the target intermittent monitoring points and the intermittent monitoring points is smaller than the preset Euler grid cell size, the high-dimensional reconstruction condition is determined to be met.

Specifically, the intermittent monitoring points may be disposed around the outer periphery of the target substance, and a predetermined number of euler mesh cells are spaced between the intermittent monitoring points and the target substance. It should be noted that, the specific value of the preset number may be set by those skilled in the art according to the actual situation, and is not limited herein. For example, the preset number is 1.

Fig. 3 is a schematic diagram illustrating a setting situation of intermittent monitoring points according to an embodiment of the present disclosure. Fig. 4 is a schematic diagram of a case where a high-dimensional reconstruction condition is satisfied, provided by an embodiment of the present disclosure. Referring to fig. 3 and 4, intermittent monitoring points are disposed at a distance L (i.e., a preset number of euler mesh units) from the high-dimensional simulation model around the high-dimensional simulation model, so that when the object intermittently moves to a distance from the intermittent monitoring points smaller than the preset euler mesh unit size, the high-dimensional reconstruction condition is satisfied.

It can be appreciated that whether the high-dimensional reconstruction condition is met currently can be automatically monitored by setting intermittent monitoring points so as to accurately grasp the high-dimensional reconstruction opportunity.

In some embodiments, the first number is a preset first fixed number.

Specifically, the specific value of the preset first fixed number may be set by those skilled in the art according to the actual situation, and is not limited herein.

In other embodiments, the first number is determined based on the mesh size of the encrypted low-dimensional simulation model and a mesh count for each kernel preset during the low-dimensional simulation.

Specifically, the grid computing amount of each kernel preset in the low-dimensional simulation process can be set by a person skilled in the art according to actual situations, and the method is not limited herein. For example, the grid computation amount per kernel is 1 ten thousand grids.

Specifically, the first number of computing resources, i.e., kernels, required during the low-dimensional simulation may be calculated by equation (4):

formula (4)

Wherein,for the first quantity->Grid computation for each kernel preset in the low-dimensional simulation process, +.>And the grid quantity of the encrypted low-dimensional simulation model.

In some embodiments, the first number is an integer greater than or equal to 2, where S130 may include:

S131, performing parallel region division on the encrypted low-dimensional simulation model by adopting an automatic region decomposition method to obtain a first number of low-dimensional simulation regions.

The grid quantity of each low-dimensional simulation area is the grid calculated quantity of each kernel preset in the low-dimensional simulation process, or the difference value between the grid quantity of each low-dimensional simulation area and the grid calculated quantity of each kernel preset in the low-dimensional simulation process is within a tolerance range.

And S132, performing parallel simulation calculation on the low-dimensional simulation areas corresponding to the cores by adopting the first number of cores to obtain the current low-dimensional data corresponding to the encrypted low-dimensional simulation model.

Specifically, the low-dimensional data is a physical quantity of each low-dimensional grid node on the encrypted low-dimensional simulation model, and is used for representing the physical state of a substance in the encrypted low-dimensional simulation model.

Correspondingly, the current low-dimensional data is the physical quantity of each low-dimensional grid node on the encrypted low-dimensional simulation model at the next time step, and is obtained by adopting a control equation to simulate and calculate.

Specifically, the current low-dimensional data corresponding to the encrypted low-dimensional simulation model can be obtained by adopting a peer-to-peer mode and a non-blocking communication method based on an MPI method. But is not limited thereto.

S133, when the fact that the high-dimensional reconstruction condition is not met is detected, returning to the S132 to continue the low-dimensional simulation process; and when the high-dimensional reconstruction condition is detected to be met, ending the low-dimensional simulation process.

It can be appreciated that the parallel computation is adopted in the low-dimensional simulation process, so that the rapid low-dimensional computation can be realized, and the requirement of rapid response of the target engineering problem can be met.

And S140, mapping the current low-dimensional data to a high-dimensional simulation model after high-dimensional reconstruction.

