Computing resource prediction method, device, equipment, medium and program product

Abstract

The present application discloses a computing resource prediction method, device, equipment, medium and program product, which relate to the field of oil and gas seismic exploration. The method includes: obtaining operation information parameters and operation cycle, predicting the computing amount of project operations based on the operation information parameters, obtaining the operation computing amount value corresponding to the project operation, obtaining benchmark information parameters, first time consumption data, first computing power value and benchmark computing amount value, calculating the second computing power value based on the ratio relationship between the first computing power value, the operation computing amount value and the benchmark computing amount value, and calculating the number of computing nodes required to process the operation information parameters based on the ratio relationship between the first time consumption data, the operation computing amount, the second computing power value, the operation cycle, the benchmark computing amount value and the candidate computing power value. The computing node number result obtained by the above method is more reliable, can improve the accuracy and efficiency of computing resource allocation, and improve the utilization rate of computing resources.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of oil and gas seismic exploration, and in particular to a computing resource prediction method, apparatus, device, medium, and program product. Background Art

[0002] Seismic exploration is a geophysical exploration method that uses artificially generated seismic waves to record, observe, and analyze reflections from subsurface elastic interfaces, leveraging differences in the elasticity and density of underground media. Migration imaging is a method that relocates reflections from gather records to the interfaces that generated them, enabling the study of underground geological rock structure and physical properties.

[0003] Kirchhoff integration prestack time migration is the most accurate imaging method in the time domain and is currently the most commonly used migration method in the seismic data processing industry. Migration calculations are a significant step in seismic data processing projects, consuming significant computing resources and machine time. Insufficient computing resources can lead to prolonged migration, impacting project schedules; while excessive resources can lead to wasted resources and increased project costs.

[0004] Computational resources for prestack time migration using the Kirchhoff integration method are usually allocated based on manual experience, which leads to large errors, low accuracy, and low resource utilization in the resource allocation process. Summary of the Invention

[0005] The embodiments of the present application provide a computing resource prediction method, apparatus, device, medium, and program product that can predict the number of computing nodes required for project operations, thereby improving the efficiency of computing resource allocation. The technical solution is as follows:

[0006] In one aspect, a computing resource prediction method is provided, the method comprising:

[0007] Obtaining job information parameters and a job cycle, wherein the job information parameters refer to parameters involved in the project job of the computing resource to be tested, and the job cycle refers to a specified completion period of the project job;

[0008] Predicting the computational effort of the project operation based on the operation information parameters to obtain a computational effort value corresponding to the project operation, wherein the computational effort value refers to the computational effort of processing the operation information parameters during the project operation;

[0009] Obtaining a benchmark information parameter, first time consumption data, a first computing power value, and a benchmark computing amount value, wherein the benchmark information parameter refers to a parameter involved in a historical project operation, the first time consumption data refers to the time data consumed in processing the benchmark information parameter, the first computing power value refers to the computing power value required to process the benchmark information parameter, and the benchmark computing amount value refers to the computing amount in the process of processing the benchmark information parameter;

[0010] Calculating a second computing capability value based on a ratio relationship among the first computing capability value, the job computing capability value, and the reference computing capability value, wherein the second computing capability value refers to a computing capability value required for processing the job information parameter;

[0011] Based on the first time-consuming data, the job computing amount, the second computing power value, the job cycle, the ratio between the benchmark computing amount value and the candidate computing power value, the number of computing nodes required to process the job information parameters is calculated. The candidate computing power value refers to the computing power value of the computing node that processes the job information parameters.

[0012] In another aspect, a computing resource prediction device is provided, the device comprising:

[0013] An acquisition module is configured to acquire operation information parameters and an operation cycle, wherein the operation information parameters refer to parameters involved in the project operation of the computing resource to be tested, and the operation cycle refers to a specified completion cycle of the project operation;

[0014] a prediction module, which predicts the computational amount of the project operation based on the operation information parameters to obtain a computational amount value corresponding to the project operation, where the computational amount value refers to the computational amount of processing the operation information parameters during the project operation;

[0015] The acquisition module acquires a benchmark information parameter, a first time consumption data, a first computing power value, and a benchmark computing amount value, wherein the benchmark information parameter refers to a parameter involved in a historical project operation, the first time consumption data refers to the time data consumed in processing the benchmark information parameter, the first computing power value refers to the computing power value required to process the benchmark information parameter, and the benchmark computing amount value refers to the computing amount in the process of processing the benchmark information parameter;

[0016] a calculation module, configured to calculate a second computing capability value based on a ratio relationship among the first computing capability value, the job computing capability value, and the reference computing capability value, wherein the second computing capability value refers to a computing capability value required for processing the job information parameter;

[0017] The computing module calculates the number of computing nodes required to process the job information parameters based on the first time-consuming data, the job computing amount, the second computing power value, the job cycle, the benchmark computing amount value and the ratio between the candidate computing power values. The candidate computing power value refers to the computing power value of the computing node that processes the job information parameters.

[0018] On the other hand, a computer device is provided, which includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, the at least one program, the code set or the instruction set is loaded and executed by the processor to implement the computing resource prediction method as described in any of the above-mentioned embodiments of the present application.

[0019] On the other hand, a computer-readable storage medium is provided, in which at least one instruction, at least one program, a code set or an instruction set is stored. The at least one instruction, the at least one program, the code set or the instruction set is loaded and executed by a processor to implement a computing resource prediction method as described in any of the above-mentioned embodiments of the present application.

[0020] In another aspect, a computer program product or computer program is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the computing resource prediction method described in any of the above embodiments.

