A method for ladder operation of data and a terminal

Abstract

The application provides a kind of data ladder operation method and terminal, through establishing transaction pool, the operation data and its operation formula of user dispatch to distribution server are carried out hierarchical disassembly according to operation order, i.e.the whole complex operation process is decomposed into multiple operation steps, and different computers are matched for each hierarchical operation step to calculate in turn, i.e.the complex operation process is dispersed to different computers for hierarchical processing, which is beneficial to improve data calculation efficiency, reduce CPU load of each computer and improve utilization of idle computer power;At the same time, hierarchical operation grading can push the formula and data required by the data in hierarchical, and each level is responsible for a data, which also ensures data privacy and solves the problem of data isolation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a data ladder operation method and terminal. BACKGROUND

[0002] Nowadays, data computing such as cloud computing and online table computing of Feishu, etc. usually utilizes the computing power of servers to help users to perform data computing, which needs to upload user tables to the server, and is easy to increase the risk of user privacy data; at the same time, it needs to rent idle computing power of servers to perform data computing, which has large computing amount and leads to low computing efficiency.

[0003] Therefore, how to effectively ensure data privacy while effectively improving the utilization rate of idle computer computing power to improve computing efficiency has gradually become a technical problem to be solved in the field of data operation. SUMMARY

[0004] The technical problem to be solved by the present application is to provide a data ladder operation method and terminal, which effectively solves the data isolation problem and improves the CPU computing power utilization rate of idle computers.

[0005] In order to solve the above technical problems, the technical scheme adopted by the present application is:

[0006] A data ladder operation method, comprising the steps of:

[0007] S1, establishing a transaction pool and obtaining operation data and operation formula of a user dispatched to a distribution server;

[0008] S2, hierarchically decomposing the operation data and operation formula according to the operation sequence to obtain hierarchical data and hierarchical formula;

[0009] S3, distributing each hierarchical data and its corresponding hierarchical formula to different computers and sequentially calculating according to the operation sequence;

[0010] S4, after the calculation is completed, calling a transaction callback mechanism to output the calculation result and return it to the corresponding user.

[0011] In order to solve the above technical problems, another technical scheme adopted by the present application is:

[0012] A data ladder operation terminal, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program:

[0013] S1, establishing a transaction pool and obtaining operation data and operation formula of a user dispatched to a distribution server;

[0014] S2, hierarchical decomposition is performed on the operation data and operation formula according to the operation sequence to obtain hierarchical data and hierarchical formula;

[0015] S3, each hierarchical data and the corresponding hierarchical formula are distributed to different computers and sequentially calculated according to the operation sequence;

[0016] S4, after the calculation is completed, a transaction callback mechanism is called to output the calculation result and return it to the corresponding user.

[0017] The data ladder operation method and terminal provided by the application have the beneficial effects that: the operation data and operation formula distributed to the distribution server by the transaction pool are hierarchically decomposed according to the operation sequence, that is, the entire complex operation process is decomposed into multiple hierarchical operation steps, and different computers are matched for each hierarchical operation step to sequentially calculate, that is, the complex operation process is dispersed to different computers for hierarchical processing, which is beneficial to improve the data calculation efficiency, reduce the CPU load of each computer, and improve the utilization rate of idle computer computing power; meanwhile, hierarchical operation grading can push the formula and data required by the data in a hierarchical manner, each level is responsible for a data, which ensures data privacy and solves the data isolation problem. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 The overall flowchart of the data ladder operation method of the embodiment of the application;

[0019] Figure 2 The structural schematic diagram of the data ladder operation terminal of the embodiment of the application.

[0020] Label explanation:

[0021] 1, a data ladder operation terminal; 2, a memory; 3, a processor. DETAILED DESCRIPTION

[0022] To explain the technical content, purposes and effects of the application in detail, the following will be described in combination with the embodiments and the accompanying drawings.

