JP4805594B2 - Program development management device - Google Patents

Description

  The present invention relates to a program development management apparatus for managing program files, and in particular, by converting a program file created in a programming language one or more times, the generation of a machine language program file is advanced step by step. The present invention relates to a program development management apparatus that manages each generation of program files when it is generated.

  At present, a source program created in a high-level language is converted into an object program through compilation, assembly, etc., and is actually used as a target program. Furthermore, various ideas have been made to develop a target program that can actually be used.

  For example, as a target program development method that reduces the burden on the program developer, there is a method of generating an object program having an intermediate expression in the process of compiling and assembling a source program. In this method, by using this intermediate representation, an object program is optimized and a target program is generated (see, for example, Patent Document 1).

  As an efficient target program development method, there is a method that can automatically confirm that a high-level language program has been assembled correctly (see, for example, Patent Document 2).

There is also a method of preparing a makefile that stores information for analyzing the contents of a project that stores a predetermined program and analyzing the contents of the makefile (see, for example, Patent Document 3).
JP 2003-208317 A JP 2003-015910 A JP 2001-216139 A

  However, in the technique disclosed in Patent Document 1, optimization is performed only on an object program in each generation of programs generated in stages from a source program.

  In the technique disclosed in Patent Document 2, optimization is performed only on the source program in each generation of programs generated in stages from the high-level language program serving as the source program.

  In the technique disclosed in Patent Document 3, optimization is performed only on a predetermined generation of programs in each generation of programs generated in stages from a source program.

  The present invention has been made in view of the above points, and provides a program development management apparatus capable of executing optimization of a target program without being aware of each generation of programs generated in stages from a source program. For the purpose.

In the present invention, in order to solve the above-described problem, as illustrated in FIG. 1, a program file created in a programming language, for example, a source file 1aa and an assembly file 1ba are converted one or more times, In a program development management device that manages a source file 1aa, an assemble file 1ba, and an object file 1ca when a machine language program file, for example, an object file 1ca is generated by advancing the generation of programming languages, Among the file names of 1aa, assemble file 1ba, and object file 1ca, file name storage means 1 for storing the file name of the program file that is the source of the latest target program, and source file 1aa Assemble file 1ba, and, as the program file editing means 2 to edit the object file 1ca, if the program file is edited, the file name of the program file of generations of ancestors of the edited program file, the file name storage means 1 And a file name management means 3 to be deleted from the program development management apparatus.

According to such a program development management device, the file name storage unit 1 uses the file name of the source file 1aa, the assembly file 1ba, and the object file 1ca as the source of the program file that is the source of the latest target program. The name is remembered. The program file is edited by the program file editing means 2. If the program file has been edited, by the file name management means 3, the file name of the generation of the program file of the ancestor of the edited program file is deleted from the file name storage means 1.

Further, in the present invention, in order to solve the above problem, as illustrated in FIG. 1, by converting a program file created in a programming language, for example, a source file 1aa and an assembly file 1ba at least once. In the program development management apparatus for managing the source file 1aa, the assemble file 1ba, and the object file 1ca in the case of generating a program file in machine language, for example, an object file 1ca, by gradually progressing the generation of the programming language, Stores the file names of the source file 1aa, the assemble file 1ba, and the object file 1ca generated by editing and converting the source file 1aa, the assemble file 1ba, and the object file 1ca. That the file name storage means 1, the source file 1aa, assemble file 1ba, and a program file editing means 2 for editing the object file 1ca, if the program file is edited, the file name of the edited program file, the file name managing means 3 for setting a flag, have a program file editing means 2 editing flag set program file, and each generation of the program file flag is generated from the set program file A program development management device is provided.

According to such a program development management device, the source file 1aa, the assemble file 1ba, and the source file 1aa, the assemble file 1ba, and the object file 1ca, which are generated by the file name storage means 1, by editing and conversion. The file name of the object file 1ca is stored. The program file is edited by the program file editing means 2. If the program file has been edited, the file name managing unit 3, the file name of the edited program file, flag is set. The program file editing means 2 edits the program file in which the flag is set and each generation of program files generated from the program file in which the flag is set.

  In the present invention, the file name storage means stores the file name of the program file that is the generation source of the latest target program, and the file name management means stores the file name of the program file of the ancestor generation of the edited program file. Then, it is deleted from the file name storage means.

  In this way, since the program file to be edited is individually managed and specified, optimization of the target program can be executed without being aware of each generation of program files generated in stages from the source file.

  In the present invention, the file name storage means stores the file names of a plurality of generations of program files, and the file name management means sets a flag to the file names of the edited program files.