Specifically, the current low-dimensional data is subjected to high-dimensional reconstruction in different directions to obtain reconstruction data corresponding to the high-dimensional simulation model. Because the reconstruction data is data which does not contain a grid structure, the reconstruction data obtained by reconstruction can be mapped to a high-dimensional Euler grid in the high-dimensional simulation model so that the high-dimensional simulation model contains a low-dimensional calculation result.

Optionally, S140 may include:

s141, performing high-dimensional reconstruction on the current low-dimensional data according to a preset direction to obtain reconstruction data.

Specifically, for the encrypted low-dimensional simulation model, if the high-dimensional material model corresponding to the low-dimensional material model is in spherical symmetry, reconstructing the low-dimensional material model into a spherical-symmetrical three-dimensional model, so as to realize Gao Weichong structure of the current low-dimensional material model corresponding to the low-dimensional material model; and if the high-dimensional material model corresponding to the low-dimensional material model is axisymmetric, reconstructing the low-dimensional material model into an axisymmetric three-dimensional model, so that Gao Weichong structure of the current low-dimensional material model corresponding to the low-dimensional material model can be realized.

For example, fig. 2 is a schematic diagram of a tangent plane of a high-dimensional reconstruction map provided by the embodiment of the present disclosure, as shown in fig. 2, if a simulation problem is a two-dimensional axisymmetric problem, the two-dimensional problem may be simplified into a one-dimensional problem, that is, the two-dimensional problem is shown as a black node in the map, after one-dimensional calculation is completed, the two-dimensional problem is subjected to data reconstruction, that is, axisymmetric rotation, so that physical states of different nodes may form a state diagram of a concentric circle shape in the map, while a two-dimensional calculation grid is a cartesian grid in the map, it can be seen that the state diagram and the grid do not correspond one to one another, and therefore high-order mapping is required, and high-dimensional data, that is, information contained in the state diagram, is interpolated into the grid.

S142, mapping reconstruction data corresponding to a first region to a high-dimensional simulation model by adopting a Lagrange interpolation method, wherein the first region is a smooth region in a calculation domain where the encrypted low-dimensional simulation model is located.

S143, mapping reconstruction data corresponding to a second region to a high-dimensional simulation model by adopting a weighted intrinsic oscillation-free interpolation method, wherein the second region is a region with a break in a calculation domain where the encrypted low-dimensional simulation model is located.

Specifically, since the encrypted low-dimensional simulation model is obtained through grid encryption, the high-dimensional grid nodes in the high-dimensional simulation model can correspond to the reconstructed data obtained by reconstructing a plurality of low-dimensional grid nodes, and therefore, a high-order weighted interpolation method can be adopted when the reconstructed data is mapped to the high-dimensional simulation model, so that the mapping accuracy is improved.

Specifically, the interpolation direction is selected as the reconstructed normal direction, i.e. the weighted interpolation template points are selected according to the reconstructed normal direction.

Specifically, since the lagrangian interpolation (lagrangian interpolation) method is small in calculation amount, a 5-order lagrangian interpolation method can be adopted for a smooth region. Since Lagrange interpolation may exhibit non-physical oscillations in the region where there is a discontinuity, weighted intrinsic non-oscillating interpolation (WENO interpolation) may be employed for the region where there is a discontinuity, and in particular, 5 th order WENO interpolation may be employed.

The calculation formula of the 5-order Lagrange interpolation method is as follows:

formula (5)

Wherein,physical quantity of high-dimensional grid node, +.>Is->The reconstruction data of the weighted interpolation template points, i.e. the low dimensional data of the corresponding position, +.>Is->Is used for the interpolation weighting coefficients of the (a).

The calculation formula of the 5-order WENO interpolation method is as follows:

formula (6)

Formula (7)

Wherein,physical quantity of high-dimensional grid node, +.>、/>、/>、/>、/>Is->The reconstruction data of the weighted interpolation template points, i.e. the low dimensional data of the corresponding position, +.>、/>、/>For interpolation coefficients +.>、/>、/>Is a coefficient of->Is->Interpolation weighting coefficients of>Is->Is used for the interpolation weighting coefficients of the (a).