[0023] Before that, the English terms appearing in this text are explained as follows:

[0024] FLOPS: Floating-point Operations Per Second, the number of floating-point operations per second, which can be used to evaluate the performance of a computer, especially in the field of scientific computing using a large number of floating-point operations.

[0025] Please refer to Figure 1 A data ladder operation method, comprising the steps of:

[0026] S1. Establish a transaction pool and obtain the computation data and its computation formulas dispatched by the user to the distribution server;

[0027] S2. The computational data and computational formulas are decomposed hierarchically according to the order of operations to obtain hierarchical data and hierarchical formulas;

[0028] S3. Allocate each level of data and its corresponding level formula to different computers and perform calculations sequentially according to the order of operations;

[0029] S4. After the calculation is completed, the transaction callback mechanism is invoked to output the calculation result and return it to the corresponding user.

[0030] As can be seen from the above description, the beneficial effects of the present invention are as follows: It provides a hierarchical data calculation method, which decomposes the calculation data and calculation formulas dispatched by users to the distribution server in a hierarchical manner according to the calculation order by establishing a transaction pool. That is, it decomposes the entire complex calculation process into multiple calculation steps, and matches different computers to each level of calculation steps for sequential calculation. In other words, it distributes the complex calculation process to different computers for hierarchical processing, which helps to improve data calculation efficiency, reduce the CPU load of each computer, and improve the utilization rate of idle computer computing power. At the same time, the hierarchical calculation can push the formulas and data required by the data hierarchically, with each level responsible for one piece of data, which also ensures data privacy and solves the data isolation problem.

[0031] Further, step S2 specifically includes:

[0032] S21. The operational formula is decomposed hierarchically according to the order of operations to obtain multiple hierarchical formulas, and stored as an array S{S1,S2,S3,…,S…} n};

[0033] S22. The computational data is also decomposed into mechanical energy levels according to the computational order to obtain multiple levels of data, and stored in array form D{D1,D2,D3,…,D…} n};

[0034] S23. Combine the hierarchical formulas and hierarchical data one by one according to the order of operations, and store them as a combined array F{F1,F2,F3,...,F...} n},in:

[0035] F1 = S1(D1);

[0036] F2 = S2(F1D2);

[0037] F3 = S3(F2D3);

[0038] ...

[0039] F n =S n (F n-1 D n );

[0040] S24, marking the complexity level of each element in the combination array in the order from low to high according to the operation sequence, denoted as {NP1, NP2, NP3, …, NPn}.

[0041] From the above description, the operation data and the operation formula needed are hierarchically disassembled according to the operation sequence, and are nested in the form of an array, which is fast and meets the subsequent sequential operation requirements.

[0042] Further, the step S3 specifically comprises:

[0043] S31, obtaining the running state of each computer in the current network, including the idle condition and the floating point operation per second FLOPS;

[0044] S32, screening to obtain idle computers, and sequentially matching the computers with corresponding FLOPS for each element in the combination array according to the data amount of the hierarchical data corresponding to each element in the combination array and the complexity of the hierarchical formula in the order from low to high according to the complexity level;

[0045] S33, performing step-by-step operation on the corresponding elements in the matched computers according to the complexity level from low to high, and outputting the final calculation result in the computer matched at the highest complexity level.

[0046] From the above description, each data after layering and the operation formula needed are matched with idle computers with corresponding computing power according to the operation complexity, and are sequentially matched according to the operation sequence, ensuring that the computers matched at each layer meet the sufficient utilization of idle computing power utilization rate; meanwhile, by matching the floating point operation per second FLOPS of the computers, it is ensured that the matched computers will not be overloaded in the subsequent data operation process.

