CN104572134A - Optimization method and optimization device - Google Patents
Optimization method and optimization device Download PDFInfo
- Publication number
- CN104572134A CN104572134A CN201510068301.2A CN201510068301A CN104572134A CN 104572134 A CN104572134 A CN 104572134A CN 201510068301 A CN201510068301 A CN 201510068301A CN 104572134 A CN104572134 A CN 104572134A
- Authority
- CN
- China
- Prior art keywords
- program
- time value
- value
- time
- successively
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Landscapes
- Debugging And Monitoring (AREA)
- Telephonic Communication Services (AREA)
Abstract
The invention provides an optimization method and an optimization device. The method comprises the following steps: the frequency that a program i in programs recorded in an application-level log subjected to first class of calls is taken as basis, the frequency of the program i subjected to first class of calls is reduced to a preset numerical value according to the basis, and until the first time value and the second time value meet the preset conditions, optimization of the service operation process is finished, wherein the first time value refers to the original operation time of the service operation process, and the second time value refers to the operation time of the service operation process after the frequency of the first class of programs subjected to first class of calls is sequentially reduced to the preset numerical value; and therefore, the programs in the application-level log are taken as optimization objects, and the time values are taken as optimization targets. Therefore, compared with the conventional optimization method, the method disclosed by the invention has the advantages that the optimization objects are clear, and the targets needing to be optimized are also clear, so that the reference of human experience is not needed, and the method has high optimization efficiency.
Description
Technical field
The application relates to computer realm, particularly relates to a kind of optimization method and device.
Background technology
In the programming stage, usually need to be optimized program, such as, reduce the working time etc. of program.Traditional optimization method, usually by specific purpose tool, obtains the operational factor of program to be optimized, is optimized according to these parameters.
And the operational factor of the data abstraction layer of the normally bottom that existing instrument obtains, these parameters primarily of which upper procedure are called, and cannot learn, therefore, cannot navigate to should emphasis optimize specific procedure, therefore, at present, the main artificial experience that relies on locates concrete optimization object and optimization aim.
Therefore, there is inefficient problem in existing optimization method.
Summary of the invention
This application provides a kind of optimization method and device, object is to solve the low problem of optimization efficiency.
To achieve these goals, this application provides following technical scheme:
A kind of optimization method, comprising:
Program i in statistics program set is performed the number of times that the first kind is called, the described first kind is called as identical the calling of input parameter, and described collection of programs is the program recorded in the application layer daily record of service operation, wherein, i=1,2 ... N, described N be greater than 0 integer;
Successively described program i is performed the number of times that the first kind calls and is reduced to default value, until when very first time value and the second time value meet pre-conditioned, determine the optimization to described service operation process, the raw runtime that described very first time value is described service operation process, described second time value is successively described program i is performed the working time that number of times that the first kind calls is reduced to the described service operation process after default value.
Alternatively, describedly successively described program i is performed number of times that the first kind calls and is reduced to second value and comprises:
According to being performed number of times that the first kind calls from big to small, described program i is sorted;
From first type 1 programming in sequence, successively described type 1 programming is performed the number of times that the first kind calls and is reduced to second value.
Alternatively, describedly pre-conditionedly to comprise:
Second time value <=very first time value * (1-P), wherein, P is default optimized proportion target, and its span is (0,1).
Alternatively, if be performed by described program i after number of times that the first kind calls is reduced to second value successively described, described very first time value and the second time value can not meet pre-conditioned, also comprise:
Upgrade the invoked number of times of level of abstraction program in described collection of programs;
Upgrade the level of abstraction program invoked wastage in bulk or weight time that invoked number of times is greater than third value;
According to wastage in bulk or weight time order from big to small, described level of abstraction program is sorted;
Be once many inquiries by the repeatedly single query conversion of the level of abstraction program in described sequence under different parameters successively, until described very first time value is compared with the 3rd time value, the ratio of saving reaches default first and can to save time ratio, and described 3rd time value is be working time of described service operation process after once many inquiries successively by the repeatedly single query conversion of the program in described sequence under different parameters.