It can be understood that the WENO interpolation and Lagrange interpolation methods are coupled, so that a high-efficiency and high-precision weighted mapping method can be realized, the accuracy of the multi-dimensional mapping process is ensured, and compared with the traditional method, the method has the advantages of low dissipation, no oscillation, high efficiency and the like, and the accurate prediction of the target engineering problem is realized.

And S150, calculating current high-dimensional data corresponding to the high-dimensional simulation model by adopting a second number of kernels until the simulation duration reaches the preset duration.

In the embodiment of the present disclosure, in the high-dimensional simulation process, specifically, the following steps may be periodically performed until the simulation duration reaches the preset duration: and calculating the current high-dimensional data corresponding to the encrypted high-dimensional simulation model by adopting a second number of kernels.

In some embodiments, the second number is a preset second fixed number.

Specifically, the specific value of the preset second fixed number may be set by those skilled in the art according to the actual situation, and is not limited herein. For example, the second fixed number is equal to the first fixed number.

In other embodiments, the second number is equal to the first number in the case where the encryption multiplier is equal to or less than the limit encryption multiplier and a minimum mesh size corresponding to the encryption multiplier meets the continuous medium assumption; under the condition that the encryption multiple is larger than the limit encryption multiple and/or the minimum grid size corresponding to the encryption multiple does not meet the continuous medium assumption, determining a second quantity according to the grid quantity of the high-dimensional simulation model, the grid calculated quantity of each kernel preset in the low-dimensional simulation process and the preset scaling multiple, wherein the second quantity can be calculated by the formula (8):

Formula (8)

Wherein,for the second quantity->Grid quantity for high-dimensional simulation model, +.>For the grid calculation amount of each kernel preset in the high-dimensional simulation process,/for each kernel>Grid computation for each kernel preset in the low-dimensional simulation process, +.>Is a preset scaling multiple.

In some embodiments, the second number is an integer greater than or equal to 2, where S150 may include:

s151, performing parallel region division on the high-dimensional simulation model by adopting an automatic region decomposition method to obtain a second number of high-dimensional simulation regions.

The grid quantity of each high-dimensional simulation area is the grid calculated quantity of each inner core preset in the high-dimensional simulation process, or the difference value between the grid quantity of each high-dimensional simulation area and the grid calculated quantity of each inner core preset in the high-dimensional simulation process is within a tolerance range.

And S152, performing parallel simulation calculation on the high-dimensional simulation areas corresponding to the cores by adopting a second number of cores to obtain current high-dimensional data corresponding to the high-dimensional simulation model.

Specifically, the high-dimensional data is a physical quantity of each high-dimensional grid node on the high-dimensional simulation model, and is used for representing the physical state of a substance in the high-dimensional simulation model.

Correspondingly, the current high-dimensional data is the physical quantity of each high-dimensional grid node on the high-dimensional simulation model at the next time step, and is obtained by adopting a control equation to simulate and calculate.

Specifically, the current high-dimensional data corresponding to the high-dimensional simulation model can be obtained by adopting a peer-to-peer mode and a non-blocking communication method based on an MPI method. But is not limited thereto.

S153, when the simulation duration does not reach the preset duration, returning to the step S152 to continue the high-dimensional simulation process; and when the simulation duration reaches the preset duration, ending the high-dimensional simulation process.

Specifically, the term that the simulation duration reaches the preset duration refers to that the target is interrupted to start simulation movement from the position where the target is initialized, and after the simulation movement reaches the preset duration, the whole multi-dimensional simulation process is ended.

It can be understood that parallel computation is adopted in the high-dimensional simulation process, so that rapid high-dimensional computation can be realized, and the requirement of rapid response of the target engineering problem is met.

Optionally, when the first number is not equal to the second number, calculating current high-dimensional data corresponding to the high-dimensional simulation model by adopting the second number of kernels until the simulation duration reaches the preset duration, further including: s160, generating a restarting task script, and suspending the current computing tasks of the first number of kernels; executing the restart task script to allocate a second number of kernels.