[0047] Further, the step S32 further comprises:

[0048] The matching of the computers with corresponding FLOPS for each element in the combination array according to the data amount of the hierarchical data corresponding to each element in the combination array and the complexity of the hierarchical formula specifically comprises:

[0049] According to the data amount of the hierarchical data corresponding to each element in the combination array, a computer with corresponding FLOPS is matched, if multiple computers are matched, then according to the complexity of the hierarchical formula, a computer with corresponding FLOPS is matched again from the multiple computers matched, as the computer for hierarchical operation of each element in the combination array.

[0050] As can be seen from the above description, the computer with corresponding computing power is matched by the calculation amount of the hierarchical data first, and then the computer with corresponding computing power is further screened by the complexity of the hierarchical formula used, which can effectively guarantee the optimal throughput of data calculation.

[0051] Further, the FLOPS of each computer in the current network per second is obtained in the step S31, specifically:

[0052] The clock period and clock frequency of each computer are obtained in advance, the number of floating point operations executable by each computer per clock period is calculated, and the clock frequency of the corresponding computer is multiplied to obtain the FLOPS of each computer per second.

[0053] As can be seen from the above description, the FLOPS per second is used to measure the speed of floating point operation of CPU, which is a key indicator in modern CPU design. The FLOPS is calculated by obtaining the clock period and clock frequency of the computer, so as to ensure that the matched computer will not be overloaded in the subsequent data operation process.

[0054] Please refer to Figure 2 A hierarchical operation terminal of data, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program:

[0055] S1, establishing a transaction pool and obtaining operation data and operation formula of a user distributed to a distribution server;

[0056] S2, the operation data and operation formula are hierarchically decomposed according to the operation sequence to obtain hierarchical data and hierarchical formula;

[0057] S3, each hierarchical data and the corresponding hierarchical formula are distributed to different computers and calculated in sequence according to the operation sequence;

[0058] S4, after the calculation is completed, the transaction callback mechanism is called to output the calculation result and return it to the corresponding user.

[0059] As described above, the beneficial effects of this invention are as follows: Based on the same technical concept, and in conjunction with the aforementioned data hierarchical calculation method, a data hierarchical calculation terminal is provided. By establishing a transaction pool, the calculation data and calculation formulas dispatched by users to the distribution server are hierarchically decomposed according to the calculation order. That is, the entire complex calculation process is decomposed into multiple calculation steps, and different computers are matched to each level of calculation steps for sequential calculation. In other words, the complex calculation process is distributed to different computers for hierarchical processing, which helps to improve data calculation efficiency, reduce the CPU load of each computer, and improve the utilization rate of idle computer computing power. At the same time, hierarchical calculation can push the formulas and data required by the data hierarchically, with each level responsible for one piece of data, which also ensures data privacy and solves the data isolation problem.

[0060] Further, step S2 specifically includes:

[0061] S21. The operational formula is decomposed hierarchically according to the order of operations to obtain multiple hierarchical formulas, and stored as an array S{S1,S2,S3,…,S…} n};

[0062] S22. The computational data is also decomposed into mechanical energy levels according to the computational order to obtain multiple levels of data, and stored in array form D{D1,D2,D3,…,D…} n};

[0063] S23. Combine the hierarchical formulas and hierarchical data one by one according to the order of operations, and store them as a combined array F{F1,F2,F3,...,F...} n},in:

[0064] F1 = S1(D1);

[0065] F2 = S2(F1D2);

[0066] F3 = S3(F2D3);

[0067] ...

[0068] F n =S n (F n-1 D n );

[0069] S24. Mark the complexity level of each element in the combined array from low to high according to the order of operations, denoted as {NP1,NP2,NP3,…,NPn}.

[0070] From the above description, the operation data and the operation formula required are hierarchically disassembled according to the operation sequence and are nested in the form of an array, which is fast and meets the subsequent sequential operation requirements.

[0071] Further, the step S3 is specifically:

[0072] S31, obtaining the running state of each computer in the current network, including the idle condition and the floating point operation per second FLOPS;

[0073] S32, screening to obtain idle computers, and sequentially matching the computer with corresponding FLOPS for each element in the combined array according to the data amount of the hierarchical data corresponding to each element in the combined array and the complexity of the hierarchical formula in the order from low to high according to the complexity level.