Alternatively, described by the repeatedly single query conversion of the program in described sequence under different parameters be successively once many inquiry after, if described very first time value is compared with the 3rd time value, the ratio of saving does not reach default first and can to save time ratio, also comprises:
Successively the row table called in described service operation process is converted to list, until described very first time value is compared with the 4th time value, the ratio of saving reaches default second and can to save time ratio.
A kind of optimization device, comprising:
Statistical module, be performed for the program i in statistics program set the number of times that the first kind calls, the described first kind is called as identical the calling of input parameter, described collection of programs is the program recorded in the application layer daily record of service operation, wherein, i=1,2 ... N, described N be greater than 0 integer;
First optimizes execution module, the number of times called for successively described program i being performed the first kind is reduced to default value, until when very first time value and the second time value meet pre-conditioned, determine the optimization to described service operation process, the raw runtime that described very first time value is described service operation process, described second time value is successively described program i is performed the working time that number of times that the first kind calls is reduced to the described service operation process after default value.
Alternatively, described first optimizes execution module and is used for being performed by described program i successively number of times that the first kind calls and is reduced to second value and comprises:
Described first optimize execution module specifically for, according to being performed number of times that the first kind calls from big to small, described program i is sorted; From first type 1 programming in sequence, successively described type 1 programming is performed the number of times that the first kind calls and is reduced to second value.
Alternatively, described first optimizes execution module is used for successively described program i being performed the number of times that the first kind calls and is reduced to default value, until when very first time value and the second time value meet pre-conditioned, determine to comprise the optimization of described service operation process:
Described first optimize execution module specifically for, successively described program i is performed the number of times that the first kind calls and is reduced to default value, until when very first time value and the second time value meet pre-conditioned, determine the optimization to described service operation process, describedly pre-conditionedly to comprise: the second time value <=very first time value * (1-P), wherein, P is default optimized proportion target, its span is (0,1).
Alternatively, also comprise:
Second optimizes execution module, if for successively described program i being performed after number of times that the first kind calls is reduced to second value described, described very first time value and the second time value can not meet pre-conditioned, upgrade the invoked number of times of level of abstraction program in described collection of programs; Upgrade the level of abstraction program invoked wastage in bulk or weight time that invoked number of times is greater than third value; According to wastage in bulk or weight time order from big to small, described level of abstraction program is sorted; Be once many inquiries by the repeatedly single query conversion of the level of abstraction program in described sequence under different parameters successively, until described very first time value is compared with the 3rd time value, the ratio of saving reaches default first and can to save time ratio, and described 3rd time value is be working time of described service operation process after once many inquiries successively by the repeatedly single query conversion of the program in described sequence under different parameters.
Alternatively, also comprise:
3rd optimizes execution module, for described by the repeatedly single query conversion of the program in described sequence under different parameters be successively once many inquiry after, if described very first time value is compared with the 3rd time value, the ratio of saving does not reach default first and can to save time ratio, successively the row table called in described service operation process is converted to list, until described very first time value is compared with the 4th time value, the ratio of saving reaches default second and can to save time ratio.
Method described in the application and device, number of times that the first kind calls is performed for foundation with the program i in the program recorded in application layer daily record, default value is reduced to according to program i being performed the number of times that the first kind calls, until when very first time value and the second time value meet pre-conditioned, determine the optimization to described service operation process, wherein, very first time value is the raw runtime of described service operation process, second time value is successively described type 1 programming is performed the working time that number of times that the first kind calls is reduced to the described service operation process after default value, visible, with the program in application layer daily record for optimization object, take time value as optimization aim, therefore, compared with existing optimization method, optimization object is clear and definite, simultaneously, optimize and need the target reached also to be clear and definite, so, without the need to reference artificial experience, there is higher optimization efficiency.
Accompanying drawing explanation
In order to be illustrated more clearly in the embodiment of the present application or technical scheme of the prior art, be briefly described to the accompanying drawing used required in embodiment or description of the prior art below, apparently, accompanying drawing in the following describes is only some embodiments of the application, for those of ordinary skill in the art, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to these accompanying drawings.