Specifically, the restart task script is regenerated according to the calculation information obtained after S140, and the calculation program (i.e., the calculation program corresponding to the current calculation task of the first number of kernels) is terminated at the same time. Based on the automatic scheduling system, executing the restarting task script to realize automatic restarting of the computing program, simultaneously distributing a second number of kernels, and then performing S150.

Thus, continuous large-scale calculation of automatic load balancing before and after dimension crossing can be realized.

According to the embodiment of the disclosure, based on the multi-dimensional parallel simulation method oriented to super-calculation, the multi-dimensional load matching and the low-dissipation high-order mapping are coupled, and meanwhile, the multi-dimensional large-scale parallel numerical simulation of the large-scale parallel method is combined, so that the solving speed and the calculating efficiency of the super-scale target engineering problem can be greatly improved, the simulation precision is improved, and the error between the simulation precision and the test is reduced. The multi-dimensional load matching can realize load matching of parallel computation in different dimensions, so that the computing efficiency is ensured, and the engineering application level is improved; the low-dissipation high-order mapping can improve the mapping precision between different dimensions through high-order weighted interpolation, avoid the dissipation of the impact wave number value and ensure the calculation precision of different dimensions; based on MPI parallelism, large-scale simulation of different dimensions is realized, the calculation scale and calculation speed are improved, and the simulation requirement of the ultra-large scale target engineering problem is met. Therefore, according to the embodiment of the disclosure, high-precision and high-efficiency simulation prediction can be effectively realized on the ultra-large scale target engineering problem.

In order to better illustrate the objects and advantages of the embodiments of the present disclosure, a detailed description will be made below of a physical scene simulation method provided by the embodiments of the present disclosure based on a specific example.

Fig. 5 is a schematic flow chart of a multi-dimensional parallel simulation process for super computing according to an embodiment of the disclosure. Referring to fig. 5, the super-computing-oriented multi-dimensional parallel simulation process includes the following steps:

s410, CEA simulation initialization.

Specifically, the initialization may be performed with reference to the content described in the foregoing S111-S116, and the content described in the intermittent monitoring point in S130, which will not be described herein.

For example, simulation initialization is performed for the explosive explosion problem, wherein the full-size calculation domain is 100m×50m×18m, the grid quantity is 24.3 million, the low-dimensional simulation model calculation domain is 75m×18m, the grid quantity is 121.5 million, and since the front section of the calculation domain has little influence on the structure when the shock wave formed by the explosive explosion propagates to the position of the earth dike structure, the high-dimensional simulation model calculation domain is set to be 50m×50m×18m, the grid quantity is 12.15 million, and fig. 6 is a schematic diagram of the high-dimensional simulation model corresponding to the explosive explosion problem provided by the embodiment of the disclosure, as shown in fig. 6, 5t TNT explosive is detonated at a position 100m away from the earth dike. Meanwhile, intermittent monitoring points are set, a control equation is set to be an Euler equation considering intensity, and the preset duration is set to be 150ms.

S420, encryption processing of the low-dimensional simulation model.

Specifically, the low-dimensional simulation model may be encrypted with reference to the content described in the foregoing S120, which is not described herein.

Illustratively, according to a load matching criterion, a limit encryption criterion and a continuous medium assumption limit, 10 times of encryption processing is carried out on a low-dimensional simulation model corresponding to the explosive explosion problem, the mesh quantity of the encrypted low-dimensional simulation model is 1.215 hundred million, and the required core number is 12150.

S430, dividing the encrypted low-dimensional simulation model into parallel areas.

Specifically, the parallel region division may be performed with reference to the content described in the foregoing S131, which is not described herein.

Illustratively, the encrypted grid quantity is 1.215 hundred million of low-dimensional simulation model parallel region division.

S440, parallel computation of the low-dimensional simulation model.

Specifically, the parallel region division may be performed with reference to the content described in the foregoing S132, which is not described herein.

S450 and Gao Weichong structure condition monitoring.