[0074] S33, sequentially performing step-by-step operation on the corresponding elements in the matched computers according to the complexity level from low to high, and outputting the final calculation result in the computer matched at the highest complexity level.

[0075] From the above description, for each data after layering and the operation formula required, the idle computer with corresponding computing power is matched according to the operation complexity, and is sequentially matched according to the operation sequence, which ensures that the computer matched at each level of operation meets the sufficient utilization of idle computing power utilization rate; meanwhile, by matching the floating point operation per second FLOPS of the computer, it is ensured that the matched computer will not be overloaded in the subsequent data operation process.

[0076] Further, the step S32 further includes:

[0077] The matching of the computer with corresponding FLOPS for each element in the combined array according to the data amount of the hierarchical data corresponding to each element in the combined array and the complexity of the hierarchical formula is specifically:

[0078] First, the computer with corresponding FLOPS is matched according to the data amount of the hierarchical data corresponding to each element in the combined array, and if multiple computers are matched, then the computer with corresponding FLOPS is further matched from the multiple matched computers according to the complexity of the hierarchical formula, as the computer for hierarchical operation of each element in the combined array.

[0079] From the above description, the computer with corresponding computing power is preferentially matched through the calculation amount of hierarchical data, and then the computer with corresponding computing power is further screened through the complexity of the hierarchical formula used, which can effectively guarantee the optimal throughput of data calculation.

[0080] Further, the step S31 obtains the floating point operation per second (FLOPS) of each computer in the current network, specifically:

[0081] The clock period and clock frequency of each computer are obtained in advance, the number of floating point operations that each computer can perform per clock period is calculated, and the clock frequency of the corresponding computer is multiplied to obtain the floating point operation per second (FLOPS) of each computer.

[0082] The application provides a data ladder operation method and a terminal, which are mainly applied to a scene of complex operation on a large amount of user data in a distributed network.

[0083] Please refer to Figure 1 The embodiment one of the application is:

[0084] A data ladder operation method, as shown in the figure, comprises the following steps: Figure 1

[0085] S1, a transaction pool is established, and operation data and operation formula of a user distributed to a distribution server are obtained.

[0086] S2, the operation data and the operation formula are hierarchically decomposed according to an operation sequence to obtain hierarchical data and hierarchical formula.

[0087] S3, each hierarchical data and the corresponding hierarchical formula are distributed to different computers and sequentially calculated according to the operation sequence.

[0088] S4, after the calculation is completed, a transaction callback mechanism is called to output the calculation result and return it to the corresponding user.

[0089] The transaction callback in the embodiment refers to creating a transaction callback for data and distributing the data to the server, and the user waits for the data processing to be completed, and other processes are processed under the hierarchical operation of the transaction pool. The transaction callback satisfies the formula sequence execution, that is, the final result is output after the data of the user is completely operated, and then the result is returned to the user. That is, the user only submits the data, and the data processing process is processed by the transaction pool and the transaction callback. The application further ensures the data privacy security and solves the data isolation problem.

[0090] ​Namely, in the embodiment, the operation data and operation formula of the user distributed to the distribution server are hierarchically decomposed according to the operation sequence by establishing a transaction pool, that is, the whole complex operation process is decomposed into multiple operation steps, and different computers are matched for each hierarchical operation step to perform sequential calculation, that is, the complex operation process is dispersed to different computers for hierarchical processing, which is beneficial to improve data calculation efficiency, reduce CPU load of each computer, and improve utilization of idle computer computing power; meanwhile, hierarchical operation grading can push the formula and data required by the data in a hierarchical manner, and each level is responsible for one data, which ensures data privacy and solves the data isolation problem.