  In this way, since the program file to be edited is individually managed and specified, optimization of the target program can be executed without being aware of each generation of program files generated in stages from the source file.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. Here, the concept of the present invention will be described first, and then specific embodiments of the present invention will be described.
First, the concept of the present invention will be described. FIG. 1 is a conceptual diagram of the present invention.

  Here, the program development management device advances the generation of the programming language step by step by converting the program file created in the programming language, for example, the source file 1aa and the assembly file 1ba at least once. Program file, for example, an object file 1ca is generated to generate a target program. Further, when tuning the target program, the program development management device manages which program file is the program file to be tuned among the source file 1aa, the assemble file 1ba, and the object file 1ca.

  Assume that an assembly file 1ba is generated by compiling the source file 1aa, an object file 1ca is generated by assembling the assembly file 1ba, and a target program is generated. The source file name 1a, the assemble file name 1b, and the object file name 1c are assumed to be the names of the source file 1aa, the assemble file 1ba, and the object file 1ca, respectively.

  As illustrated in FIG. 1, the program development management apparatus includes a file name storage unit 1, a program file editing unit 2, a file name management unit 3, a source file 1aa, an assemble file 1ba, and an object file 1ca. Further, the file name storage means 1 stores only the file name of the program file to be tuned among the source file name 1a, the assemble file name 1b, and the object file name 1c. The file name storage unit 1, the program file editing unit 2, and the file name management unit 3 communicate with each other.

  The file name storage means 1 stores only the file name of the program file to be tuned, which is the source of the latest target program among the file names of the source file 1aa, the assemble file 1ba, and the object file 1ca.

The program file editing means 2 edits the source file 1aa, the assemble file 1ba, and the object file 1ca.
When the program file is edited and applied as the latest program file, the file name management unit 3 deletes the file name of the ancestor generation of the edited program file from the file name storage unit 1.

Next, the operation of the conceptual diagram of FIG. 1 will be described.
For example, first, assume that the source of the latest target program is the source file 1aa. Therefore, it is assumed that only the source file name 1a is stored in the file name storage unit 1.

  Thereafter, the assembly file 1ba is edited by the program file editing means 2 in order to optimize the target program. The file name of the source file 1aa of the ancestor generation of the edited assemble file 1ba is deleted from the file name storage means 1 by the file name management means 3. Here, since the program file to be tuned has been changed from the source file 1aa to the assemble file 1ba, the file name storage means 1 stores the file name of the assemble file 1ba to be tuned.

  Further, the source file 1aa is edited by the program file editing means 2 in order to optimize the target program. Here, since there is no program file of the ancestor generation of the edited source file 1aa, the file name storage means 1 is not edited by the file name management means 3.

  In this way, since the program file to be edited is individually managed and specified, optimization of the target program can be executed without being aware of each generation of program files generated in stages from the source file 1aa.

  Further, since the program file to be edited and the program files of a plurality of generations generated from the program file can be freely edited, editing with a high degree of freedom for the target program can be performed.

  The assemble file 1ba is generated from the source file 1aa and the object file 1ca is generated from the assemble file 1ba. However, the object file 1ca may be generated from the assemble file 1ba that is not generated from the source file 1aa.

Hereinafter, specific embodiments of the present invention will be described. Here, the first embodiment to the second embodiment will be sequentially described.
{First embodiment}
First, the first embodiment will be described. FIG. 2 is a system configuration diagram of the first embodiment.

  Here, the program development management device advances the generation of the programming language step by step by converting the program file created in the programming language, for example, the source file 11a and the assembly file 11b at least once. Program file, for example, the object file 11c is generated, and the target program is generated. Further, when tuning the target program, the program development management device manages which program file is the program file to be tuned among the source file 11a, the assemble file 11b, and the object file 11c.

  Assume that an assembly file 11b is generated by compiling the source file 11a, an object file 11c is generated by assembling the assembly file 11b, and a target program is generated. The source file name 18a, the assemble file name 18b, and the object file name 18c are assumed to be the names of the source file 11a, the assemble file 11b, and the object file 11c, respectively.

  The program development management apparatus includes a file management unit 11, an option management unit 12, a build unit 13, a debug unit 14, a performance information database 15, a tuning unit 16, a feedback unit 17, and a file name storage unit 18. The file management unit 11, the option management unit 12, the build unit 13, the debug unit 14, the performance information database 15, the tuning information database 15a, the tuning unit 16, the feedback unit 17, and the file name storage unit 18 communicate with each other.