Fig. 1 is a kind of optimization method process flow diagram disclosed in the embodiment of the present application;
Fig. 2 is the process flow diagram of the embodiment of the present application another optimization method disclosed;
Fig. 3 is that program is called the schematic diagram of time consumed;
Fig. 4 is that program is called the schematic diagram of time consumed;
Fig. 5 is the process flow diagram of the embodiment of the present application another optimization method disclosed;
Fig. 6 is the schematic diagram of the relation of single inquiry and many inquiries;
The structural representation of Fig. 7 a kind of optimization device disclosed in the embodiment of the present application.
Embodiment
Optimization method described in the embodiment of the present application and device, can be applied in the system using Rule language to build, such as blue extra large engine (Blue Ocean Engineering, BoEing) system.
Rule program is by the program of forth generation programming language AppBuilder Rule language compilation, Rule language provides basic variable declarations, variable assignments, establishment data structure etc. to support the function of programming as other most of programming language, Rule program can be compiled into multiple programming language as required, as COBOL, C#, JAVA etc., in BoEing system, background program is compiled into COBOL code, and foreground program is compiled into C# code.
Bank transaction in BoEing system is made up of a series of Rule program, due to the process that each Rule program similarly is more in C language, if a C programmer is on average made up of 10 processes, a bank transaction is completed by 15 c programs, approximately will need 150 programs when so realizing by Rule program.
The exploitation of BoEing system does not adopt traditional vertical shaft type development mode in addition, but whole system is divided into multiple module, each module sets up corresponding development group, each development group has cooperated bank transaction, this division of labor become more meticulous makes the hierarchical structure of system more clear, the function of each module can be done more powerful, the focus of developer is more concentrated, but the function that the next program of vertical shaft type development mode completes may need to split in multiple module now, is jointly completed by multiple program.
Directly results in a bank transaction in BoEing system by the feature of above-mentioned two aspects to need to have been come by a large amount of Rule programs, as the payment beforehand transaction of providing a loan in application is made up of jointly 1230 event-driven layer Rule and 714 data abstraction layer Rule, sharply increasing of program quantity brings larger difficulty to performance evaluation.After performance test is reported out, the time loss that bottom database program is total and the average execution speed of single-length fixed point instruction (Million instructions Per Second can only be seen, MIPS) consume, clear and definite optimization object cannot be obtained according to this result.
The object of method described in the embodiment of the present application and device is, automatic clear and definite optimization object, saves the process of artificial dependence experience determination optimization object, improves the efficiency optimized.
Below in conjunction with the accompanying drawing in the embodiment of the present application, be clearly and completely described the technical scheme in the embodiment of the present application, obviously, described embodiment is only some embodiments of the present application, instead of whole embodiments.Based on the embodiment in the application, those of ordinary skill in the art are not making the every other embodiment obtained under creative work prerequisite, all belong to the scope of the application's protection.
A kind of optimization method disclosed in the embodiment of the present application, as shown in Figure 1, comprising:
S101: the program i in statistics program set is performed the number of times that the first kind is called, wherein, i=1,2 ... N, described N be greater than 0 integer;
In the present embodiment, the first kind is called as identical the calling of input parameter, such as, suppose when Rule_A calls Rule_B, the concrete input data of Rule_B are A, and when Rule_C calls Rule_B, the concrete input data of Rule_B are also A, then Rule_B has been performed the first kind and has called, and being performed the number of times that the first kind calls is 2.
If when Rule_A calls Rule_C, the concrete input data of Rule_C are A (or B), then can not determine whether Rule_C has been performed the first kind and has called, as long as Rule_B call Rule_C and concrete input data are A (or B) time, could determine that Rule_C is performed the number of times that the first kind calls is 2.
In the present embodiment, collection of programs is the program recorded in the application layer daily record of service operation.
S102: successively program i is performed the number of times that the first kind calls and is reduced to default value, until when very first time value and the second time value meet pre-conditioned, determined the optimization to described service operation process.
Wherein, very first time value is the raw runtime of service operation process, and described second time value is successively type 1 programming is performed the working time that number of times that the first kind calls is reduced to the service operation process after default value.
In the present embodiment, in program i, both can comprise event layers program and level of abstraction program.