[0091] Embodiment two of the application is:

[0092] A ladder operation method of data, based on the above embodiment one, in the embodiment, the step S2 is specifically:

[0093] S21, the operation formula is hierarchically decomposed according to the operation sequence, a plurality of hierarchical formulas are obtained, and stored in an array form S{S1, S2, S3, …, S n}。

[0094] S22, the operation data is also mechanically hierarchically decomposed according to the operation sequence, a plurality of hierarchical data are obtained, and stored in an array form D{D1, D2, D3, …, D n}。

[0095] S23, the hierarchical formula and the hierarchical data are combined according to the operation sequence one by one, and stored in a combined array form F{F1, F2, F3, …, F n},wherein:

[0096] F1=S1(D1);

[0097] F2=S2(F1D2);

[0098] F3=S3(F2D3);

[0099] …

[0100] F n =S n (F n-1 D n )。

[0101] S24, each element in the combined array is marked with a complexity level according to the operation sequence from low to high, and is recorded as{NP1, NP2, NP3, …, NP n}。

[0102] Namely, according to the operation sequence, the operation data and the operation formula required are hierarchically disassembled and nested in the form of an array, which is fast and meets the subsequent sequential operation requirements.

[0103] Meanwhile, in the embodiment, the step S3 is specifically:

[0104] S31, obtain the running state of each computer in the current network, including the idle condition and the floating point operation per second FLOPS, wherein the floating point operation per second FLOPS is obtained as follows:

[0105] The clock period and the clock frequency of each computer are obtained in advance, the number of floating point operations that can be executed by each computer per clock period is calculated, and the clock frequency of the corresponding computer is multiplied to obtain the floating point operation per second FLOPS of each computer.

[0106] Namely, the floating point operation per second FLOPS is used to measure the speed of CPU executing floating point operation, which is a key indicator in modern CPU design:

[0107] One MFLOPS (megaFLOPS) is equal to one million (= 10^6) floating point operations per second;

[0108] One GFLOPS (gigaFLOPS) is equal to one billion (= 10^9) floating point operations per second;

[0109] One TFLOPS (teraFLOPS) is equal to one trillion (= 10^12) floating point operations per second;

[0110] One PFLOPS (petaFLOPS) is equal to one quadrillion (= 10^15) floating point operations per second.

[0111] Meanwhile, FLOPS can also be replaced by OPS (Tera Operations Per Second, the number of operations that the processor can perform per second):

[0112] One MOPS (Million Operation Per Second) represents one million (10^6) operations that the processor can perform per second;

[0113] One GOPS (Giga Operations Per Second) represents one billion (10^9) operations that the processor can perform per second;

[0114] One TOPS (Tera Operations Per Second) represents one trillion (10^12) operations that the processor can perform per second.

[0115] By acquiring the clock period and clock frequency of the computer, the FLOPS is calculated, and it is ensured that the matched computer will not be overloaded in the subsequent data operation process.

[0116] S32, the idle computer is screened, and according to the data amount of the hierarchical data corresponding to each element in the combination array and the complexity of the hierarchical formula, the computer with corresponding FLOPS is matched for each element in the combination array in order from low to high according to the complexity level.

[0117] S33, according to the complexity level from low to high, the corresponding element is operated in the matched computer in order, and the final calculation result is output in the computer matched at the highest complexity level.

[0118] That is, for each data after layering and the operation formula needed, the computer with corresponding computing power is matched according to the operation complexity, and the computer is matched in order according to the operation sequence, so as to ensure that the computer matched in each layer operation meets the sufficient utilization of idle computing power utilization rate; meanwhile, by matching the floating point operation number FLOPS of the computer per second, it is ensured that the matched computer will not be overloaded in the subsequent data operation process.

[0119] In this embodiment, the step S32 further includes:

[0120] According to the data amount of the hierarchical data corresponding to each element in the combination array and the complexity of the hierarchical formula, the computer with corresponding FLOPS is matched for each element in the combination array, specifically.