The file management unit 11 manages a source file 11a, an assembly file 11b, and an object file 11c as program files for each generation.
The option management unit 12 manages settings such as compilation, assembly, decompilation, and disassembly. Based on the setting, the build unit 13 converts the program file created in the programming language, for example, the source file 11a and the assembly file 11b one or more times to advance the generation of the programming language step by step. A word program file, for example, an object file 11c is generated.

  The debug unit 14 debugs the generated target program. The performance information database 15 stores the debugging result as performance information of the target program. Based on the performance information of the target program, that is, the performance information for the source file 11a, the assemble file 11b, and the object file 11c, the tuning unit 16 generates tuning information that can be easily understood by the program developer, and the tuning information database 15a. Remember me. Thereafter, the tuning unit 16 generates tuning data for each based on the tuning information. The tuning unit 16 tunes the source file 11a, the assemble file 11b, and the object file 11c based on each tuning data. Here, performance information and tuning data are classified for each source file 11a, assembly file 11b, and object file 11c.

  When the program file is edited and applied as the latest program file, the feedback unit 17 deletes the file name of the ancestor generation program file of the edited program file from the file name storage unit 18. As a result, the file name storage unit 18 stores the file name of the program file that is the source of the latest target program among the file names of the source file 11a, the assemble file 11b, and the object file 11c.

Note that the tuning may be executed manually or may be automatically executed by setting tuning conditions in advance.
Next, the operation of the system configuration diagram of FIG. 2 will be described.

  For example, first, assume that the source of the latest target program is the source file 11a. Therefore, it is assumed that only the source file name 18a is stored in the file name storage unit 18.

  Thereafter, the assembly file 11b is tuned by the tuning unit 16 in order to optimize the target program. The file name of the source file 11a of the ancestor generation of the tuned assemble file 11b is deleted from the file name storage unit 18 by the feedback unit 17. Here, since the program file to be tuned is changed from the source file 11a to the assemble file 11b, the file name storage unit 18 stores the file name of the assemble file 11b to be tuned.

  Further, the source file 11a is tuned by the tuning unit 16 in order to optimize the target program. Here, since there is no ancestor generation program file of the tuned source file 11 a, the file name storage unit 18 is not edited by the feedback unit 17.

Next, the contents stored in the file name storage unit 18 will be described. FIG. 3 is a diagram illustrating a change example of the storage contents of the file name storage unit according to the first embodiment.
First, as illustrated in FIG. 3A, seven file names are stored in the file name storage unit 18. Specifically, “a.c”, “b.c”, and “c.c” are stored as the source file name 18 a in the file name storage unit 18, and “ds” as the assembled file name 18 b. , And “es” are stored, and “fo” and “go” are stored as the object file name 18c.

  Thereafter, as illustrated in FIG. 3B, “c.s” generated by assembling “c.c” is tuned, and “c.c” is deleted in the file name storage unit 18. “Cs” which is the generation source of the latest target program is stored.

Next, the performance information database 15 will be described. FIG. 4 is a diagram illustrating an example of the performance information database.
As illustrated in FIG. 4, the performance information database 15 includes an address, an instruction name, a function name, a Taken number, a Not_Taken number, and a Taken rate. In the performance information database 15, the items arranged in the horizontal direction are related to each other.

  The address is an address of an instruction in the object file 11c generated from the assemble file 11b. The instruction name is the name of the instruction in the assemble file 11b generated from the source file 11a. The function name is the name of the instruction in the source file 11a. The number of Taken is the number of hit failures in branch prediction of a branch instruction. The Not_Take number is the number of successful hits in branch prediction of a branch instruction. The Taken rate is the probability of success in branch prediction of a branch instruction.

Next, the tuning information database 15a will be described. FIG. 5 is a diagram illustrating an example of a tuning information database.
The tuning information database 15a includes file types and tuning information as illustrated in FIG. In the tuning information database 15a, the items arranged in the horizontal direction are related to each other.

  The file type indicates one of a source file 11a, an assemble file 11b, and an object file 11c. The tuning information is tuning information for each generation program file. For example, the tuning information for the source file 11a is File. This is information that the Taken rate of the if clause in the 12th row of c is 87%. The tuning information for the assemble file 11b is File. This is information that the Taken rate of the Bne branch instruction on line 84 of s is 87%. The tuning information for the object file 11c is File. This is information that the Taken rate of 0x56 instruction (Bne branch instruction) at address 0x1234 at o is 87%.

Next, an example of data cache tuning data will be described. FIG. 6 is a diagram illustrating an example of data cache tuning data.
As illustrated in FIG. 6, the data cache tuning data 41 includes pre-tuning data, post-tuning data, and tuning conditions for the source file 11a, the assembly file 11b, and the object file 11c.