Specifically, the high-dimensional reconstruction condition monitoring can be performed according to the content of the high-dimensional reconstruction condition monitored by the intermittent monitoring points with reference to the foregoing S130, which is not described herein.

Specifically, if the high-dimensional reconstruction condition is satisfied, S460 is performed; if the high-dimensional reconstruction condition is not satisfied, execution returns to S440.

S460, gao Weichong construct a map.

Specifically, the high-dimensional reconstruction can be performed with reference to the content described in the foregoing S140.

Illustratively, through S430-S460, parallel region division and parallel calculation of the encrypted low-dimensional simulation model with the grid quantity of 1.215 hundred million are completed, and when the high-dimensional reconstruction condition is met, high-dimensional reconstruction mapping is performed according to the physical problem characteristics, so that the high-dimensional simulation model containing the low-dimensional information is obtained. Fig. 7 is a diagram of a calculation result of a low-dimensional simulation model after encryption at a typical time according to an embodiment of the present disclosure. Fig. 8 is a high-dimensional simulation model containing low-dimensional information provided by an embodiment of the present disclosure.

S470, whether the encryption multiple is corrected.

Specifically, if the encryption multiple is not corrected, S480 is executed; if the encryption multiple is corrected, S490 is performed.

S480, parallel area division of the high-dimensional simulation model.

Specifically, the parallel region division may be performed with reference to the content described in the foregoing S151, which is not described herein.

Illustratively, since the encryption multiple is not corrected, the number of kernels used in the low-dimensional simulation process and the high-dimensional simulation process is the same (i.e., the first number is equal to the second number), so that the parallel region division is directly performed on the high-dimensional simulation model with the mesh number of 12.15 hundred million according to the first number.

S490, calculating a second number.

Specifically, the number of kernels used in the high-dimensional simulation process, i.e., the second number, may be calculated with reference to the content described in the second number in the foregoing S150, and will not be described herein.

Illustratively, since the encryption multiple is modified, the number of kernels used in the high-dimensional simulation process by the high-dimensional simulation model with the mesh size of 12.15 billion needs to be recalculated.

S510, dividing parallel areas of the high-dimensional simulation model.

Illustratively, parallel region partitioning is performed for a high-dimensional simulation model with a mesh size of 12.15 billion according to the second number.

S520, generating a restarting task script.

S530, automatic restarting is performed based on super-calculation scheduling.

Specifically, the restart task script and the automatic restart may be performed with reference to the content described in the foregoing S160.

S540, parallel computation of the high-dimensional simulation model.

Specifically, the parallel region division may be performed with reference to the content described in the foregoing S152, which is not described herein.

S550, whether the simulation time length is longer than a preset time length.

Specifically, it is determined whether the simulation duration is greater than a preset duration, if so, the actual number is simulated, and if not, the process returns to S540.

Illustratively, through S540-S550, parallel computation is completed for a high-dimensional simulation model with a mesh size of 12.15 hundred million. Fig. 9 is a graph of the results of a typical time-of-day high-dimensional simulation model calculation provided by an embodiment of the present disclosure, showing the three-dimensional shock wave wavefront morphology, from which it can be seen that the shock wave propagates forward along the dike and bypasses the dike after contacting the dike. Therefore, the ultra-calculation-oriented multidimensional parallel simulation method of the embodiment of the disclosure can effectively perform high-efficiency and high-precision numerical simulation on the ultra-large scale explosion problem, ensure the precision of simulation prediction while reducing the grid quantity and the calculation scale and improving the solving speed, and meet the requirements of quick response and accurate prediction.

