JPH05346858A - Program editing method - Google Patents

Abstract

PURPOSE:To provide the program editing method which can efficiently develop a program with a single program language concerning the program conventionally developed with each of plural program languages corresponding to functions for each section in a digital computer system. CONSTITUTION:A first compiler 11 analyzes a source program 10 containing plural functions provided with the functions of difference use for each function, and the source program is divided and converted into next-stage programs 12 (12a-12c) described in the plural program languages corresponding to the use. Further, the next-stage programs 12 are converted to machine wood programs 14 (14a-14c) by second compilers 13 (13a-13c) corresponding to the respective program languages. Moreover, such a method is applied to a cross compiler environment as well, and a program used for plural computers is obtained by editing a single program on a single computer.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a compiler in a digital computer, and more particularly to a compiler and a program editing method when a plurality of programming languages operate in the same computer system.

[0002]

2. Description of the Related Art Conventionally, a program editing method such as a compiler for converting a source program written in a high-level programming language such as FORTRAN into a lower language such as a machine language has been known and used. The conventional program editing method for converting a source program written in a high-level programming language into a lower language using a compiler will be described below. FIG. 9 is a diagram showing a data flow in a conventional compiler. Source program 12b is FORT
It is a program written in RAN. Compiler 13
Reference numeral b is a program that performs a grammatical check of a source program described in FORTRAN and the like, and performs a process (compile) of translating it into a machine language. The object program 14b converts the source program 12b into the compiler 13b.
It is a program translated into machine language by. All of the above programs exist in the form of files in the storage device on the same computer.

A computer user first uses a source program 12 by using a document creation program called an editor.
Create b in FORTRAN. After the creation of the source program 12b is completed, the computer user specifies the source program 12b to be compiled, and the compiler 13
Start b. The activated compiler 13b performs grammar check, variable check, etc. of the source program 12b, and then translates the source program 12b into a machine language to create an object program 14b.

[Problems to be Solved by the Invention]

Since the conventional program editing method is configured as described above, one type of compiler can handle only one type of high-level programming language. When a high-level programming language is used, a single language can be used for programs for most purposes. On the other hand, each high-level programming language is suitable for its purpose and unsuitable, and for example, FORTRAN is suitable for numerical operation, but input / output (I / I) closely related to computer hardware is used.
It is not suitable for applications such as handling O). For this reason, for example, using FORTRAN, numerical operation and I /
When a program having a mixture of O operations is created, a problem arises in which an object program is generated which has a high numerical calculation process but a slow I / O operation process.

In order to solve the above problem, the computer user analyzes the specifications of the source program, describes the numerical operation part in FORTRAN, and writes the I / O operation part in C, for example.
There is a method to describe in. When this method is used, the computer user is required to have a high level of ability to analyze the source program specifications and write each part in a different high-level programming language, which causes a problem in developing the source program. there were. The present invention has been made in view of the above problems of the prior art, and automatically selects the optimum high-level programming language for the processing of each part from a single source program, and Can be converted into the program of the next stage described in, and the above conversion can be applied to a computer network composed of a plurality of computers, and moreover, the resources of the entire computer network can be effectively used when performing the conversion processing. An object of the present invention is to provide a program editing method capable of performing.

[0006]

In order to solve the above problems, a program editing method according to the present invention has a plurality of high-level programming language editing functions for a source program written in one high-level programming language. Then, analyze the processing contents of each part of the source program, and according to the processing contents, among the plurality of programming languages,
An optimum next-stage programming language is selected for the part, and the source program is converted into the selected programming language for each part. Also,
The conversion processing according to the processing content is performed according to the type of computer to which the processing is applied.

Further, in a computer network composed of a plurality of computers and using different computers for each processing content, it is possible to create a program written in the programming language of the next stage corresponding to each computer. Characterize. Further, in one computer, analyze the processing contents for each part of the source program, and select the optimum next-stage programming language for that part from among the plurality of programming languages according to the processing contents, It is characterized in that the source program is converted into a selected programming language for each of the parts, and the converted program is further converted into a machine language in another computer.

[0008]

With respect to a source program written in one kind of high-level programming language, it has a plurality of high-level programming language editing functions, analyzes the processing contents of each part of the source program, and according to the processing contents, An optimum high-level program according to the processing content is selected by selecting the optimum next-stage programming language for that part from among a plurality of programming languages and converting the above source program into the selected programming language for each part. It is possible to select the language and to realize portability between the computers of the converted program. In addition, since the conversion processing according to the processing content is performed according to the type of computer to be applied, it has higher functions and performance than the standard high-level programming language to machine language conversion compiler. It enables the use of a compiler created specifically for the computer.