  Here, the tuning content cannot be expressed by the source file 11a and the object file 11c, but can be expressed by the assemble file 11b. In the case of automatic tuning, a pre-data load instruction is added when a tuning condition that a data cache miss is 80% or more is satisfied. Thereafter, since the assemble file 11b is tuned, the file name storage unit 18 stores the file name of the assemble file 11b.

Next, an example of branch tuning data will be described. FIG. 7 is a diagram showing an example of branch tuning data.
As illustrated in FIG. 7, the branch tuning data 42 includes data before tuning, data after tuning, and tuning conditions for the source file 11a, the assembly file 11b, and the object file 11c.

  Here, the tuning content can be expressed by a source file 11a, an object file 11c, and an assembly file 11b. When tuning is performed automatically, if the tuning condition that the execution frequency of the “if” or “else” clause is 80% or more is satisfied, “#pragma statement if 87” is added for the source file 11a, and hint bit for the assemble file 11b. “00” is changed to “11”, and in the case of the object file 11c, the hint bit is changed from “00” to “11”. Thereafter, the file name storage unit 18 stores the file name of the program file that has been tuned.

Note that the hint bit is implemented in hardware as a branch prediction function, and the expression "#pragma statement if 87" can be compiled.
Next, processing of the program development management device according to the first embodiment will be described. FIG. 8 is a flowchart illustrating an example of processing of the program development management device according to the first embodiment.

  {Step S11} The file management unit 11 starts managing the source file 11a, the assembly file 11b, and the object file 11c as the program files for each generation.

{Step S12} The file name storage unit 18 stores the file name of the program file that is the source of the latest target program.
{Step S13} The build unit 13 advances the generation of the programming language step by step by converting the program file created in the programming language, for example, the source file 11a and the assembly file 11b at least once. Program file, for example, the object file 11c is generated.

  {Step S14} The tuning unit 16 acquires performance information for the source file 11a, the assemble file 11b, and the object file 11c as program files of each generation.

{Step S15} The tuning unit 16 generates tuning data for each.
{Step S16} The tuning unit 16 determines whether or not to tune. When tuning, the process proceeds to step S17, and when not tuned, the process ends.

{Step S17} The tuning unit 16 tunes the source file 11a, the assemble file 11b, and the object file 11c.
{Step S18} When the program file is edited and applied as the latest program file, the feedback unit 17 deletes the file name of the ancestor generation program file of the edited program file from the file name storage unit 18. .

Next, tuning data generation processing will be described. FIG. 9 is a flowchart illustrating an example of a tuning data generation process.
{Step S15a} The tuning unit 16 generates tuning information based on the performance information for the source file 11a, the assemble file 11b, and the object file 11c.

  {Step S15b} The tuning unit 16 determines whether tuning information has been generated. If so, the process proceeds to step S15c. If not, the process ends.

  {Step S15c} The tuning unit 16 determines whether or not the tuning content can be expressed by the source file 11a. If it can be expressed, the process proceeds to step S15d, and if it cannot be expressed, the process proceeds to step S15e.

{Step S15d} The tuning unit 16 generates tuning data for the source file 11a.
{Step S15e} The tuning unit 16 determines whether or not the tuning content can be expressed by the assemble file 11b. If it can be expressed, the process proceeds to step S15f, and if it cannot be expressed, the process proceeds to step S15g.

{Step S15f} The tuning unit 16 generates tuning data for the assemble file 11b.
{Step S15g} The tuning unit 16 determines whether or not the tuning content can be expressed by the object file 11c. If it can be expressed, the process proceeds to step S15h. If it cannot be expressed, the process ends.

{Step S15h} The tuning unit 16 generates tuning data for the object file 11c.
{Step S15i} The tuning unit 16 stores tuning data for each.

Next, the tuning process will be described. FIG. 10 is a flowchart illustrating an example of tuning processing.
{Step S17a} The tuning unit 16 determines whether to tune the source file 11a. When tuning, the process proceeds to step S17b, and when not tuned, the process proceeds to step S17c.

{Step S17b} The tuning unit 16 tunes the source file 11a based on performance information or tuning information.
{Step S17c} The tuning unit 16 determines whether or not to tune the assemble file 11b. When tuning, the process proceeds to step S17d, and when not tuned, the process proceeds to step S17e.

{Step S17d} The tuning unit 16 tunes the assemble file 11b based on performance information or tuning information.
{Step S17e} The tuning unit 16 tunes the object file 11c based on performance information or tuning information.

Next, the storage process will be described. FIG. 11 is a flowchart illustrating an example of the storage process.
{Step S18a} The feedback unit 17 determines whether the file name of the program file that is the generation source of the latest target program has been stored. If stored, the process ends. If not stored, the process proceeds to step S18b.