Fig. 10 is a schematic structural diagram of a super-computing-oriented multi-dimensional parallel simulation device according to an embodiment of the present disclosure, where the super-computing-oriented multi-dimensional parallel simulation device may be understood as the electronic device or a part of functional modules in the electronic device. As shown in fig. 10, the super-computing-oriented multi-dimensional parallel simulation apparatus 100 includes:

the initialization module 1010 is configured to perform simulation initialization for the target engineering problem, where the simulation initialization includes establishing a high-dimensional simulation model and a low-dimensional simulation model corresponding to the target engineering problem;

An encryption module 1020 for encrypting the grid on the low-dimensional simulation model;

a first calculation module 1030, configured to calculate current low-dimensional data corresponding to the encrypted low-dimensional simulation model using a first number of kernels until it is monitored that the high-dimensional reconstruction condition is satisfied;

the reconstruction mapping module 1040 is configured to perform high-dimensional reconstruction on the current low-dimensional data and then map the current low-dimensional data to a high-dimensional simulation model;

the second calculating module 1050 is configured to calculate, using a second number of kernels, current high-dimensional data corresponding to the high-dimensional simulation model until the simulation duration reaches a preset duration;

In another embodiment of the present disclosure, the encryption module 1020 may include:

the first determining submodule is used for determining encryption multiples according to the grid quantity of the low-dimensional simulation model, the grid quantity of the high-dimensional simulation model and a preset scaling factor;

the first encryption submodule is used for encrypting the grids on the low-dimensional simulation model according to the encryption multiple under the condition that the encryption multiple is smaller than or equal to the limit encryption multiple and the minimum grid size corresponding to the encryption multiple meets the continuous medium assumption;

The minimum grid size corresponding to the encryption multiple is the size of the grid with the minimum size after the low-dimensional simulation model is encrypted according to the encryption multiple.

In yet another embodiment of the present disclosure, the encryption module 1020 may further include:

the first correction submodule is used for correcting the encryption multiple to be the limit encryption multiple under the condition that the encryption multiple is larger than the limit encryption multiple;

and the second encryption sub-module is used for encrypting the grid on the low-dimensional simulation model according to the corrected encryption multiple.

the second determining submodule is used for determining a correction coefficient according to the minimum grid size corresponding to the encryption multiple and the grid size under the assumption limit of the continuous medium under the condition that the minimum grid size corresponding to the encryption multiple does not meet the assumption of the continuous medium;

the second correction submodule is used for correcting the encryption multiple according to the correction coefficient;

and the third encryption sub-module is used for encrypting the grid on the low-dimensional simulation model according to the corrected encryption multiple.

In yet another embodiment of the present disclosure, the first number and the second number are each integers of 2 or more.

In yet another embodiment of the present disclosure, the reconstruction mapping module 1040 may include:

the reconstruction submodule is used for carrying out high-dimensional reconstruction on the current low-dimensional data according to a preset direction to obtain reconstruction data;

the first mapping submodule is used for mapping reconstruction data corresponding to a first region to the high-dimensional simulation model by adopting a Lagrange interpolation method, wherein the first region is a smooth region in a calculation domain where the encrypted low-dimensional simulation model is located;

and the second mapping sub-module is used for mapping the reconstruction data corresponding to the second region to the high-dimensional simulation model by adopting a weighted intrinsic oscillation-free interpolation method, wherein the second region is a region with a break in a calculation domain where the encrypted low-dimensional simulation model is located.

In yet another embodiment of the present disclosure, the apparatus further comprises:

the generating and suspending module is used for calculating current high-dimensional data corresponding to the high-dimensional simulation model by adopting a second number of kernels under the condition that the first number is not equal to the second number, until the simulation duration reaches the preset duration: generating a restarting task script, and suspending the current computing tasks of the first number of kernels;

and the execution module is used for executing the restarting task script to allocate a second number of kernels.

The device provided in this embodiment can execute the method of any one of the above embodiments, and the execution mode and the beneficial effects thereof are similar, and are not described herein again.

The embodiment of the disclosure also provides an electronic device, which comprises: a memory in which a computer program is stored; a processor for executing the computer program, which when executed by the processor can implement the method of any of the above embodiments.

By way of example, fig. 11 is a schematic structural diagram of an electronic device in an embodiment of the present disclosure. Referring now in particular to fig. 11, a schematic diagram of an electronic device 1100 suitable for use in implementing embodiments of the present disclosure is shown. The electronic device 1100 in the embodiments of the present disclosure may include, but is not limited to, mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), P second numbers P (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and the like, and stationary terminals such as digital TVs, desktop computers, and the like. The electronic device shown in fig. 11 is merely an example, and should not impose any limitations on the functionality and scope of use of embodiments of the present disclosure.