Further, in a computer network composed of a plurality of computers and using different computers depending on the processing contents, by creating a program written in the next-stage programming language corresponding to each computer, It is possible to obtain an optimum object program for the entire computer network. Further, in one computer, analyze the processing contents for each part of the source program, and select the optimum next-stage programming language for that part from among the plurality of programming languages according to the processing contents, For each of the above parts, convert the above source program into the selected programming language,
In another computer, further, by converting the converted program into a machine language, it is possible to effectively use the computer resources of the entire computer network.

[0010]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment of the present invention will be described with reference to the drawings. FIG. 2 is a diagram showing the configuration of the processor element (PE) shown in FIG. The present invention is
A symbol processing computer as shown in FIG. 2 (E in FIG.
U) and a man-machine interface and a computer (RM in FIG. 2) used for storage etc. are applied to PE of a parallel computer network. The evaluator (EU) and the resource manager (RM) are connected by the RM interface (RM I / F), and send function data from the EU to the RM. Further, information data is sent from the EU through the main memory interface (MM I / F) and sent to the main memory (MM) and RM.
The RM I / F and MM I / F are connected to the ALU via an internal bus (IBUS) in the EU. Also, RM
The I / F is connected to the WS via an evaluator interface (EU I / F) in the RM. EU is A
LU, masker, shifter, local memory, register file, user stack, dispatch table, instruction cache, system controller, control memory, IBUS, diagnostic statics, RM I / F and MM I / F. WSR3000, diagnostics interface, communication interface, secondary storage, M
M, a pointer manipulator, an EU I / F, and a CRT display.

FIG. 3 is a summary of the part relating to the present invention of the computer network of FIG. 2 for explaining the present embodiment. In FIG. 3, the first computer 2 is the RM of FIG.
2, the CPU 20 corresponds to WS, the memory 22 corresponds to the MM in FIG. 2, the display device 24 corresponds to the CRT display in FIG.
The communication interface 25 corresponds to the RM I / F in FIG. The magnetic storage device 26 corresponds to the secondary storage of FIG.
Note that the bus 21 and the I / O 23 are omitted in FIG.

In FIG. 3, the CPU 20 has a memory 22.
Executing the programs stored in
And a CPU that controls the first computer 2 and its peripheral circuits. The bus 21 is a bus for transmitting information between the respective parts of the first computer 2. The memory 22 is a memory that stores programs and data. I / O2
An I / O control device 3 operates the display device 24, the communication interface 25, and the magnetic storage device 26.
The display device 24 is a CRT display that displays the output data of the first computer 2. Communication interface 2
Reference numeral 5 is an interface for communicating with other computers. The magnetic storage device 26 is a storage device that stores data.

FIG. 4 is a diagram showing a program conversion method of the program editing method of the present invention. Source program 10
Is a source program written in LISP. The first compiler 11 converts the source program 10 into the LIS
It is a compiler that converts each high-level programming language such as P, FORTRAN, and C. Next-stage program 12a,
b and c are programs in each high-level programming language generated by the first compiler 11, LISP, FORTRAN, and C for each partial portion of the source program 10. The second compilers 13a, b, c are compilers for converting the next-stage programs 12a, b, c into machine language. The machine language programs 14a, b, c are the first compiler 11 and the second compilers 13a, b,
c is a machine language program generated by converting the source program 10. Source program 1 above
0 and each compiler are stored in the magnetic storage device 26 of the first computer 2 and read out in the memory 22 as needed. The next-stage programs 12a, b, c and machine language programs 14a, b, c are sequentially generated according to the operations of the first compiler 11 and the second compilers 13a, b, c, and are stored in the memory 22 or the magnetic storage device 26. Remembered.

The operation of the program editing method of the present invention will be described below. FIG. 5 is a flowchart of the processing of the program conversion method of the program editing method of the present invention.
In step 01 (S01), the first compiler 1
1 performs a grammatical check of the source program 10 and the like, and then analyzes each function included in the source program 10. The source program 10 is assumed to include the functions 10a to 10d that realize the functions shown in FIG. Here, the function 10a is a function for receiving the data A from another computer connected to the communication interface 25,
The function 10b is a function for performing symbol processing on the data A, the function 10c is a function for performing numerical operation processing on the data A, and the function 10d is for displaying the above processing results on the display device 24, and It is assumed that the function is for storing in the storage device 26. The first compiler 11 classifies the types of instructions included in each function 10a to 10d of the source program 10, and counts the number of list processing instructions, the number of numerical calculation instructions, and the number of I / O instructions of each function. To do. Here, the function 10a includes three list processing instructions, ten numerical calculation instructions, and 50 I / O instructions, and the function 10b includes list processing instructions 200, numerical calculation instructions 40,
The function 10c includes 0 I / O instructions, the function 10c includes 20 list processing instructions, the numerical calculation instruction 400, and the I / O instruction 0, and the function 10d includes 10 list processing instructions and the numerical calculation. It is assumed that the result of including 10 instructions for I / O and 500 instructions for I / O has occurred.