{Step S18b} The feedback unit 17 deletes the file name of the program file of the ancestor generation of the edited program file from the file name storage unit 18.
{Step S18c} The file name storage unit 18 stores the file name of the program file from which the latest target program is generated.

  In this way, since the program file to be tuned is individually managed and specified, the target program can be optimized without being aware of each generation of program files that are generated in stages from the source file 11a.

  In addition, since the program file to be tuned and the multiple generation program files generated from the program file can be freely tuned, the most effective program file can be selected and the target program can be tuned with a high degree of freedom. .

  The assemble file 11b is generated from the source file 11a and the object file 11c is generated from the assemble file 11b. However, the object file 11c may be generated from the assemble file 11b that is not generated from the source file 11a.

It is also possible to tune without using tuning information that is easy for program developers to understand.
{Second Embodiment}
Next, a second embodiment will be described. FIG. 12 is a system configuration diagram of the second embodiment.

  Here, the program development management apparatus advances the generation of the programming language step by step by converting the program file created in the programming language, for example, the source file 21a and the assembly file 21b at least once. Program file, for example, the object file 21c is generated, and the target program is generated. Further, when tuning the target program, the program development management device manages which program file is the program file to be tuned among the source file 21a, the assemble file 21b, and the object file 21c.

  Assume that an assembly file 21b is generated by compiling the source file 21a, an object file 21c is generated by assembling the assembly file 21b, and a target program is generated. The source file name 29a, the assemble file name 29b, and the object file name 29c are assumed to be the names of the source file 21a, the assemble file 21b, and the object file 21c, respectively.

  The program development management apparatus includes a file management unit 21, an option management unit 22, a build unit 23, a debug unit 24, a performance information database 25, a tuning information database 25a, a tuning unit 26, a build target flag setting unit 28, and a file name storage The unit 29 is configured. The file management unit 21, the option management unit 22, the build unit 23, the debug unit 24, the performance information database 25, the tuning unit 26, the build target flag setting unit 28, and the file name storage unit 29 communicate with each other.

The file management unit 21 manages a source file 21a, an assembly file 21b, and an object file 21c as program files for each generation.
The option management unit 22 manages settings such as compilation, assembly, decompilation, and disassembly. The build unit 23 based on the setting and a build target flag to be described later performs step-by-step programming by converting a program file created in a programming language, for example, the source file 21a and the assembly file 21b at least once. The language generation is advanced to generate a machine language program file, for example, an object file 21c.

  The debug unit 24 debugs the generated target program. The performance information database 25 stores the debug result as performance information of the target program. Based on the performance information of the target program, that is, the performance information for the source file 21a, the assemble file 21b, and the object file 21c, the tuning unit 26 generates tuning information that can be easily understood by the program developer, and the tuning information database 25a. Remember me. Thereafter, the tuning unit 26 generates tuning data for each based on the tuning information. The tuning unit 26 tunes the source file 21a, the assemble file 21b, and the object file 21c based on each tuning data. Here, the performance information and the tuning data are classified for each source file 21a, assemble file 21b, and object file 21c.

  When the program file is edited and applied as the latest program file, the build target flag setting unit 28 sets the build target flag to the file name of the edited program file. As a result, the file name storage unit 29 stores the file name of the program file that is the generation source of the latest target program in which the build target flag is set. The file name storage unit 29 stores the file names of the source file 21a, the assemble file 21b, and the object file 21c that are generated by editing and converting the program file.

Note that the tuning may be executed manually or may be automatically executed by setting tuning conditions in advance.
The build target flag may be set manually by the program developer, or may be set automatically by setting conditions in advance.

Next, the operation of the system configuration diagram of FIG. 12 will be described.
For example, first, assume that the source of the latest target program is the source file 21a. Therefore, it is assumed that the build target flag is set in the source file name 29a. Further, it is assumed that the file name storage unit 29 stores a source file name 29a, an assemble file name 29b, and an object file name 29c.

  Thereafter, the assembly file 21b is tuned by the tuning unit 26 in order to optimize the target program. A build target flag is set by the build target flag setting unit 28 to the file name of the tuned assemble file 21b. Here, the program file to be tuned is changed from the source file 21a to the assemble file 21b.

  Further, the source file 21a is tuned by the tuning unit 26 in order to optimize the target program. Since the build target flag is already set in the file name of the tuned source file 21a, the build target flag is not edited.