As shown in fig. 11, the electronic device 1100 may include a processing means (e.g., a central processor, a graphics processor, etc.) 1101 that may perform various appropriate actions and processes according to a program stored in a read-only memory (RO second number) 1102 or a program loaded from a storage means 1108 into a random access memory (RA second number) 1103. In the RA second number 1103, various programs and data required for the operation of the electronic device 1100 are also stored. The processing device 1101, the RO second number 1102 and the RA second number 1103 are connected to each other by a bus 1104. An input/output (I/O) interface 1105 is also connected to bus 1104.

In general, the following devices may be connected to the I/O interface 1105: input devices 1106 including, for example, a touch screen, touchpad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, and the like; an output device 1107 including, for example, a Liquid Crystal Display (LCD), a speaker, a vibrator, and the like; storage 1108, including for example, magnetic tape, hard disk, etc.; and a communication device 1109. The communication means 1109 may allow the electronic device 1100 to communicate wirelessly or by wire with other devices to exchange data. While fig. 11 illustrates an electronic device 1100 having various means, it is to be understood that not all illustrated means are required to be implemented or provided. More or fewer devices may be implemented or provided instead.

In particular, according to embodiments of the present disclosure, the processes described above with reference to flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program embodied on a non-transitory computer readable medium, the computer program comprising program code for performing the method shown in the flow chart. In such an embodiment, the computer program may be downloaded and installed from the network via the communication means 1109, or from the storage means 1108, or from the RO second number 1102. The above-described functions defined in the methods of the embodiments of the present disclosure are performed when the computer program is executed by the processing device 1101.

It should be noted that the computer readable medium described in the present disclosure may be a computer readable signal medium or a computer readable storage medium, or any combination of the two. The computer readable storage medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples of the computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RA second number), a read-only memory (RO second number), an erasable programmable read-only memory (EPRO second number or flash memory), an optical fiber, a portable compact disc read-only memory (CD-RO second number), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this disclosure, a computer-readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In the present disclosure, however, the computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, with the computer-readable program code embodied therein. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: electrical wires, fiber optic cables, RF (radio frequency), and the like, or any suitable combination of the foregoing.

In some embodiments, the client, server may communicate using any currently known or future developed network Protocol, such as HTTP (HyperText transfer Protocol) and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include local area networks ("LA first quantity"), wide area networks ("WA first quantity"), the internet (e.g., the internet), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future developed networks.

The computer readable medium may be contained in the electronic device; or may exist alone without being incorporated into the electronic device.

The computer readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to: performing simulation initialization for the target engineering problem, wherein the simulation initialization comprises the steps of establishing a low-dimensional simulation model and a high-dimensional simulation model corresponding to the target engineering problem;

encrypting the grid on the low-dimensional simulation model;

Computer program code for carrying out operations of the present disclosure may be written in one or more programming languages, including, but not limited to, an object oriented programming language such as Java, S second number alltalk, C++, and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a local area network (LA first number) or a wide area network (WA first number), or may be connected to an external computer (e.g., connected through the internet using an internet service provider).

The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

The units involved in the embodiments of the present disclosure may be implemented by means of software, or may be implemented by means of hardware. Wherein the names of the units do not constitute a limitation of the units themselves in some cases.

The functions described above herein may be performed, at least in part, by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: a Field Programmable Gate Array (FPGA), an Application Specific Integrated Circuit (ASIC), an Application Specific Standard Product (ASSP), a system on a chip (SOC), a Complex Programmable Logic Device (CPLD), and the like.

In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a random access memory (RA second amount), a read-only memory (RO second amount), an erasable programmable read-only memory (EPRO second amount or flash memory), an optical fiber, a portable compact disc read-only memory (CD-RO second amount), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

The embodiments of the present disclosure further provide a computer readable storage medium, where a computer program is stored, where the computer program, when executed by a processor, may implement a method according to any one of the foregoing embodiments, and the implementation manner and beneficial effects of the method are similar, and are not described herein again.