[0021] The beneficial effects of the technical solutions provided in the embodiments of the present application include at least:

[0022] By obtaining the operation information parameters involved in the project operation, predicting the computational amount of the project operation based on the operation information parameters, obtaining the computational amount value corresponding to the project operation, obtaining the benchmark information parameters and other information involved in the historical project operation, calculating the computing power value required to process the current project operation, and further calculating the number of computing nodes required to process the current project operation, it has guiding significance for the rational and efficient allocation of high-performance computing resources for pre-stack time migration imaging. Compared with the traditional method of allocating resources based on manual experience, the allocation quantity of computing resources is more reliable, which can improve the accuracy and efficiency of computing resource allocation and improve the utilization rate of computing resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0024] Figure 1 is a schematic diagram of the first stage of a computing resource prediction method provided by an exemplary embodiment of the present application;

[0025] Figure 2 is a schematic diagram of the second stage of a computing resource prediction method provided by an exemplary embodiment of the present application;

[0026] Figure 3 is a schematic diagram of the third stage of a computing resource prediction method provided by an exemplary embodiment of the present application;

[0027] Figure 4 is a flowchart of a computing resource prediction method provided by an exemplary embodiment of the present application;

[0028] Figure 5 This is a flowchart of a method for obtaining the number of computing nodes provided by an exemplary embodiment of the present application;

[0029] Figure 6 is a flowchart of a method for allocating based on the number of computing nodes provided by an exemplary embodiment of the present application;

[0030] Figure 7 is a structural block diagram of a computing resource allocation device provided by an exemplary embodiment of the present application;

[0031] Figure 8 is a structural block diagram of a computing resource allocation device provided by another exemplary embodiment of the present application;

[0032] Figure 9 It is a structural block diagram of a computer device provided by an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0033] In order to make the objectives, technical solutions and advantages of this application clearer, the implementation methods of this application will be further described in detail below with reference to the accompanying drawings.

[0034] Seismic exploration refers to a geophysical exploration method that uses the differences in elasticity and density of underground media caused by artificial excitation to observe and analyze the propagation patterns of seismic waves generated by artificial earthquakes underground to infer the properties and morphology of underground rock formations.

[0035] Seismic exploration is the most important geophysical method and the most effective in solving oil and gas exploration problems. It is a crucial tool for surveying oil and natural gas resources before drilling. It is also widely used in coalfield and engineering geological surveys, regional geological research, and crustal studies.

[0036] Reflection seismic analysis studies the structure and physical properties of underground geological strata by exciting elastic waves in a specific manner at the surface and recording the reflected waves from subsurface elastic interfaces within a specific surface area (aperture). Therefore, it can also be considered a backscattering problem. Due to the characteristics of reflection seismic observation, its imaging problem is divided into two steps: the first step is to record the reflected waves reaching the surface in a specific manner, and the second step is to process the observed data using a computer using specific calculation methods to produce an image of the reflecting interface that reflects the location and reflection coefficient of the subsurface geological layer. Seismic migration technology is a technique that optimizes the imaging of the reflecting interface during this second step.

[0037] Seismic migration can be performed pre-stack or post-stack. Pre-stack migration is to relocate the reflection waves in the common shot gather records or common offset gather records to the reflection interface that produced them and to converge the diffraction waves to the diffraction point that produced them. When projecting the reflection waves back to the reflection interface and converging the diffraction waves to the diffraction point, the effects of the propagation process, such as diffusion and attenuation, must be removed. Finally, a seismic waveform profile that can reflect the characteristics of the interface reflection coefficient and is correctly relocated is obtained, namely the migration profile. Post-stack migration is performed on the basis of the horizontal stacking profile. In order to solve the problem that the inclined reflection layer in the horizontal stacking profile cannot be correctly relocated and the diffraction wave cannot be fully converged, the concept of explosive reflection surface is adopted to achieve the correct relocation of the inclined reflection layer and the complete convergence of the diffraction wave.

[0038] In seismic exploration, the most commonly used migration imaging method is Kirchhoff integration prestack time migration. The advantages of Kirchhoff integration prestack time migration are its high speed and flexibility, allowing it to offset any set of samples and adapt to irregular field observation systems. Kirchhoff integration prestack time migration is based on the non-zero offset equation for point diffraction and sums the amplitudes along the diffraction travel-time trajectory at non-zero offsets.

[0039] The parameters involved in the Kirchhoff integration method prestack time migration are important factors that directly affect the computational complexity of the migration calculation process.

[0040] In recent years, with the widespread adoption of broadband, wide-azimuth, and high-density data acquisition, the volume of raw seismic data has rapidly increased, and with it the complexity of migration calculations. Migration calculations are a particularly computationally intensive and time-consuming step in seismic data processing projects. Insufficient computing resources can lead to excessive migration times, impacting project deadlines. Excessive resources, on the other hand, can lead to wasted resources, underutilization, and increased project costs. Rationally allocating high-performance computing resources based on actual needs presents a significant challenge.

[0041] Therefore, a method is needed to accurately predict the computing resources required for integral prestack time migration, providing a reference for the rational allocation of high-performance computing resources. Currently known resource allocation methods rely on experienced staff to manually allocate computing resources for project tasks. This approach can easily lead to uneven resource allocation, resulting in low efficiency and accuracy, and easily wasteful resource allocation.

[0042] In the embodiment of the present application, when performing the offset imaging link of the seismic exploration project operation, by obtaining the operation information parameters involved in the project operation, the calculation amount of the project operation is predicted based on the operation information parameters, the operation calculation value corresponding to the project operation is obtained, and the benchmark information parameters and other information involved in the historical project operation are obtained. The computing power value required for processing the current project operation is calculated, and the number of computing nodes required for processing the current project operation is further calculated, which can provide an accurate basis for computing resource allocation and improve the efficiency of computing resource allocation.

[0043] The method provided in the embodiment of the present application obtains the number of computing nodes required for the current project operation, which needs to go through the following three stages.

[0044] Indicative, such as Figure 1 As shown, Figure 1 This is a schematic diagram of the first stage of the computing resource prediction method provided in an embodiment of the present application.

[0045] The first stage is to predict the amount of calculation.

[0046] A first parameter group 100 is obtained, wherein the first parameter group 100 includes the following parameters: current project operation information parameters 101 , benchmark project operation information parameters 102 , benchmark project operation total time 103 , and benchmark project operation computing capability value 104 .