Next, the contents stored in the file name storage unit 29 will be described. FIG. 13 is a diagram illustrating a change example of the contents stored in the file name storage unit according to the second embodiment.
First, as illustrated in FIG. 13A, seven file names are stored in the file name storage unit 29. Specifically, “a.c”, “b.c”, and “c.c” are stored as the source file name 29a in the file name storage unit 29, and “d.s” is stored as the assembled file name 29b. , And “es” are stored, and “fo” and “go” are stored as the object file name 29c.

  After that, as illustrated in FIG. 13B, “ac”, “bc”, “cc”, “ds”, and “es” are built, and the latest “%” Is set in “ac”, “bc”, “cc”, “ds”, and “es”, which are the generation sources of the target program.

Next, processing of the program development management device according to the second embodiment will be described. FIG. 14 is a flowchart illustrating an example of processing of the program development management device according to the second embodiment.
{Step S21} The file management unit 21 starts managing the source file 21a, the assembly file 21b, and the object file 21c as program files for each generation.

{Step S22} The file name storage unit 29 stores the file name of the program file from which the latest target program is generated.
{Step S23} The build unit 23 automatically recognizes the build target flag.

  {Step S24} The build unit 23 advances the generation of the programming language step by step by converting the program file created in the programming language, for example, the source file 21a and the assembly file 21b at least once. Program file, for example, the object file 21c is generated.

  {Step S25} The file name storage unit 29 automatically stores the file names of the source file 21a, the assemble file 21b, and the object file 21c generated by editing and converting the program file.

  {Step S26} The tuning unit 26 acquires performance information for the source file 21a, the assemble file 21b, and the object file 21c as the program files for each generation.

{Step S27} The tuning unit 26 generates tuning data for each.
{Step S28} The tuning unit 26 determines whether or not to tune. When tuning, the process proceeds to step S29, and when not tuned, the process ends.

{Step S29} The tuning unit 26 tunes the source file 21a, the assemble file 21b, and the object file 21c.
{Step S30} When the program file is edited and applied as the latest program file, the build target flag setting unit 28 sets the build target flag in the file name of the edited program file.

Here, the process of step S27 and the process of step S29 are respectively the same as the process of step S15 and the process of step S17 in the first embodiment.
Note that the same tuning content as that executed for a certain program file can be executed for other generations of program files.

  In this way, the program file to be tuned is individually managed and specified by the build target flag, so that the target program can be optimized without being aware of each generation of the program file generated in stages from the source file 21a. Can be executed.

  In addition, since it is possible to freely tune the program file to be tuned with the build target flag set and the multiple generation program files generated from the program file, the most effective program file can be selected, and the target program can be selected. Tuning with a high degree of freedom is possible.

  The assemble file 21b is generated from the source file 21a and the object file 21c is generated from the assemble file 21b. However, the object file 21c may be generated from the assemble file 21b that is not generated from the source file 21a.

It is also possible to tune without using tuning information that is easy for program developers to understand.
In addition, the program development management apparatus according to the second embodiment can execute the same tuning content as the tuning content executed for a certain program file for a program file of another generation. The contents stored in the file name storage unit 29 in this case will be described. FIG. 15 is a diagram illustrating another example of changes stored in the file name storage unit according to the second embodiment.

  First, as illustrated in FIG. 15A, the file name storage unit 29 stores the same storage content as that illustrated in FIG. 13B and is tuned to “% ac”, In addition, “*” is set in “% go”.

  After that, as illustrated in FIG. 15B, the same tuning contents as those executed for “% a.c *” and “% g.o *” are stored in programs of other generations. The file is executed, and “a.s *” and “a.o *” are generated for “% a.c *”, and “g.s” is generated for “% g.o *”. * ”Is generated.

  Here, the same tuning content as that executed for a certain program file is executed by compiling, assembling, decompiling, disassembling, etc., when executing for other generations of program files. In this case, “*” is not set and stored in the file name storage unit 29. Further, when a fine tuning without a redundant part such as partially inserting an assembly language into the C language is successful, “*” is set and stored in the file name storage unit 29. In addition, when a fine tuning with no redundant part is executed and fails, a non-build target flag or the like is set and stored in the file name storage unit 29 or not stored in the file name storage unit 29. Both “a.s *” and “a.s” may be stored, or only “a.s *” may be stored.

  In this way, all generations of program files with “*” set have the same logic, and therefore, generations of program files with “*” set can be targeted for tuning.

  In addition, when executing tuning contents similar to the tuning contents executed for a program file for other generations of program files, it may be executed manually or automatically by setting conditions in advance. May be executed.