Optimization method described in the present embodiment, the program comprised with application layer daily record is optimization object, under identical input parameter, being performed the number of times that calls for foundation with program, take time value as optimization aim, therefore, compared with the existing optimization method to Rule program, with clearly defined objective, object is clear and definite, and means are clear and definite, therefore, without the need to the access of artificial experience, thus there is higher efficiency.
The embodiment of the present application another optimization method disclosed, as shown in Figure 2, comprises following concrete steps:
S201: the program i in statistics program set is performed the number of times that the first kind is called, wherein, i=1,2 ... N, described N be greater than 0 integer;
S202: according to being performed number of times that the first kind calls from big to small, described program i is sorted;
S203: from first type 1 programming in sequence, successively described type 1 programming is performed the number of times that the first kind calls and is reduced to second value, until the second time value <=very first time value * (1-P), wherein, P is default optimized proportion target, its span is (0,1).Such as, P can be 0.3, and before optimizing, elapsed time is 0.2 second, then optimization aim is 0.14 second.
In the present embodiment, very first time value is the raw runtime of service operation process, can be obtained by operation business.
Second time value is successively program i is performed the working time that number of times that the first kind calls is reduced to the service operation process after default value, can obtain in the following manner: from system-level daily record, obtain a Rule program called time value consumed each time, determine the time value calling consumption be subtracted further, use very first time value to deduct the total time value calling consumption be subtracted, be the second time value.
Such as shown in Fig. 3, calling the time that Rule_B consumes is approximately t2-t1, if deduct after a Rule_B calls, then the second time value is that very first time value deducts t2-t1.
It should be noted that, Rule_A may carry out some variable assignments and calculating operation after having called Rule_B, then just performs Rule_C, and therefore the value of t2-t1 is slightly larger than the actual execution time of Rule_B, and actual test error is within 1%.
Usually, classical stack algorithm can be adopted to obtain each program and to be called the time value consumed, such as Fig. 4 again, the allocating time of Rule_C is t3-t2, the allocating time of the allocating time of Rule_E to be the allocating time of t5-t4, Rule_D be t5-t3, Rule_B is t5-t2.
Optimization method described in the present embodiment take time value as optimization aim, meets pre-conditioned once very first time value and the second time value, determine that optimizing process completes, therefore, it is possible to when ensureing optimization aim, the transformation for the program in system is minimum, and, optimized proportion target can be arranged according to actual needs, can avoid the phenomenon excessively optimized, and reduces the change amount optimized as far as possible, not only can save human cost, and can risk be reduced.
The embodiment of the present application another optimization method disclosed, as shown in Figure 5, comprises following concrete steps:
S501: each program recorded in statistics application level logs is performed the number of times that the first kind is called;
In Rule program, when Rule_A calls Rule_B, form is Rule_A->Rule_B: input data.
S502: according to being performed number of times that the first kind calls from big to small, each program is sorted;
S503：n＝1；
S504: judge whether n<=M sets up, wherein M is total number of program, if so, performs S505, if not, performs S508;
S505: program n is performed the number of times that the first kind calls and is reduced to 1;
S506: judge whether the second time value <=very first time value * (1-P) sets up, and if so, completes optimizing process, if not, performs S507;
S507:n=n+1, returns and performs S504;
S508: the invoked number of times of level of abstraction program in refresh routine set;
S509: upgrade the level of abstraction program invoked wastage in bulk or weight time that invoked number of times is greater than third value;
S510: according to wastage in bulk or weight time order from big to small, sorts to described level of abstraction program;
S511: be once many inquiries by the repeatedly single query conversion of level of abstraction program under different parameters in sequence successively, until described very first time value is compared with the 3rd time value, the ratio of saving reaches default first and can to save time ratio;
Wherein, the 3rd time value is be working time of described service operation process after once many inquiries successively by the repeatedly single query conversion of the program in described sequence under different parameters.
It should be noted that, after optimizing, the execution number of times of some event layers Rule becomes 1, the number of times of the level of abstraction Rule that it calls needs corresponding minimizing, such as, event layers Rule_A have invoked level of abstraction Rule_I and Rule_J, Rule_A own has been repeated to call (each called time its input just the same) 3 times in concluding the business, then removed by Rule_A after repeating for twice to call, the call number of Rule_I and Rule_J in concluding the business will reduce 2 times respectively.