[0121] First, the computer with corresponding FLOPS is matched according to the data amount of the hierarchical data corresponding to each element in the combination array, if multiple computers are matched, then the computer with corresponding FLOPS is matched again from the multiple matched computers according to the complexity of the hierarchical formula, as the computer for hierarchical operation of each element in the combination array.

[0122] That is, the computer with corresponding computing power is matched by the calculation amount of the hierarchical data first, and then the computer with corresponding computing power is further screened by the complexity of the hierarchical formula used, because the data amount is large but the formula is simple, so the computer with corresponding computing power is matched by the calculation amount of the data first, and when the data amount is large and the formula is also complex, since some formulas are designed to calculate data, variance, analogy, etc., multiple calculations consume computing power, therefore, on the basis of considering the data amount, it is further known how many hierarchical formulas are nested in the data, that is, the complexity of the formula, to further screen the best computer, which can effectively guarantee the optimal throughput of data calculation.

[0123] Referring to Figure 2 Embodiment three of the present application is:

[0124] A terminal 1 for ladder operation of data, comprising a memory 2, a processor 3, and a computer program stored in the memory 2 and executable on the processor 3, wherein the processor 3 completes the steps in the method for ladder operation of data in one of Embodiment one or Embodiment two when executing the computer program.

[0125] In summary, the present application provides a method and terminal for ladder operation of data, wherein the operation data and operation formula of a user assigned to a distribution server are hierarchically decomposed according to the operation sequence by establishing a transaction pool, i.e., the entire complex operation process is decomposed into multiple hierarchical operation steps, and the computers with corresponding computing power are matched for each hierarchical operation step according to the running state of each computer in the current network for sequential calculation, i.e., the complex operation process is dispersed to different computers for hierarchical processing, which is conducive to improving data calculation efficiency, reducing CPU load of each computer, and improving utilization of idle computer computing power; at the same time, hierarchical operation grading can hierarchically push the formula and data required by the data, and each level is responsible for one data, which ensures data privacy and solves the data isolation problem.

[0126] The above is only an embodiment of the present application, and does not limit the patent scope of the present application, and any equivalent transformation or direct or indirect application in related technical fields using the content of the present application specification and drawings is also included in the patent protection scope of the present application.

Claims

1. A method for stepwise data calculation, characterized in that, Including the following steps: S1. Establish a transaction pool and obtain the computation data and its computation formulas dispatched by the user to the distribution server; S2. The computational data and computational formulas are decomposed hierarchically according to the order of operations to obtain hierarchical data and hierarchical formulas; S3. Allocate each level of data and its corresponding level formula to different computers and perform calculations sequentially according to the order of operations; S4. After the calculation is completed, the transaction callback mechanism is invoked to output the calculation result and return it to the corresponding user; Step S2 specifically involves: S21. The operational formula is decomposed hierarchically according to the order of operations to obtain multiple hierarchical formulas, and stored as an array S{S1,S2,S3,…,S…} n }; S22. The computational data is also decomposed hierarchically according to the computational order to obtain multiple levels of data, and stored in array form D{D1,D2,D3,…,D…} n }; S23. Combine the hierarchical formulas and hierarchical data one by one according to the order of operations, and store them as a combined array F{F1,F2,F3,...,F...} n },in: F1=S1(D1); F2 = S2(F1D2); F3 = S3(F2D3); …… F n =S n (F n-1 D n ); S24. Label the complexity level of each element in the combined array from low to high according to the order of operations, denoted as {NP1,NP2,NP3,…,NPn}; Step S3 specifically involves: S31. Obtain the operating status of each computer in the current network, including idle status and floating-point operations per second (FLOPS); S32. Select available computers and match them with computers of corresponding FLOPS for each element in the combination array according to the complexity level from low to high, based on the amount of data of the level data corresponding to each element in the combination array and the complexity of the level formula. S33. Perform stepwise operations on the corresponding elements in the matched computers in order of increasing complexity level, and output the final calculation result in the computer matched with the highest complexity level.