  Further, the above processing functions can be realized by a computer. In that case, a program describing the processing contents of the functions that the program development management apparatus should have is provided. By executing the program on a computer, the above processing functions are realized on the computer. The program describing the processing contents can be recorded on a computer-readable recording medium. Examples of the computer-readable recording medium include a magnetic recording device, an optical disc, a magneto-optical recording medium, and a semiconductor memory. Examples of the magnetic recording device include a hard disk drive (HDD), a flexible disk (FD), and a magnetic tape. Optical discs include DVD (Digital Versatile Disc), DVD-RAM (Random Access Memory), CD-ROM (Compact Disc Read Only Memory), CD-R (Recordable) / RW (ReWritable), and the like. Magneto-optical recording media include MO (Magneto-Optical disk).

  When distributing the program, for example, a portable recording medium such as a DVD or a CD-ROM in which the program is recorded is sold. Further, the program can be stored in a storage device of the server computer, and the program can be transferred from the server computer to another computer via the network.

  The computer that executes the program stores, for example, the program recorded on the portable recording medium or the program transferred from the server computer in its own storage device. Then, the computer reads the program from its own storage device and executes processing according to the program. The computer can read the program directly from the portable recording medium and execute processing according to the program. Further, each time the program is transferred from the server computer, the computer can sequentially execute processing according to the received program.

Next, the hardware configuration of the program development management apparatus as a computer will be described. FIG. 16 is a diagram illustrating an example of a hardware configuration of a computer.
The entire computer 300 is controlled by a CPU (Central Processing Unit) 301. A RAM 302, a hard disk drive 303, a graphic processing device 304, an input interface 305, and a communication interface 306 are connected to the CPU 301 via a bus 307.

  The RAM 302 temporarily stores at least part of an OS (Operating System) program and application programs to be executed by the CPU 301. The RAM 302 stores various data necessary for processing by the CPU 301. The HDD 303 stores an OS and application programs.

  A monitor 310 is connected to the graphic processing device 304. In addition, the graphic processing device 304 displays an image on the screen of the monitor 310 in accordance with a command from the CPU 301. A keyboard 320 and a mouse 330 are connected to the input interface 305. The input interface 305 transmits a signal sent from the keyboard 320 or the mouse 330 to the CPU 301 via the bus 307.

  The communication interface 306 is connected to the network 340. The communication interface 306 transmits and receives data to and from other computers via the network 340.

With the hardware configuration described above, the processing functions of the embodiment of the present invention can be realized.
(Supplementary note 1) When a program file created in a programming language is converted one or more times to generate a machine language program file by progressively generating a programming language generation, each generation of program files is managed. In the program development management device,
File name storage means for storing the file name of the program file that is the generation source of the latest target program among the file names of the multiple generation program files;
Program file editing means for editing each program file,
When the program file is edited and applied as the latest program file, the file name management means for deleting the file name of the ancestor generation of the edited program file from the file name storage means,
A program development management device comprising:

  (Additional remark 2) The said program file edit means edits the program file used as the production | generation source of the said latest target program, and the program file of each generation produced | generated from the said program file, The additional remark 1 characterized by the above-mentioned. Program development management device.

(Supplementary note 3) When a program file created in a programming language is converted one or more times by a computer to generate a machine language program file by stepping up the generation of the programming language, the program file for each generation In the program development management method for managing
A step of storing a file name of a program file, which is a generation source of the latest target program, among file names of a plurality of generations of program files;
A program file editing means for editing each program file;
A step of deleting a file name of a program file of an ancestor generation of the edited program file from the file name storage means when the program file is edited and applied as the latest program file; When,
A program development management method comprising:

(Supplementary note 4) When a program file created in a programming language is converted one or more times, the generation of a machine language is generated by stepping up the generation of the programming language step by step, and the program file of each generation is managed In the program development management program,
On the computer,
A step of storing a file name of a program file, which is a generation source of the latest target program, among file names of a plurality of generations of program files;
A program file editing means for editing each program file;
A step of deleting a file name of a program file of an ancestor generation of the edited program file from the file name storage means when the program file is edited and applied as the latest program file; When,
A program development management program characterized by causing a process including

(Supplementary note 5) When a program file created in a programming language is converted one or more times to generate a machine language program file by stepping up the generation of the programming language, the program file of each generation is managed. In a computer-readable recording medium recording a program development management program,
On the computer,
A step of storing a file name of a program file, which is a generation source of the latest target program, among file names of a plurality of generations of program files;
A program file editing means for editing each program file;
A step of deleting a file name of a program file of an ancestor generation of the edited program file from the file name storage means when the program file is edited and applied as the latest program file; When,
The computer-readable recording medium which recorded the program development management program characterized by performing the process containing these.