The relation of single inquiry and many inquiries as shown in Figure 6, level of abstraction Rule_I has been called twice, each input parameter is different, be then twice single inquiry, use more than one to inquire about by Rule_A after optimization disposablely to inquire 2 and record and be put in global buffer, Rule_B and Rule_C no longer calls level of abstraction Rule_I in directly obtaining from buffer memory.Can find out, be once many inquiries by the repeatedly single query conversion of level of abstraction program under different parameters, can save time.
In the present embodiment, the first ratio of can saving time is repeatedly that single query conversion becomes more than 1 time to inquire about savable time scale, can be designated as Q, priority is 1, and span is (0,1), this value is empirical value, can be provided by user (such as program test personnel).Such as, inquire about 3 different customer informations from customer information table, 3 times singles inquiry elapsed time is 0.004 second, and the elapsed time of more than 1 time inquiry is 0.003 second, Q=0.25 now.
S512: after being once many inquiries successively by the repeatedly single query conversion of the program in described sequence under different parameters, if described very first time value is compared with the 3rd time value, the ratio of saving does not reach default first and can to save time ratio, successively the row table called in described service operation process is converted to list, until described very first time value is compared with the 4th time value, the ratio of saving reaches default second and can to save time ratio.
Second ratio of can saving time can be saved time ratio for data query when row table being converted to list, R can be designated as, priority is that 2 (value is larger, priority is lower), span is (0,1), this value is empirical value, can be provided by user (such as program test personnel).Such as, row list structure type is as table 1:
Table 1
| Account | Type | Value |
| A | 0001 | 10 |
| A | 0002 | 15 |
| A | ... | ... |
After converting list to, structure type is as table 2:
Table 2
| Account | Class1 | Value 1 | Type 2 | Value 2 | ... |
| A | 0001 | 10 | 0002 | 15 | ... |
It should be noted that, because not all physics table can be converted to list, can in advance the physics table that may be converted to list be identified out.
Optimization method described in this enforcement, progressively uses different targets to be optimized, and its each optimizing process all has clear and definite optimization aim, so exploitativeness is strong.What is more important, adopts different optimization aim successive optimization, and effect of optimization can be used good and transform little optimization aim, saves human cost as far as possible and reduces risk.
With said method embodiment accordingly, the embodiment of the present application also discloses a kind of optimization device, as shown in Figure 7, comprising:
Statistical module 701, the number of times that the first kind calls is performed for the program i in statistics program set, the described first kind is called as identical the calling of input parameter, described collection of programs is the program recorded in the application layer daily record of service operation, wherein, i=1,2 ... N, described N be greater than 0 integer;
First optimizes execution module 702, the number of times called for successively described program i being performed the first kind is reduced to default value, until when very first time value and the second time value meet pre-conditioned, determine the optimization to described service operation process, the raw runtime that described very first time value is described service operation process, described second time value is successively described program i is performed the working time that number of times that the first kind calls is reduced to the described service operation process after default value.
Particularly, first optimizes execution module is performed the specific implementation that number of times that the first kind calls is reduced to second value successively and can is by described program i: according to being performed number of times that the first kind calls from big to small, sort to described program i; From first type 1 programming in sequence, successively described type 1 programming is performed the number of times that the first kind calls and is reduced to second value.
First optimizes execution module is performed the number of times that the first kind calls successively and is reduced to default value by described program i, until when very first time value and the second time value meet pre-conditioned, determined to the optimization specific implementation of described service operation process can be: successively described program i is performed the number of times that the first kind calls and is reduced to default value, until when very first time value and the second time value meet pre-conditioned, determine the optimization to described service operation process, describedly pre-conditionedly to comprise: the second time value <=very first time value * (1-P), wherein, P is default optimized proportion target, its span is (0, 1).