In step 02 (S02), the first compiler 11 sequentially converts the function 10a into C, which is a programming language suitable for I / O operation, based on the result of the number of instructions obtained as a result of the processing of S01. Step 03 to do
Proceed to. The process proceeds to step 04 to convert the function 10b into LISP which is a programming language suitable for list processing. The process proceeds to step 04 to convert the function 10c into FORTRAN, which is a programming language suitable for numerical calculation processing. In order to convert the function 10d into C which is a programming language suitable for I / O operation, the process proceeds to step 04. Is performed.

In step 03 (S03), the first compiler 11 executes the function 10 from the source program 10.
The portion b is cut out to obtain the next-stage program 12a.
Here, since the source program 10 is described in LISP, only the function is cut out and the processing related to the interface with other functions is performed, but the conversion is not performed. In step 04 (S04), the first compiler 1
1 cuts out the portion of the function 10c from the source program 10 and converts it into FORTRAN to obtain the second next-stage program 12b. In step 05 (S05), the first compiler 11 cuts out the functions 10a and 10d of the source program 10 and converts them into C to obtain the next-stage program 12c. In step 05 (S05), it is determined whether the processing has been completed for all functions.

The next-stage programs 12a to 12c obtained by the above operation are respectively LISP compiler 13
a, the FORTRAN compiler 13b, and the C compiler 13c are converted into machine language programs 14a to 14c. From the above, each function 10 of the source program 10
The next-stage programs 12a, b, c and machine language programs a, b, c for the first computer 2 optimized for a to d are obtained. Here, the specifications of the high-level programming languages, LISP, FORTRAN and C used in this embodiment are completely compliant with the standard specifications in the field of information processing. If necessary, the next-stage program generated by the program editing method of this embodiment can be easily ported to another computer.

The second embodiment will be described below. Figure 7
FIG. 3 is a diagram in which the essential points of the computer network of FIG. 2 to which this embodiment is applied are extracted. The second computer 3 corresponds to the EU in FIG. The second computer 3 is used as a dedicated computer for performing symbol processing at high speed. Second calculator 3
For simplification of description, it is assumed that the configuration of 1 is the same as that of the first computer 2. RM I / F30 and MM
The I / F 31 is an interface for connecting the first computer 2 and the second computer 3 and transmitting / receiving various information. The first computer 2 and the second computer 3 transmit and receive necessary information via either of them.

The first computer 2 and the second computer 3 are
It is assumed that each has a computer environment for carrying out the program editing method according to the present invention shown in the first embodiment. However, the first compiler 1 of the second computer 3
1. LISP compiler 13a and FORTRAN
The compiler 13b specialized for the second computer 3 corresponds to the enhanced and unique instruction. In each of the above computers, the source program 10 developed for the first computer 2 and the source program 10 developed for the second computer 3 are machine language programs 14a and 14b in the same manner as in the first embodiment. , C. As described above, the most optimized machine language programs 14a, 14b, 14c can be obtained according to the type of computer.

The third embodiment will be described below. The present embodiment provides a so-called cross compilation environment. FIG. 8 is a diagram showing a program conversion method of the program editing method of this embodiment. The operating computer environment of the program editing method of this embodiment is provided in the first computer 2 of FIG. The source program 10 is written in LISP, and functions that operate on the first computer 2 and the second computer 3 coexist. For each of the functions, it is set which of the above computers operates in a fixed format. It is assumed that The first compiler 11a is the source program 1
Based on the above setting of 0, each part of the source program 10 is divided into the first computer 2 and the second computer 3. The source programs 10a and 10b are respectively the first computer 2
, A source program 1 divided for the second computer 3
It is 0. The first compilers 11a and 11b are first compilers 11 specialized for the first computer 2 and the second computer 3, respectively. It is assumed that the first compiler 11b corresponds to the instructions of each programming language reinforced for the second computer 3. The next-stage programs 12a, 12b, 12c are next-stage programs for the first computer 2. The next-stage programs 12d, e, f are the next-stage programs for the second computer 3. The second compilers 13a, 13b, 13c are compilers for LISP, FORTRAN, and C for the first computer 2. Second compiler 13d, e, f
For LISP for the second computer 3, FORTRAN
And C compilers. These are the second computer 3
Shall correspond to the enhanced instructions for.