(Supplementary note 6) When a program file created in a programming language is converted one or more times, a generation of machine language program files is generated in stages, and program files for each generation are managed. In the program development management device,
File name storage means for storing file names of a plurality of generations of program files generated by editing and converting the program file;
Program file editing means for editing each program file,
When the program file is edited and applied as the latest program file, file name management means for setting a flag to the file name of the edited program file;
A program development management device comprising:

  (Additional remark 7) The program development of Additional remark 6 which further has an edit reflection means to perform the edit content similar to the edit content performed with respect to the said program file with respect to the program file of another generation Management device.

  (Supplementary note 8) The program development management device according to supplementary note 6, wherein the program file editing means edits the program file in which the flag is set, and each generation program file generated from the program file. .

(Supplementary note 9) When a program file created in a programming language is converted one or more times by a computer, the generation of a machine language is generated by stepping up the generation of the programming language step by step. In the program development management method for managing
A file name storage means for storing file names of a plurality of generations of program files generated by editing and converting the program file;
A program file editing means for editing each program file;
A file name management means, when the program file is edited and applied as the latest program file, setting a flag to the file name of the edited program file;
A program development management method comprising:

(Supplementary note 10) When a program file created in a programming language is converted one or more times to generate a machine language program file by progressively generating the programming language generation, the program file of each generation is managed. In the program development management program,
On the computer,
A file name storage means for storing file names of a plurality of generations of program files generated by editing and converting the program file;
A program file editing means for editing each program file;
A file name management means, when the program file is edited and applied as the latest program file, setting a flag to the file name of the edited program file;
A program development management program characterized by causing a process including

(Supplementary note 11) When a program file created in a programming language is converted one or more times to generate a machine language program file by progressively generating a programming language generation, the program file for each generation is managed. In a computer-readable recording medium recording a program development management program,
On the computer,
A file name storage means for storing file names of a plurality of generations of program files generated by editing and converting the program file;
A program file editing means for editing each program file;
A file name management means, when the program file is edited and applied as the latest program file, setting a flag to the file name of the edited program file;
The computer-readable recording medium which recorded the program development management program characterized by performing the process containing these.

It is a conceptual diagram of this invention. It is a system configuration figure of a 1st embodiment. It is a figure which shows the example of a change of the memory content of the file name memory | storage part in 1st Embodiment. It is a figure which shows the example of a performance information database. It is a figure which shows the example of a tuning information database. It is a figure which shows the example of the tuning data of a data cache. It is a figure which shows the example of the tuning data of a branch. It is a flowchart which shows the example of a process of the program development management apparatus of 1st Embodiment. It is a flowchart which shows the example of the production | generation process of tuning data. It is a flowchart which shows the example of a tuning process. It is a flowchart which shows the example of a memory | storage process. It is a system configuration figure of a 2nd embodiment. It is a figure which shows the example of a change of the memory content of the file name memory | storage part in 2nd Embodiment. It is a flowchart which shows the example of a process of the program development management apparatus of 2nd Embodiment. It is a figure which shows the other example of a change of the memory content of the file name memory | storage part in 2nd Embodiment. It is a figure which shows the example of the hardware constitutions of a computer.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 File name memory | storage means 1a Source file name 1b Assemble file name 1c Object file name 1aa Source file 1ba Assemble file 1ca Object file 2 Program file edit means 3 File name management means

Claims (4)

A program development management device that manages program files for each generation when a program file created in a programming language is converted one or more times to generate a program file in machine language by stepping through the generations of the programming language. In
File name storage means for storing the file name of the program file that is the generation source of the latest target program among the file names of the multiple generation program files;
Program file editing means for editing each program file,
If the program file has been edited, and the file name management means that the file name of the program file of generations of ancestors of the edited the program file is deleted from the file name storage means,
A program development management device comprising: The program file editing means edits a program file that is a generation source of the latest target program, and a program file of each generation generated from the program file that is a generation source of the latest target program. The program development management apparatus according to claim 1. A program development management device that manages program files for each generation when a program file created in a programming language is converted one or more times to generate a program file in machine language by stepping through the generations of the programming language. In
File name storage means for storing file names of a plurality of generations of program files generated by editing and converting the program file;
Program file editing means for editing each program file,
If the program file has been edited, the file name of the edited the program file, and the file name managing means for setting the flag,
Have
The program file editing means edits a program file in which the flag is set and a program file for each generation generated from the program file in which the flag is set.
A program development management device characterized by that.   4. The program development management apparatus according to claim 3, further comprising edit reflection means for executing edit contents similar to the edit contents executed for the program file for program files of other generations.

Images