[0047] The current project operation information parameter 101 in the first parameter group 100 is input into the offset calculation amount prediction formula 110, and the output is the current project operation calculation amount 105. The benchmark project operation information parameter 102 in the first parameter group 100 is input into the offset calculation amount prediction formula 110, and the output is the benchmark project operation calculation amount 106.

[0048] After obtaining the current project operation calculation amount 105 and the benchmark project operation calculation amount 106, the first parameter group 100 is updated to obtain the second parameter group 120, wherein the second parameter group 120 includes the following parameters: current project operation information parameters 101, benchmark project operation information parameters 102, benchmark project operation total time 103, benchmark project operation computing power value 104, current project operation calculation amount 105, benchmark project operation calculation amount 106.

[0049] Indicative, such as Figure 2 As shown, Figure 2 It is a schematic diagram of the second stage of the computing resource prediction method provided in an embodiment of the present application.

[0050] The second stage is to predict the computing power value.

[0051] The reference project operation information parameters 102 in the second parameter group 120 are input into the computing capacity value prediction formula 210 , and the current project operation computing capacity value 107 is output.

[0052] After obtaining the current project operation computing power value 107, the second parameter group 120 is updated to obtain the third parameter group 130, wherein the third parameter group 130 includes the following parameters: current project operation information parameter 101, benchmark project operation information parameter 102, benchmark project operation total time 103, benchmark project operation computing power value 104, current project operation calculation amount 105, benchmark project operation calculation amount 106, current project operation computing power value 107.

[0053] Indicative, such as Figure 3 As shown, Figure 3 This is a schematic diagram of the third stage of the computing resource prediction method provided in an embodiment of the present application.

[0054] The third stage is to predict the number of computing nodes.

[0055] A fourth parameter group 140 is obtained, wherein the fourth parameter group 140 includes the following parameters: a current project operation cycle 108 and a candidate computing node computing capability value 109 .

[0056] The total time consumption of the benchmark project operation 103, the calculation amount of the current project operation 105, the calculation amount of the benchmark project operation 106, the computing capacity value of the current project operation 107 in the third parameter group 130 and the current project operation cycle 108 and the computing capacity value of the candidate computing node 109 in the fourth parameter group 140 are input into the computing node quantity prediction formula 310, and the computing node quantity 311 is output.

[0057] Combined with the above-mentioned noun introduction and implementation environment description, the computing resource prediction method provided in the embodiment of this application is described, please refer to Figure 4 , which shows a flowchart of a computing resource prediction method provided by an exemplary embodiment of the present application, and takes the method applied in a terminal as an example for explanation, such as Figure 4 As shown, the method includes:

[0058] Step 401: Obtain operation information parameters and operation cycle.

[0059] The job information parameters refer to the parameters involved in the project job of the computing resource to be tested, and the job cycle refers to the specified completion cycle of the project job.

[0060] Performing migration imaging processing on the current project requires allocating a certain amount of computing resources to the project. This project contains many parameters. The job information parameters are crucial for determining the computational load of the project. A higher computational load indicates a higher need for computing resources.

[0061] The operation cycle of the current project operation is generally specified, and the project operation needs to be completed within the operation cycle.

[0062] Optionally, the job information parameters include seven types of parameters: first aperture, second aperture, CMP line number, CMP dot number, sample dot number, CMP line pitch, and CMP dot pitch.

[0063] The migration aperture refers to the distribution range of the seismic data used for migration imaging. For a specific imaging point, the migration aperture refers to the number of seismic traces around it that participate in the diffraction stacking. When the migration aperture is too small, the migration results can only ensure the imaging of low-angle structures and high signal-to-noise ratio characteristics, but cannot ensure the imaging of steep-angle structures. When the migration aperture is too large, the migration can ensure the imaging of steep-angle structures, but the phase axis continuity of the profile will be reduced, the resolution and signal-to-noise ratio will be reduced, and a large amount of computation time will be consumed.

[0064] The first aperture is also called the X aperture, which is the half aperture value in the X direction. The second aperture is also called the Y aperture, which is the half aperture value in the Y direction (X and Y represent two directions, the CMP line direction and the CMP point direction). The two values ​​can be the same or different.

[0065] CMP is the abbreviation of Common Middle Point, that is, CMP data set, which means extracting data with a common center point from different shot sets to form a new set.

[0066] The number of CMP lines (CMP Lines) refers to the number of CMP bus lines of the offset output. The number of CMP points (CMP Points) refers to the number of CMP points contained in a CMP line of the offset output. The CMP line interval (CMP Line Interval) refers to the distance between two adjacent CMP lines in the CMP grid. The CMP point interval (CMP Point Interval) refers to the distance between two adjacent CMP points in the CMP grid.

[0067] The number of samples refers to the number of sampling points for each data channel.

[0068] The value corresponding to each parameter in the job information parameters is preset in advance.

[0069] Optionally, the operation cycle of the current project operation is 30 days.

[0070] The operation cycle of the current project work is also preset in advance, based on the urgency of the current project work and the completion deadline specified when the current project work is arranged.

[0071] It is worth noting that the numerical value of the job cycle of the current project job can be arbitrary, the cycle unit of the job cycle of the current project job can be arbitrary, the numerical value corresponding to each parameter in the job information parameters of the current project job can be arbitrary, the unit of each parameter in the job information parameters of the current project job can be arbitrary, and the job information parameters of the current project job can contain any number and type of parameter information, but only the above seven parameters are more important for the impact on computing resources, and this embodiment does not limit this.

[0072] Step 402: Predict the computational amount of the project operation based on the operation information parameters to obtain the computational amount value corresponding to the project operation.

[0073] The operation calculation value refers to the calculation amount generated when processing the operation information parameters during the project operation.

[0074] Predicting the computational workload of current project tasks can help better allocate computing resources.

[0075] A computational workload prediction model is established, the job information parameters of the current project job are input into the computational workload prediction model, and the output is the computational workload value corresponding to the current job project.