2. The stepwise data calculation method according to claim 1, characterized in that, Step S32 also includes: The step of matching a computer with corresponding FLOPS for each element of the combined array based on the data volume of the hierarchical data corresponding to each element and the complexity of the hierarchical formula is as follows: First, the corresponding FLOPS computer is matched according to the amount of data of the hierarchical data corresponding to each element in the combined array. If multiple computers are matched, then the corresponding FLOPS computer is matched again from the multiple matched computers according to the complexity of the hierarchical formula, and used as the computer for hierarchical operation of each element in the combined array.

3. The stepwise data calculation method according to claim 1, characterized in that, In step S31, obtaining the number of floating-point operations per second (FLOPS) for each computer in the current network specifically involves: The clock cycle and clock frequency of each computer are obtained in advance. The number of floating-point operations that each computer can perform per clock cycle is calculated and multiplied by the clock frequency of the corresponding computer to obtain the number of floating-point operations per second (FLOPS) for each computer.

4. A data tiered processing terminal, characterized in that, Includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, performs the following steps: S1. Establish a transaction pool and obtain the computation data and its computation formulas dispatched by the user to the distribution server; S2. The computational data and computational formulas are decomposed hierarchically according to the order of operations to obtain hierarchical data and hierarchical formulas; S3. Allocate each level of data and its corresponding level formula to different computers and perform calculations sequentially according to the order of operations; S4. After the calculation is completed, the transaction callback mechanism is invoked to output the calculation result and return it to the corresponding user; Step S2 specifically involves: S21. The operational formula is decomposed hierarchically according to the order of operations to obtain multiple hierarchical formulas, and stored as an array S{S1,S2,S3,…,S…} n }; S22. The computational data is also decomposed hierarchically according to the computational order to obtain multiple levels of data, and stored in array form D{D1,D2,D3,…,D…} n }; S23. Combine the hierarchical formulas and hierarchical data one by one according to the order of operations, and store them as a combined array F{F1,F2,F3,...,F...} n },in: F1=S1(D1); F2 = S2(F1D2); F3 = S3(F2D3); …… F n =S n (F n-1 D n ); S24. Label the complexity level of each element in the combined array from low to high according to the order of operations, denoted as {NP1,NP2,NP3,…,NPn}; Step S3 specifically involves: S31. Obtain the operating status of each computer in the current network, including idle status and floating-point operations per second (FLOPS); S32. Select available computers and match them with computers of corresponding FLOPS for each element in the combination array according to the complexity level from low to high, based on the amount of data of the level data corresponding to each element in the combination array and the complexity of the level formula. S33. Perform stepwise operations on the corresponding elements in the matched computers in order of increasing complexity level, and output the final calculation result in the computer matched with the highest complexity level.

5. A data tiered processing terminal according to claim 4, characterized in that, Step S32 also includes: The step of matching a computer with corresponding FLOPS for each element of the combined array based on the data volume of the hierarchical data corresponding to each element and the complexity of the hierarchical formula is as follows: First, the corresponding FLOPS computer is matched according to the amount of data of the hierarchical data corresponding to each element in the combined array. If multiple computers are matched, then the corresponding FLOPS computer is matched again from the multiple matched computers according to the complexity of the hierarchical formula, and used as the computer for hierarchical operation of each element in the combined array.

6. A data tiered processing terminal according to claim 4, characterized in that, In step S31, obtaining the number of floating-point operations per second (FLOPS) for each computer in the current network specifically involves: The clock cycle and clock frequency of each computer are obtained in advance. The number of floating-point operations that each computer can perform per clock cycle is calculated and multiplied by the clock frequency of the corresponding computer to obtain the number of floating-point operations per second (FLOPS) for each computer.

Images