Alternatively, device described in the present embodiment, can also comprise:
Second optimizes execution module 703, if for successively described program i being performed after number of times that the first kind calls is reduced to second value described, described very first time value and the second time value can not meet pre-conditioned, upgrade the invoked number of times of level of abstraction program in described collection of programs; Upgrade the level of abstraction program invoked wastage in bulk or weight time that invoked number of times is greater than third value; According to wastage in bulk or weight time order from big to small, described level of abstraction program is sorted; Be once many inquiries by the repeatedly single query conversion of the level of abstraction program in described sequence under different parameters successively, until described very first time value is compared with the 3rd time value, the ratio of saving reaches default first and can to save time ratio, and described 3rd time value is be working time of described service operation process after once many inquiries successively by the repeatedly single query conversion of the program in described sequence under different parameters;
And, 3rd optimizes execution module 704, for described by the repeatedly single query conversion of the program in described sequence under different parameters be successively once many inquiry after, if described very first time value is compared with the 3rd time value, the ratio of saving does not reach default first and can to save time ratio, successively the row table called in described service operation process is converted to list, until described very first time value is compared with the 4th time value, the ratio of saving reaches default second and can to save time ratio.
If the function described in the embodiment of the present application method using the form of SFU software functional unit realize and as independently production marketing or use time, can be stored in a computing equipment read/write memory medium.Based on such understanding, the part of the part that the embodiment of the present application contributes to prior art or this technical scheme can embody with the form of software product, this software product is stored in a storage medium, comprising some instructions in order to make a computing equipment (can be personal computer, server, mobile computing device or the network equipment etc.) perform all or part of step of method described in each embodiment of the application.And aforesaid storage medium comprises: USB flash disk, portable hard drive, ROM (read-only memory) (ROM, Read-OnlyMemory), random access memory (RAM, Random Access Memory), magnetic disc or CD etc. various can be program code stored medium.
In this instructions, each embodiment adopts the mode of going forward one by one to describe, and what each embodiment stressed is the difference with other embodiment, between each embodiment same or similar part mutually see.
To the above-mentioned explanation of the disclosed embodiments, professional and technical personnel in the field are realized or uses the application.To be apparent for those skilled in the art to the multiple amendment of these embodiments, General Principle as defined herein when not departing from the spirit or scope of the application, can realize in other embodiments.Therefore, the application can not be restricted to these embodiments shown in this article, but will meet the widest scope consistent with principle disclosed herein and features of novelty.
Claims (10)
1. an optimization method, is characterized in that, comprising:
Program i in statistics program set is performed the number of times that the first kind is called, the described first kind is called as identical the calling of input parameter, and described collection of programs is the program recorded in the application layer daily record of service operation, wherein, i=1,2.......N, described N be greater than 0 integer;
Successively described program i is performed the number of times that the first kind calls and is reduced to default value, until when very first time value and the second time value meet pre-conditioned, determine the optimization to described service operation process, the raw runtime that described very first time value is described service operation process, described second time value is successively described program i is performed the working time that number of times that the first kind calls is reduced to the described service operation process after default value.
2. method according to claim 1, is characterized in that, describedly successively described program i is performed number of times that the first kind calls and is reduced to second value and comprises:
According to being performed number of times that the first kind calls from big to small, described program i is sorted;
From first type 1 programming in sequence, successively described type 1 programming is performed the number of times that the first kind calls and is reduced to second value.
3. method according to claim 1 and 2, is characterized in that, describedly pre-conditionedly to comprise:
Second time value <=very first time value * (1-P), wherein, P is default optimized proportion target, and its span is (0,1).
4. method according to claim 1, is characterized in that, if be performed by described program i after number of times that the first kind calls is reduced to second value successively described, described very first time value and the second time value can not meet pre-conditioned, also comprise:
Upgrade the invoked number of times of level of abstraction program in described collection of programs;
Upgrade the level of abstraction program invoked wastage in bulk or weight time that invoked number of times is greater than third value;
According to wastage in bulk or weight time order from big to small, described level of abstraction program is sorted;
Be once many inquiries by the repeatedly single query conversion of the level of abstraction program in described sequence under different parameters successively, until described very first time value is compared with the 3rd time value, the ratio of saving reaches default first and can to save time ratio, and described 3rd time value is be working time of described service operation process after once many inquiries successively by the repeatedly single query conversion of the program in described sequence under different parameters.