[0076] The formula corresponding to the calculation amount prediction model is as follows: Formula 1:

[0077] Cx=(π*Ax*Ay*Nl*Np*Nsp) / (Dl*Dp)

[0078] Among them, Ax is the first aperture, Ay is the second aperture, Nl is the number of CMP lines, Np is the number of CMP points, Nsp is the number of sample points, Dl is the CMP line spacing, Dp is the CMP point spacing, Cx is the calculation value corresponding to the current job project, and π is the coefficient constant.

[0079] Among them, for the parameters involved in Formula 1, the subscript x represents that the direction of the aperture is the x direction, the subscript y represents that the direction of the aperture is the y direction; the subscript l represents the number of lines, and l is the abbreviation of line; the subscript sp represents the number of samples, and sp is the abbreviation of sample; the subscript p represents the number of points, and p is the abbreviation of point.

[0080] Input one of the parameters in the operation information parameters into the formula corresponding to the above-mentioned calculation amount prediction model, and output the predicted operation calculation amount value.

[0081] It is worth noting that the formula corresponding to the above-mentioned computational quantity prediction model can not only calculate the computational quantity value of the current project job, but also calculate the computational quantity values ​​corresponding to other project jobs; the above-mentioned method of predicting the computational quantity of project jobs can be arbitrary, that is, including but not limited to inputting the job information parameters into the computational quantity prediction model; the coefficient constants involved in the formula corresponding to the above-mentioned computational quantity prediction model can be any numerical value that conforms to the actual situation, and the formula corresponding to the computational quantity prediction model includes but is not limited to Formula 1 in the above example; when inputting the job information parameters into the formula corresponding to the computational quantity prediction model, the numerical value and unit corresponding to each parameter in the job information parameters can be arbitrary, and this embodiment does not limit this.

[0082] Step 403: Obtain benchmark information parameters, first time consumption data, first computing capability value, and benchmark computing value.

[0083] Among them, the benchmark information parameters refer to the parameters involved in historical project operations, the first time-consuming data refers to the time data consumed in processing the benchmark information parameters, the first computing power value refers to the computing power value required to process the benchmark information parameters, and the benchmark computing value refers to the computing amount in the process of processing the benchmark information parameters.

[0084] Optionally, the baseline information parameters are a set of information parameters randomly obtained from the information parameter library, and the baseline information parameters correspond to the parameters involved in the first historical project operation. The information parameter library contains multiple sets of information parameter data, including but not limited to information parameter groups, and other types of information, wherein the information parameter groups are all measured in different historical time periods.

[0085] During the historical time period, when the first historical project operation is processed, the values ​​and units corresponding to all parameters in the benchmark information parameters are recorded, and the time consumed in the processing is recorded, that is, the first time-consuming data.

[0086] Optionally, the information parameter library also includes a first computing power value, which refers to the computing power value required when processing the benchmark information parameters, that is, the computing power value of the computing cluster used when processing the first historical project job.

[0087] Optionally, the computing resource used when processing the first historical project job is a first computing cluster, which includes at least one first computing node. Each first computing node in the first computing cluster is the same, that is, each first computing node has the same parameters, namely, the first computing parameters.

[0088] The computing parameter is an important parameter for measuring the computing capability of each computing node. The first computing parameter is used to measure the computing capability of each first computing node. The computing capability of all first computing nodes represents the first computing capability of the first computing cluster.

[0089] A first computing parameter of a first computing cluster is obtained, and a first computing capability value is calculated based on the first computing parameter. The first computing cluster refers to a collection of computing nodes that processes the reference information parameters.

[0090] Based on the benchmark information parameters, the computational load of the historical project operation is predicted to obtain the benchmark computational load corresponding to the historical project operation. Similar to step 402 above, the seven parameters in the benchmark information parameters are input into Formula 1 corresponding to the computational load prediction model to obtain the benchmark computational load corresponding to the first historical project operation.

[0091] It is worth noting that the baseline information parameters can be information parameters obtained when processing any historical project job within any historical time period. A group of information parameters are randomly obtained from the information parameter library as baseline information parameters, that is, any group of information parameters in the information parameter library can be used as baseline information parameters. The information parameter library can contain any number of information parameter groups. The information parameter groups contained in the information parameter library can be information parameters obtained when processing any historical project job within any historical time period. This embodiment does not limit this.

[0092] It is worth noting that the computing cluster used to process the above historical project jobs can be of any type, the number of computing nodes in the computing cluster can be any, and the computing parameters corresponding to the computing nodes can be any, which are not limited in this embodiment.

[0093] Step 404 : Based on the product of the first computing capability value and the operation computing capability value, the second computing capability value is calculated by performing a quotient operation on the reference computing capability value.

[0094] The second computing capability value refers to the computing capability value required to process the operation information parameters. The first computing capability value, the operation calculation value, and the reference calculation value are input into the computing capability prediction model to output the second computing capability value.

[0095] The formula corresponding to the computing power prediction model is as follows: Formula 2:

[0096] Fx=Frv*Cx / Crv

[0097] Among them, Fx is the second computing capability value, Cx is the job computing value of the current project job, Frv is the first computing capability value, and Crv is the benchmark computing value.

[0098] Specifically, the first computing capability value and the job computing value are multiplied to obtain a first product; the first product is divided by the benchmark computing value to obtain a second computing capability value, wherein the first product is the numerator and the benchmark computing value is the denominator.

[0099] Step 405 , based on the product of the first time consumption data, the job computing amount and the second computing capacity value, the product of the job cycle, the benchmark computing amount value and the candidate computing capacity value is divided to calculate the number of computing nodes required to process the job information parameters.

[0100] The candidate computing capability value refers to the computing capability value of the computing node that processes the job information parameters.

[0101] The first time consumption data, the operation calculation amount, the product of the second computing capacity value, the operation cycle, the benchmark computing capacity value and the candidate computing capacity value are input into the computing node quantity prediction model, and the computing node quantity is output.