Next, the operation will be described. Source program 1
0 is the first computer 2 by the first compiler 11b.
And the second computer 3 are divided into respective parts. At this time, the first compiler 11b divides each function according to the above-mentioned distinction inserted when the source program 10 is created. As a result, the source programs 10a and 10b are generated. With respect to the source programs 10a and 10b, final machine language programs 14a to 14f are generated by the same procedure as in the first embodiment. Of these, the machine language programs 14d to f are for the second computer 3. The machine language programs 14d to 14f are moved to the second computer 3 and operate. In the present embodiment, the above distinction is inserted into the source program 10 in advance, but the first compiler 11a may be configured to automatically distinguish. Alternatively, the source programs 10a and 10b may be separately developed from the beginning.

The fourth embodiment will be described below. In FIG. 8, the first computer 2 includes the second compilers 13d to 13f.
The program editing method except for the computer environment is prepared, and only the second compilers 13d to 13f are prepared for the second computer 3.
In the first computer 2, machine language programs 14a, 14b, 14c for the first computer 2 are created in the same manner as in the first embodiment. Further, the next-stage programs 12d to 12f for the second computer 3 are created by the same method as in the third embodiment, and transferred to the second computer 3. On the second computer 3, the next-stage programs 12d-f are applied to the second compilers 13d-f to create final machine language programs 14d-f.

The program editing method of the present invention is not limited to the above-mentioned embodiment, and various other configurations can be adopted. Further, the above computer environment is an example. In particular, the number and types of computers, the programming language used in the next-stage program 12 generated by the first compiler 11, and the programming language of the source program 10 are not limited to those described here.

[0024]

As described above, according to the program editing method of the present invention, a source program written in one kind of high-level programming language has an editing function of a plurality of high-level programming languages, The processing content of each part was analyzed, and the optimum next-stage programming language was selected for that part from among the above-mentioned programming languages according to the processing content, and the source program was selected for each part. By converting to a program language, it is possible to select an optimum high-level programming language according to the processing content, and to provide a program editing method that realizes portability between computers of the converted program. In addition, since the conversion processing according to the processing content is performed according to the type of computer to be applied, it has higher functions and performance than the standard high-level programming language to machine language conversion compiler. Provide a program editing method that can be used by a compiler created specifically for a computer.

Further, in a computer network composed of a plurality of computers and using different computers for each processing content, by creating a program described in the programming language of the next stage corresponding to each computer, Provided is a program editing method capable of obtaining an optimum object program for the entire computer network. Further, in one computer, analyze the processing contents for each part of the source program, and select the optimum next-stage programming language for that part from among the plurality of programming languages according to the processing contents, For each of the parts, the source program is converted into a selected programming language, and in another computer, by further converting the converted program into a machine language,
There is provided a program editing method capable of effectively utilizing the computer environment of the entire computer network.

[Brief description of drawings]

FIG. 1 is a diagram showing a connection form of a computer network to which the present invention is applied.

FIG. 2 is a configuration diagram of a computer network to which the present invention is applied.

FIG. 3 is a configuration diagram of a computer to which the present invention is applied.

FIG. 4 is a diagram showing a program conversion method of the program editing method of the present invention.

FIG. 5 is a flowchart of processing of a program conversion method of the program editing method of the present invention.

FIG. 6 is a diagram showing a configuration of a source program.

FIG. 7 is a diagram showing a configuration of a computer network according to a third embodiment of this invention.

FIG. 8 is a diagram showing a program conversion method of a program editing method according to a third embodiment of the present invention.

FIG. 9 is a diagram showing a conventional program editing method.

[Explanation of symbols]

1 ... Program editing method of the present invention 2 ... First computer to which the present invention is applied 3 ... Second computer to which the present invention is applied 10 ... Source program 11 ... First Compiler 12 ... Next-stage high-level programming language program 13 ... Second compiler 14 ... Final machine language program 20 ... CPU 21 ... Bus 22 ... Memory 23 ... I / O 24 ... Display device 25 ... Communication interface 26 ... Magnetic storage device 30 ... RM I / F 31 ... MM I / F

Claims (4)

[Claims] 1. A source program written in one type of high-level programming language, having a plurality of high-level programming language editing functions, analyzing the processing contents of each part of the source program,
According to the processing content, the optimum next-stage programming language is selected for the part from the plurality of programming languages, and the source program is converted into the selected programming language for each part. A program editing method characterized by converting a converted program into a machine language. 2. The program editing method according to claim 1, wherein the conversion processing according to the processing content is performed according to the type of computer to which the processing is applied. 3. The program editing method according to claim 2, wherein the computer network is composed of a plurality of computers and uses different computers for each of the processing contents, in the programming language of the next stage corresponding to each computer. A program editing method characterized by creating a written program. 4. The program editing method according to claim 3, wherein the processing content of each part of the source program is analyzed in one computer, and the part of the plurality of programming languages is selected according to the processing content. For each part, the source program is converted into the selected programming language, and in another computer, the converted program is converted into a machine language. Characterized program editing method.