5. method according to claim 4, it is characterized in that, described by the repeatedly single query conversion of the program in described sequence under different parameters be successively once many inquiry after, if described very first time value is compared with the 3rd time value, the ratio of saving does not reach default first and can to save time ratio, also comprises:
Successively the row table called in described service operation process is converted to list, until described very first time value is compared with the 4th time value, the ratio of saving reaches default second and can to save time ratio.
6. an optimization device, is characterized in that, comprising:
Statistical module, the number of times that the first kind calls is performed for the program i in statistics program set, the described first kind is called as identical the calling of input parameter, described collection of programs is the program recorded in the application layer daily record of service operation, wherein, i=1,2.......N, described N be greater than 0 integer;
First optimizes execution module, the number of times called for successively described program i being performed the first kind is reduced to default value, until when very first time value and the second time value meet pre-conditioned, determine the optimization to described service operation process, the raw runtime that described very first time value is described service operation process, described second time value is successively described program i is performed the working time that number of times that the first kind calls is reduced to the described service operation process after default value.
7. device according to claim 6, is characterized in that, described first optimizes execution module is used for being performed by described program i successively number of times that the first kind calls and is reduced to second value and comprises:
Described first optimize execution module specifically for, according to being performed number of times that the first kind calls from big to small, described program i is sorted; From first type 1 programming in sequence, successively described type 1 programming is performed the number of times that the first kind calls and is reduced to second value.
8. the device according to claim 6 or 7, it is characterized in that, described first optimizes execution module is used for successively described program i being performed the number of times that the first kind calls and is reduced to default value, until when very first time value and the second time value meet pre-conditioned, determine to comprise the optimization of described service operation process:
Described first optimize execution module specifically for, successively described program i is performed the number of times that the first kind calls and is reduced to default value, until when very first time value and the second time value meet pre-conditioned, determine the optimization to described service operation process, describedly pre-conditionedly to comprise: the second time value <=very first time value * (1-P), wherein, P is default optimized proportion target, its span is (0,1).
9. device according to claim 6, is characterized in that, also comprises:
Second optimizes execution module, if for successively described program i being performed after number of times that the first kind calls is reduced to second value described, described very first time value and the second time value can not meet pre-conditioned, upgrade the invoked number of times of level of abstraction program in described collection of programs; Upgrade the level of abstraction program invoked wastage in bulk or weight time that invoked number of times is greater than third value; According to wastage in bulk or weight time order from big to small, described level of abstraction program is sorted; Be once many inquiries by the repeatedly single query conversion of the level of abstraction program in described sequence under different parameters successively, until described very first time value is compared with the 3rd time value, the ratio of saving reaches default first and can to save time ratio, and described 3rd time value is be working time of described service operation process after once many inquiries successively by the repeatedly single query conversion of the program in described sequence under different parameters.
10. device according to claim 9, is characterized in that, also comprises:
3rd optimizes execution module, for described by the repeatedly single query conversion of the program in described sequence under different parameters be successively once many inquiry after, if described very first time value is compared with the 3rd time value, the ratio of saving does not reach default first and can to save time ratio, successively the row table called in described service operation process is converted to list, until described very first time value is compared with the 4th time value, the ratio of saving reaches default second and can to save time ratio.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510068301.2A CN104572134B (en) | 2015-02-10 | 2015-02-10 | A kind of optimization method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510068301.2A CN104572134B (en) | 2015-02-10 | 2015-02-10 | A kind of optimization method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104572134A true CN104572134A (en) | 2015-04-29 |