[0102] The formula corresponding to the calculation node quantity prediction model is as follows: Formula 3:

[0103] N=(Trv*Cx*Fx) / (t*Crv*Fy)

[0104] Among them, N is the number of computing nodes, Trv is the first time-consuming data, Cx is the job computing amount of the current project job, Fx is the second computing power value, t is the job cycle, Crv is the benchmark computing amount value, and Fy is the candidate computing power value.

[0105] It is worth noting that the formula corresponding to the above-mentioned computing node quantity prediction model can not only calculate the number of computing nodes for the current project job, but also calculate the number of computing nodes corresponding to other project jobs; the above-mentioned method of predicting the number of computing nodes for project jobs can be arbitrary, that is, including but not limited to inputting job information parameters into the computing node quantity prediction model; the formula corresponding to the computing node quantity prediction model includes but is not limited to Formula 3 in the above example; when the above-mentioned parameters are input into the formula corresponding to the computing node quantity prediction model, the numerical value and unit corresponding to each parameter can be arbitrary, and this embodiment does not limit this.

[0106] In summary, the method provided in the embodiment of the present application obtains the operation information parameters involved in the project operation, predicts the computing amount of the project operation based on the operation information parameters, obtains the operation computing amount value corresponding to the project operation, obtains the benchmark information parameters and other information involved in the historical project operation, calculates the computing power value required for processing the current project operation, and further calculates the number of computing nodes required for processing the current project operation. It has guiding significance for the reasonable and efficient allocation of high-performance computing resources for pre-stack time migration imaging. Compared with the traditional method of allocating resources based on manual experience, the allocation quantity result of the computing resources is more reliable, can improve the accuracy and efficiency of computing resource allocation, and improve the utilization rate of computing resources.

[0107] The method provided in this embodiment obtains the benchmark information parameters and the first time-consuming data corresponding to the historical project operation, inputs the benchmark information parameters into the calculation amount prediction formula, and obtains the benchmark calculation amount value corresponding to the historical project operation, which provides reference data for the operation calculation amount of the current project operation, making the predicted calculation amount value result more accurate.

[0108] The method provided in this embodiment calculates the second computing power value by taking the quotient of the benchmark computing value based on the product of the first computing power value and the job computing value, and can accurately estimate the computing power value of the computing node required to process the current project job.

[0109] The method provided in this embodiment establishes a calculation quantity prediction model, inputs the operation information parameters into the calculation quantity prediction model, multiplies or divides all parameters in the operation information parameters, and obtains the operation calculation quantity value corresponding to the project operation. The operation calculation quantity value obtained by the above method is more accurate and improves the efficiency of estimation.

[0110] In some embodiments, when allocating computing resources for a project job, it is necessary to preset the type of computing cluster in order to obtain the number of computing nodes required to process the job information parameters. Figure 5 This is a flowchart of a method for obtaining the number of computing nodes provided by an exemplary embodiment of the present application. Figure 5As shown, the method includes the following steps.

[0111] Step 501: Obtain a computing cluster sample library.

[0112] The computing cluster sample library refers to a library containing computing clusters of different types. The computing cluster sample library contains at least one computing cluster; each sample cluster contains at least one computing node, and each computing node is of the same type.

[0113] Step 502: Determine a computing cluster in the computing cluster sample library as a candidate computing cluster.

[0114] Optionally, the computing cluster sample library includes five computing clusters, namely: a first computing cluster, a second computing cluster, a third computing cluster, a fourth computing cluster, and a fifth computing cluster.

[0115] Optionally, the fifth computing cluster is selected as the candidate computing cluster, the fifth computing cluster includes five fifth computing nodes, and each fifth computing node is of the same type.

[0116] It is worth noting that the computing cluster sample library can contain any number and type of computing clusters, the number of computing nodes contained in each computing cluster can be arbitrary, the computing clusters selected as candidate computing clusters in the computing cluster sample library can be arbitrary, the type of candidate computing clusters can be arbitrary, and the number of computing nodes in the candidate computing cluster can be arbitrary, and this embodiment does not limit this.

[0117] Step 503: Obtain candidate computing parameters of the candidate computing cluster, and calculate the candidate computing capability value based on the candidate computing parameters.

[0118] The candidate computing cluster includes at least one candidate computing node, and the candidate computing nodes are all of the same type. Each candidate computing node has the same candidate computing parameters, and the candidate computing parameters of each candidate computing node represent the computing capability value of each candidate computing node, which together constitute the computing capability value of the candidate computing cluster.

[0119] Step 504: multiply the first time consumption data, the operation calculation amount, and the second computing capability value to obtain a second product.

[0120] The same as in step 405 above, the first time consumption data, the operation calculation amount and the second computing capacity value are input into formula three.

[0121] Step 505 : multiply the operation cycle, the benchmark computing value, and the candidate computing capability value to obtain a third product.

[0122] The same as in step 405 above, the operation cycle, the benchmark computing value and the candidate computing capability value are input into the step of formula three.

[0123] Step 506: Divide the second product by the third product to obtain the number of computing nodes required to process the job information parameters.

[0124] Among them, the second product is the numerator and the third product is the denominator.

[0125] The process of obtaining the number of computing nodes is the same as that in step 405 above.

[0126] To summarize, by obtaining a computing cluster sample library, a computing cluster is determined in the computing cluster sample library as a candidate computing cluster; based on the candidate computing parameters of the candidate computing cluster, the candidate computing capacity is calculated, and further based on the first time consumption data, the job computing amount, the second computing capacity value, the first time consumption data, the job computing amount and the second computing capacity value, the number of computing nodes required to process the job information parameters is calculated, thereby improving the efficiency of computing resource allocation.

[0127] The method provided in this embodiment obtains a computing cluster sample library and determines a computing cluster in the computing cluster sample library as a candidate computing cluster; based on the candidate computing parameters of the candidate computing cluster, the candidate computing capacity is calculated, and an accurate candidate computing capacity estimation result can be obtained. By comparing it with the computing capacity value required to process the current project job, an accurate computing resource allocation result can be obtained.

[0128] The method provided in this embodiment obtains candidate computing parameters of a candidate computing cluster, calculates a candidate computing capacity value based on the candidate computing parameters, multiplies the first time-consuming data, the job computing amount, and the second computing capacity value to obtain a second product; multiplies the job cycle, the benchmark computing amount value, and the candidate computing capacity value to obtain a third product; divides the second product by the third product to obtain the number of computing nodes required to process the job information parameters, thereby improving the efficiency of computing resource allocation.

[0129] In some embodiments, in order to improve the efficiency of resource allocation and maximize the utilization of computing resources allocated to each project job, the computing parameters of different types of candidate computing clusters will be calculated separately to determine how many computing nodes each computing cluster requires, so as to reasonably allocate computing resources. Figure 6 FIG. 1 is a flow chart of a method for allocating data based on the number of computing nodes provided by an exemplary embodiment of the present application. Figure 6 As shown, the method includes the following steps.

[0130] Step 601: Obtain a computing cluster sample library.

[0131] Same as step 501 above.

[0132] Step 602: Determine at least one computing cluster in the computing cluster sample library as a candidate computing cluster.

[0133] Optionally, the computing cluster sample library includes three types of computing clusters:

[0134] Computing cluster A: includes 4 Class A computing nodes;

[0135] Computing cluster B: includes 5 Class B computing nodes;

[0136] Computing cluster C: includes 6 C-type computing nodes.

[0137] Optionally, all of the above three types of computing clusters are selected as candidate computing clusters.

[0138] It is worth noting that the computing cluster sample library can contain any number and type of computing clusters, the number of computing nodes contained in each computing cluster can be arbitrary, the computing clusters selected as candidate computing clusters in the computing cluster sample library can be arbitrary, the type and number of candidate computing clusters can be arbitrary, and the number of computing nodes in the candidate computing cluster can be arbitrary, and this embodiment does not limit this.

[0139] Step 603 : Based on the candidate computing parameters of each candidate computing cluster, a candidate computing capability value corresponding to each candidate computing cluster is calculated.

[0140] Optionally, the candidate computing capability value corresponding to computing cluster A is 6, the candidate computing capability value corresponding to computing cluster B is 5, and the candidate computing capability value corresponding to computing cluster C is 4.

[0141] It is worth noting that the candidate computing parameters of each candidate computing cluster can be arbitrary, that is, the numerical values ​​and units of the candidate computing parameters can be arbitrary, and the candidate computing capability values ​​obtained based on the candidate computing parameters of each candidate computing cluster can be arbitrary, that is, the numerical values ​​and units of the candidate computing capability values ​​can be arbitrary. This embodiment does not limit this.

[0142] Step 604 : Based on the first time consumption data, the job computing amount, the second computing capacity value, the job cycle, the benchmark computing amount value, and the candidate computing capacity value, the number of computing nodes corresponding to each candidate computing node is calculated.

[0143] The process of calculating the number of computing nodes corresponding to each candidate computing node is the same as steps 504 to 506 above.

[0144] Optionally, the computing volume of the current project job is 25 unit values. When processing the current project job, if computing cluster A is selected as the candidate computing cluster, the candidate computing capability value of each Class A computing node is 6, then 5 Class A computing nodes are required, and there are only 4 Class A computing nodes in computing cluster A. Therefore, when processing the current project job, the number of computing nodes in computing cluster A is insufficient; if computing cluster B is selected as the candidate computing cluster, the candidate computing capability value of each Class B computing node is 5, then 5 Class B computing nodes are required, and there are a total of 5 Class B computing nodes in computing cluster B. Therefore, when processing the current project job, the number of computing nodes in computing cluster B is sufficient; if computing cluster C is selected as the candidate computing cluster, the candidate computing capability value of each Class C computing node is 4, then 7 Class C computing nodes are required, and there are only 6 Class C computing nodes in computing cluster C. Therefore, when processing the current project job, the number of computing nodes in computing cluster C is insufficient; in summary, computing cluster B can be selected as the computing cluster to process the current project job.

[0145] It is worth noting that when different types of computing clusters are selected as candidate computing clusters to process the current project job, the number of computing nodes required can be arbitrary. The above number of computing nodes is only an example. The computing amount of the current project job can be arbitrary, and the unit value corresponding to the computing amount can be arbitrary. This embodiment does not limit this.

[0146] Step 605 : Allocate computing resources to the project job based on the number of computing nodes corresponding to each candidate computing node.

[0147] Optionally, the known computing cluster sample library includes three types of computing clusters:

[0148] Computing cluster A: includes 4 Class A computing nodes; each Class A computing node has a candidate computing capability value of 6;

[0149] Computing cluster B: includes 5 Class B computing nodes, each of which has a candidate computing capability of 5.

[0150] Computing cluster C: includes 6 Class C computing nodes, and the candidate computing capability value of each Class C computing node is 4.

[0151] Optionally, the calculation amount of the current project job is 25 units of value.

[0152] After calculation, if computing cluster A is selected as the candidate computing cluster, 5 Class A computing nodes are required; if computing cluster B is selected as the candidate computing cluster, 5 Class B computing nodes are required; if computing cluster C is selected as the candidate computing cluster, 7 Class C computing nodes are required.

[0153] There are not enough Class A computing nodes in computing cluster A, and there are not enough Class C computing nodes in computing cluster C. Therefore, computing cluster B is selected as the candidate computing cluster. Using computing cluster B requires occupying 5 Class B computing nodes to allocate resources for project jobs, that is, allocating 5 Class B computing nodes to the current project jobs for processing.

[0154] To sum up, the method provided in the embodiment of the present application calculates the computing power value of each computing cluster in the computing cluster sample library, and further calculates the number of computing nodes required to use each computing cluster to process the current project job, compares them, and reasonably allocates resources for the project job, thereby improving the utilization rate of computing resources.

[0155] Figure 7 This is a structural block diagram of a computing resource allocation device provided by an exemplary embodiment of the present application. Figure 7 As shown, the device includes:

[0156] An acquisition module 710 is configured to acquire job information parameters and a job cycle, wherein the job information parameters refer to parameters involved in the project job of the computing resource to be tested, and the job cycle refers to a specified completion period of the project job;

[0157] A prediction module 720 is configured to predict the computational effort of the project operation based on the operation information parameters to obtain a computational effort value corresponding to the project operation, where the computational effort value refers to the computational effort of processing the operation information parameters during the project operation.

[0158] The acquisition module 710 is further configured to acquire a benchmark information parameter, a first time consumption data, a first computing power value, and a benchmark computing amount value, wherein the benchmark information parameter refers to a parameter involved in a historical project operation, the first time consumption data refers to the time data consumed in processing the benchmark information parameter, the first computing power value refers to the computing power value required to process the benchmark information parameter, and the benchmark computing amount value refers to the amount of computing required in processing the benchmark information parameter;

[0159] a calculation module 730 configured to calculate a second computing capability value based on a ratio relationship among the first computing capability value, the job computing capability value, and the reference computing capability value, wherein the second computing capability value refers to a computing capability value required for processing the job information parameter;

[0160] The calculation module 730 is also used to calculate the number of computing nodes required to process the job information parameters based on the first time-consuming data, the job calculation amount, the second computing power value, the job cycle, the benchmark computing value and the ratio between the candidate computing power value, and the candidate computing power value refers to the computing power value of the computing node that processes the job information parameters.

[0161] In an optional embodiment, the acquisition module 710 is also used to obtain the benchmark information parameters and the first time-consuming data; perform a computing amount prediction on the historical project job based on the benchmark information parameters to obtain a benchmark computing amount value corresponding to the historical project job; obtain a first computing parameter of a first computing cluster, and calculate the first computing power value based on the first computing parameter, where the first computing cluster refers to a collection of computing nodes that process the benchmark information parameters.

[0162] In an optional embodiment, the calculation module 730 is further used to multiply the first computing power value and the operation computing value to obtain a first product; and divide the first product by the reference computing value to obtain the second computing power value.

[0163] In an optional embodiment, if Figure 8 As shown, the calculation module 730 further includes:

[0164] An acquiring unit 731 is configured to acquire candidate computing parameters of a candidate computing cluster, and calculate the candidate computing capability value based on the candidate computing parameters, wherein the candidate computing cluster includes at least one candidate computing node, and the candidate computing nodes all belong to the same type;

[0165] a multiplication unit 732 configured to multiply the first time consumption data, the job calculation value, and the second computing capability value to obtain a second product;

[0166] The multiplication unit 732 is further configured to multiply the operation cycle, the benchmark computing value, and the candidate computing capability value to obtain a third product;

[0167] The division unit 733 is configured to divide the second product by the third product to obtain the number of computing nodes required to process the job information parameters.

[0168] In an optional embodiment, the acquisition unit 731 is further used to obtain a computing cluster sample library, which contains at least one computing cluster; in the computing cluster sample library, a computing cluster is determined as the candidate computing cluster; based on the candidate computing parameters of the candidate computing cluster, the candidate computing capability is calculated.

[0169] In an optional embodiment, the prediction module 720 is further used to establish a calculation amount prediction model; input the operation information parameters into the calculation amount prediction model to obtain the operation calculation amount value corresponding to the project operation.

[0170] In summary, the device provided in this embodiment obtains the operation information parameters involved in the project operation, predicts the computing amount of the project operation based on the operation information parameters, obtains the operation computing amount value corresponding to the project operation, obtains the benchmark information parameters and other information involved in the historical project operation, calculates the computing power value required to process the current project operation, and further calculates the number of computing nodes required to process the current project operation. It has guiding significance for the reasonable and efficient allocation of high-performance computing resources for pre-stack time migration imaging. Compared with the traditional method of allocating resources based on manual experience, the allocation quantity result of the computing resources is more reliable, which can improve the accuracy and efficiency of computing resource allocation and improve the utilization rate of computing resources.

[0171] It should be noted that the computing resource prediction device provided in the above embodiment is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0172] Figure 9 The following is a block diagram of a computer device 900 according to an exemplary embodiment of the present application. The computer device 900 may be a laptop or a desktop computer. The computer device 900 may also be referred to as a user device, a portable terminal, a laptop terminal, a desktop terminal, or other names.

[0173] Typically, the computer device 900 includes a processor 901 and a memory 902 .

[0174] The processor 901 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 901 may be implemented in at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), or PLA (Programmable Logic Array). The processor 901 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the awake state, also known as a CPU (Central Processing Unit); the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor 901 may be integrated with a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 901 may also include an AI processor, which is used to process computing operations related to machine learning.

[0175] The memory 902 may include one or more computer-readable storage media, which may be non-transitory. The memory 902 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices and flash memory storage devices. In some embodiments, the non-transitory computer-readable storage medium in the memory 902 is used to store at least one instruction, which is executed by the processor 901 to implement the computing resource prediction method provided in the method embodiment of the present application.

[0176] In some embodiments, the computer device 900 further includes other components, which can be understood by those skilled in the art. Figure 9 The structure shown in the figure does not constitute a limitation on the terminal 900, and the terminal 900 may include more or fewer components than shown in the figure, or combine certain components, or adopt a different component arrangement.

[0177] An embodiment of the present application also provides a computer device, which includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, code set or instruction set, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by the processor to implement the computing resource prediction method provided by the above-mentioned method embodiments.

[0178] An embodiment of the present application also provides a computer-readable storage medium, which stores at least one instruction, at least one program, code set or instruction set, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by a processor to implement the computing resource prediction method provided by the above-mentioned method embodiments.

[0179] Embodiments of the present application further provide a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the computing resource prediction method described in any of the above embodiments.

[0180] Optionally, the computer-readable storage medium may include: a read-only memory (ROM), a random access memory (RAM), a solid-state drive (SSD), or an optical disk. Among them, the random access memory may include a resistance random access memory (ReRAM) and a dynamic random access memory (DRAM). The serial numbers of the above embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0181] Those skilled in the art will understand that all or part of the steps to implement the above embodiments may be accomplished by hardware, or by a program to instruct the relevant hardware, and the program may be stored in a computer-readable storage medium, which may be a read-only memory, a disk, or an optical disk, etc.

[0182] The above description is merely an optional embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A computing resource prediction method, characterized in that: The method comprises: Obtaining job information parameters and a job cycle, wherein the job information parameters refer to parameters involved in the project job of the computing resource to be tested, and the job cycle refers to a specified completion period of the project job; Predicting the computational effort of the project operation based on the operation information parameters to obtain a computational effort value corresponding to the project operation, wherein the computational effort value refers to the computational effort of processing the operation information parameters during the project operation; Obtaining a benchmark information parameter, first time consumption data, a first computing power value, and a benchmark computing amount value, wherein the benchmark information parameter refers to a parameter involved in a historical project operation, the first time consumption data refers to the time data consumed in processing the benchmark information parameter, the first computing power value refers to the computing power value required to process the benchmark information parameter, and the benchmark computing amount value refers to the computing amount in the process of processing the benchmark information parameter; Calculating a second computing capability value based on a ratio relationship among the first computing capability value, the job computing capability value, and the reference computing capability value, wherein the second computing capability value refers to a computing capability value required for processing the job information parameter; Based on the first time-consuming data, the job computing amount, the second computing power value, the job cycle, the ratio between the benchmark computing amount value and the candidate computing power value, the number of computing nodes required to process the job information parameters is calculated. The candidate computing power value refers to the computing power value of the computing node that processes the job information parameters.

2. The method according to claim 1, characterized in that The obtaining of the benchmark information parameter, the first time consumption data, the first computing capability value, and the benchmark computing value includes: Acquire the benchmark information parameters and the first time-consuming data; Performing a computational load prediction on the historical project operation based on the benchmark information parameters to obtain a benchmark computational load value corresponding to the historical project operation; A first computing parameter of a first computing cluster is obtained, and the first computing capability value is calculated based on the first computing parameter. The first computing cluster refers to a set of computing nodes that processes the benchmark information parameter.

3. The method according to claim 1, characterized in that The calculating the second computing capability value based on the ratio relationship among the first computing capability value, the job computing capability value, and the benchmark computing capability value includes: multiplying the first computing capability value and the operation calculation amount value to obtain a first product; The first product is divided by the reference computing value to obtain the second computing capability value.

4. The method according to claim 1, wherein The calculating the number of computing nodes required to process the job information parameters based on the first time consumption data, the job computing amount, the second computing capability value, the job cycle, the benchmark computing amount value, and the ratio between the candidate computing capability values ​​includes: Obtaining candidate computing parameters of a candidate computing cluster, and calculating the candidate computing capability value based on the candidate computing parameters, wherein the candidate computing cluster includes at least one candidate computing node, and the candidate computing nodes all belong to the same type; multiplying the first time consumption data, the job calculation value, and the second computing power value to obtain a second product; multiplying the operating cycle, the benchmark computing value, and the candidate computing capability value to obtain a third product; The second product is divided by the third product to obtain the number of computing nodes required to process the job information parameters.

5. The method according to claim 4, characterized in that The acquiring candidate computing parameters of the candidate computing cluster, and calculating the candidate computing capability value based on the candidate computing parameters, further includes: Obtaining a computing cluster sample library, wherein the computing cluster sample library includes at least one computing cluster; Determining a computing cluster in the computing cluster sample library as the candidate computing cluster; The candidate computing capability is calculated based on the candidate computing parameters of the candidate computing cluster.

6. The method according to claim 1, characterized in that The performing of a computational amount prediction on the project operation based on the operation information parameters to obtain a computational amount value corresponding to the project operation includes: Establish a computational workload prediction model; The operation information parameters are input into the calculation amount prediction model to obtain the operation calculation amount value corresponding to the project operation.

7. A computing resource prediction device, characterized in that: The device comprises: An acquisition module is configured to acquire operation information parameters and an operation cycle, wherein the operation information parameters refer to parameters involved in the project operation of the computing resource to be tested, and the operation cycle refers to a specified completion cycle of the project operation; a prediction module, which predicts the computational amount of the project operation based on the operation information parameters to obtain a computational amount value corresponding to the project operation, where the computational amount value refers to the computational amount of processing the operation information parameters during the project operation; The acquisition module acquires a benchmark information parameter, a first time consumption data, a first computing power value, and a benchmark computing amount value, wherein the benchmark information parameter refers to a parameter involved in a historical project operation, the first time consumption data refers to the time data consumed in processing the benchmark information parameter, the first computing power value refers to the computing power value required to process the benchmark information parameter, and the benchmark computing amount value refers to the computing amount in the process of processing the benchmark information parameter; a calculation module, configured to calculate a second computing capability value based on a ratio relationship among the first computing capability value, the job computing capability value, and the reference computing capability value, wherein the second computing capability value refers to a computing capability value required for processing the job information parameter; The computing module calculates the number of computing nodes required to process the job information parameters based on the first time-consuming data, the job computing amount, the second computing power value, the job cycle, the benchmark computing amount value and the ratio between the candidate computing power values. The candidate computing power value refers to the computing power value of the computing node that processes the job information parameters.

8. A computer device, characterized in that: The computer device includes a processor and a memory, wherein the memory stores at least one program, and the at least one program is loaded and executed by the processor to implement the computing resource prediction method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that The storage medium stores at least one program, and the at least one program is loaded and executed by the processor to implement the computing resource prediction method according to any one of claims 1 to 6.

10. A computer program product, characterized in that The method comprises a computer program, which implements the computing resource prediction method according to any one of claims 1 to 6 when the computer program is executed by a processor.

Images