It should be noted that in this document, relational terms such as "first" and "second" and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.

The foregoing is merely a specific embodiment of the disclosure to enable one skilled in the art to understand or practice the disclosure. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the disclosure. Thus, the present disclosure is not intended to be limited to the embodiments shown and described herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. The multi-dimensional parallel simulation method for the super calculation is characterized by comprising the following steps of:

performing simulation initialization for a target engineering problem, wherein the simulation initialization comprises the steps of establishing a low-dimensional simulation model and a high-dimensional simulation model corresponding to the target engineering problem;

encrypting the grid on the low-dimensional simulation model;

mapping the current low-dimensional data to the high-dimensional simulation model after high-dimensional reconstruction;

2. The method of claim 1, wherein encrypting the mesh on the low-dimensional simulation model comprises:

determining encryption multiple according to the grid quantity of the low-dimensional simulation model, the grid quantity of the high-dimensional simulation model and a preset scaling factor;

under the condition that the encryption multiple is smaller than or equal to a limit encryption multiple and the minimum grid size corresponding to the encryption multiple meets the continuous medium assumption, encrypting the grids on the low-dimensional simulation model according to the encryption multiple;

and the minimum grid size corresponding to the encryption multiple is the size of the grid with the minimum size after the low-dimensional simulation model is encrypted according to the encryption multiple.

3. The method of claim 2, wherein encrypting the mesh on the low-dimensional simulation model further comprises:

correcting the encryption multiple to be the limit encryption multiple under the condition that the encryption multiple is larger than the limit encryption multiple;

And encrypting the grid on the low-dimensional simulation model according to the corrected encryption multiple.

4. The method of claim 2, wherein encrypting the mesh on the low-dimensional simulation model further comprises:

under the condition that the minimum grid size corresponding to the encryption multiple does not meet the continuous medium assumption, determining a correction coefficient according to the minimum grid size corresponding to the encryption multiple and the grid size under the continuous medium assumption limit;

correcting the encryption multiple according to the correction coefficient;

5. The method of claim 1, wherein the first number and the second number are integers greater than or equal to 2.

6. The method of claim 1, wherein said mapping said current low-dimensional data to said high-dimensional simulation model after high-dimensional reconstruction comprises:

performing high-dimensional reconstruction on the current low-dimensional data according to a preset direction to obtain reconstruction data;

mapping reconstruction data corresponding to a first region to the high-dimensional simulation model by adopting a Lagrange interpolation method, wherein the first region is a smooth region in a calculation domain where the encrypted low-dimensional simulation model is located;

And mapping reconstruction data corresponding to a second region to the high-dimensional simulation model by adopting a weighted intrinsic oscillation-free interpolation method, wherein the second region is a region with a break in a calculation domain where the encrypted low-dimensional simulation model is located.

7. The method according to claim 1, wherein, in the case that the first number is not equal to the second number, calculating, by using the second number of kernels, current high-dimensional data corresponding to the high-dimensional simulation model until the simulation duration reaches the preset duration, further includes:

generating a restarting task script, and suspending the current computing tasks of the first number of kernels;

and executing the restart task script to allocate the second number of kernels.

8. The utility model provides a multi-dimensional parallel simulation device towards super computing which characterized in that includes:

the system comprises an initialization module, a simulation module and a control module, wherein the initialization module is used for performing simulation initialization aiming at a target engineering problem, and the simulation initialization comprises the steps of establishing a high-dimensional simulation model and a low-dimensional simulation model corresponding to the target engineering problem;

The reconstruction mapping module is used for mapping the current low-dimensional data to the high-dimensional simulation model after high-dimensional reconstruction;

9. An electronic device, comprising:

a processor and a memory, wherein the memory has stored therein a computer program which, when executed by the processor, performs the method of any of claims 1-7.

10. A computer readable storage medium, characterized in that the storage medium has stored therein a computer program which, when executed by a processor, implements the method according to any of claims 1-7.