| CN104572134B CN104572134B (en) | 2018-03-06 |
Family
ID=53088300
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510068301.2A Active CN104572134B (en) | 2015-02-10 | 2015-02-10 | A kind of optimization method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104572134B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106293809A (en) * | 2016-07-28 | 2017-01-04 | 东软集团股份有限公司 | The method and device of optimization program |
| CN110143635A (en) * | 2019-04-03 | 2019-08-20 | 杭州电子科技大学 | A kind of optimizing operation method and device of multi-stage Flash Desalination System for Seawater |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5995754A (en) * | 1997-10-06 | 1999-11-30 | Sun Microsystems, Inc. | Method and apparatus for dynamically optimizing byte-coded programs |
| CN101441564A (en) * | 2008-12-04 | 2009-05-27 | 浙江大学 | Method for implementing reconfigurable accelerator custom-built for program |
| US7957950B2 (en) * | 2008-02-28 | 2011-06-07 | Oki Semiconductor Co., Ltd. | Hard/soft cooperative verifying simulator |
| CN102419715A (en) * | 2011-08-29 | 2012-04-18 | 展讯通信（上海）有限公司 | Code compression method and device as well as code execution method |
| CN103648116A (en) * | 2013-11-13 | 2014-03-19 | 上海交通大学 | Low-power-consumption wireless-sensor-network optimization method based on program analysis |
-
2015
- 2015-02-10 CN CN201510068301.2A patent/CN104572134B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5995754A (en) * | 1997-10-06 | 1999-11-30 | Sun Microsystems, Inc. | Method and apparatus for dynamically optimizing byte-coded programs |
| US7957950B2 (en) * | 2008-02-28 | 2011-06-07 | Oki Semiconductor Co., Ltd. | Hard/soft cooperative verifying simulator |
| CN101441564A (en) * | 2008-12-04 | 2009-05-27 | 浙江大学 | Method for implementing reconfigurable accelerator custom-built for program |
| CN102419715A (en) * | 2011-08-29 | 2012-04-18 | 展讯通信（上海）有限公司 | Code compression method and device as well as code execution method |
| CN103648116A (en) * | 2013-11-13 | 2014-03-19 | 上海交通大学 | Low-power-consumption wireless-sensor-network optimization method based on program analysis |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106293809A (en) * | 2016-07-28 | 2017-01-04 | 东软集团股份有限公司 | The method and device of optimization program |
| CN110143635A (en) * | 2019-04-03 | 2019-08-20 | 杭州电子科技大学 | A kind of optimizing operation method and device of multi-stage Flash Desalination System for Seawater |
| CN110143635B (en) * | 2019-04-03 | 2021-08-10 | 杭州电子科技大学 | Optimized operation method and device of multistage flash evaporation seawater desalination system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104572134B (en) | 2018-03-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110058883B (en) | CNN acceleration method and system based on OPU | |
| US10963292B2 (en) | Techniques to manage virtual classes for statistical tests | |
| Guo et al. | A survey of FPGA-based neural network accelerator | |
| CN112001618B (en) | Method for integrating and optimizing construction period assignment, order acceptance and production scheduling | |
| CN102521056B (en) | Task allocation device and task allocation method | |
| CN101446962B (en) | Data conversion method, device thereof and data processing system | |
| CN105868019A (en) | Automatic optimization method for performance of Spark platform | |
| US11657302B2 (en) | Model selection in a forecasting pipeline to optimize tradeoff between forecast accuracy and computational cost | |
| CN106202092B (en) | Data processing method and system | |
| CN108334408B (en) | Code execution method and device, terminal equipment and computer readable storage medium | |
| CN104636202A (en) | Computer system and scheduling method thereof | |
| US20180285011A1 (en) | Memory side accelerator thread assignments | |
| CN1853164B (en) | Combinational method for developing building blocks of DSP compiler | |
| CN110597292A (en) | Unmanned aerial vehicle cluster operation planning method and device | |
| CN112446471B (en) | Convolution acceleration method based on heterogeneous many-core processor | |
| CN103678303A (en) | Method and system used for group set inquiry of database | |
| CN104503381B (en) | A kind of Optimization Scheduling of the production assembling process of mobile phone | |
| CN115081760A (en) | Processing resource management optimization method, device, terminal and storage medium | |
| CN104572134A (en) | Optimization method and optimization device | |
| CN105405002A (en) | Formula data configuration method and system based on SAP platform | |
| CN116932174B (en) | Dynamic resource scheduling method, device, terminal and medium for EDA simulation task | |
| CN113849503A (en) | Open big data processing system, method and medium | |
| Pu et al. | MPEFT: A novel task scheduling method for workflows | |
| Hu et al. | Hydro:{Surrogate-Based} Hyperparameter Tuning Service in Datacenters | |
| CN101071489A (en) | Work flow management system and method for realizing work process automation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |