JPH05313910A - Machine word instruction generating system - Google Patents

Abstract

PURPOSE:To improve the executing speed of a system function by performing no evolution of a machine word instruction train which calls out the system function and by evolving the processing to be carried out by the function system with the machine word instruction train. CONSTITUTION:A machine word instruction generating system consists of a compiler control means 1 which controls an entire compiler, an in-line decision processing means 2, an in-line enable system function name list table 3, the in-line deciding result addition information 4, an in-line evolution processing means 5, and a machine word instruction train table 6. The means 2 compares an inputted function calling function name with the table 3 and sets the in-line unable information to the information 4 it the inputted function name is contained in the table 3. Then, the means 2 transfers its control to the syntactic meaning analyzing result output processing. As a result, an object having a high executing speed can be obtained.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a machine language instruction generation system for compiling a source program.

[0002]

2. Description of the Related Art In a conventional machine language instruction generation system, a system function is provided to a programmer in a unit of an independent compilation unit, and the system function is executed by a machine language instruction string of a function call.

[0003]

The above-mentioned conventional machine language instruction generation method requires a machine language instruction sequence associated with the call of a system function, and furthermore, within the system function, all arguments passed to the system function are used. Since it is performing processing for the system function, the judgment processing for distinguishing the processing for the argument passed to the system function must be performed. It has the drawback that it cannot generate an object that is required and has a high execution speed.

[0004]

A machine language instruction generation system of the present invention is a machine language instruction generation system of a compiler which reads a source program and outputs an object program,
Whether or not to perform inline expansion so that the processing to be performed by the system function is expanded inline with the machine language instruction string instead of the function call without expanding the machine language instruction that calls the system function provided to the programmer. In-line determination processing means for determining, in-line expandable system function name list table, in-line determination result additional information holding a determination result of whether or not to perform in-line expansion by the in-line determination processing means, and in-line determination result addition It is configured to include an inline expansion processing means for expanding an inline machine language instruction string by information and a machine language instruction string table for expanding an optimum machine language instruction string by the inline expansion processing means.

[0005]

Embodiments of the present invention will now be described with reference to the drawings. FIG. 1 is a block diagram showing an embodiment of a machine language instruction generation system of the present invention. As shown in FIG. 1, in this embodiment, compiler control means 1 for managing the entire compiler 1, inline determination processing means 2, inlineable system function name list table 3, inline determination result additional information 4, inline expansion processing means 5 are provided. , A machine language instruction sequence table 6.

FIG. 2 is a diagram showing an example of a system function name list table 3 that can be inlined. As shown in FIG. 2, inline system function name list table 3
Is a function name that can be inlined (1), (2), ...
(N) is included.

FIG. 3 is a diagram showing an example of the inline determination result additional information 4. As shown in FIG. 3, the inline determination result additional information 4 includes a determination result of whether to perform inline and is added to the conventional text information.

FIG. 4 is a diagram showing an example of the machine language instruction sequence table 6. As shown in FIG. 4, the machine language instruction sequence table 6 is composed of a system function name to be inlined, a condition number of arguments, a condition of arguments, and a machine language instruction sequence to be expanded.

FIG. 5 is a flow chart showing an example of the processing of this embodiment. As shown in FIG. 5, in this embodiment, the source input syntactic and semantic analysis process 7 reads the source program from the source program file 8. In the source input file end determination process 9, a process of determining whether or not the source program exists is performed. When the source program does not exist, the control is transferred to the syntactic and semantic analysis result input process 10. When the source program exists, the function call determination processing 11 determines whether or not the input source program is a function call. If it is a function call, the control is transferred to the inline determination processing means 2.

If it is not a function call, in the syntactic and semantic analysis result output processing 12, the syntactic and semantic analysis result file 13 is output.
Then, the result of the syntactic and semantic analysis is output, and the control is transferred to the source input syntactic and semantic analysis processing 7. The inline determination processing means 2 compares the input function name of the function call with the inlineable system function name list table 3, and as a result,
When there is an input function name in the inline-capable system function name list table 3, the inline-disabled information is set in the inline determination result additional information 4, and the control is transferred to the syntactic and semantic analysis result output process 12.

In the syntactic and semantic analysis result input processing 10, the syntactic and semantic analysis result is read from the syntactic and semantic analysis result file 13. In the syntactic / semantic analysis result file end determination process 14, a process of determining whether or not a syntactic / semantic analysis result exists, and when the syntactic / semantic analysis result exists, control is transferred to the machine language expansion process 15. In the machine language expansion processing 15,
In the inline function determination processing 16, it is determined whether the input syntactic and semantic analysis result is a function call and the inline determination result additional information 4 is inline enabled or disabled, and the function call is performed and the inline determination result additional information 4 is inline enabled. If it is information, the inline expansion processing means 5
Transfer control to.

If it is not a function call or if the inline determination result additional information 4 is a function call and inline incapable information, control is transferred to the machine language expansion output processing 17. The inline expansion processing means 5 searches the machine language instruction string table 6 for an optimum machine language instruction string based on the input function name and argument condition, and transfers control to the machine language expansion output processing 17. In the machine language expansion output process 17, after outputting the machine language command to the machine language expansion file 18, the control is returned to the syntactic and semantic analysis result file input process 10. Syntax semantic analysis result file judgment processing 14
Then, when the syntactic and semantic analysis result does not exist, the processing ends.

FIG. 6 is a flow chart showing an example of processing of the inline determination processing means 2. As shown in FIG. 6, the inline determination processing means 2 performs processing 2-1 for obtaining inlineable system function names from the inlineable system function name list table 3. Process 2-2 for the result
Then, if there is a system function name that can be inlined, the process of determining whether it matches the input function name 2-
3 is performed, and if they match, a determination 2-4 is made as to whether or not further inline expansion should be performed. When inline expansion is performed, inline determination result addition information 4 is set to inline enabled information, and inline determination result addition is performed. The information 4 is added to the conventional text (2-5), and the process ends.

When the inline expansion is not performed, the inline determination result additional information 4 is set to the inline-disabled information, the inline determination result additional information 4 is added to the conventional text (2-6), and the process is terminated. If the input function name does not match, the control is returned to the process 2-1 for obtaining the inline expandable system function name from the inline expandable system function name list table 3. When the system function name that can be expanded inline does not exist, the inline determination result additional information 4 is set to the information that cannot be inlined, the inline determination result additional information 4 is added to the conventional text, and the process ends. .

FIG. 7 is a flow chart showing an example of the processing of the inline expansion processing means 5. Inline expansion processing means 5
Performs a process 5-1 for obtaining a system function name for inline expansion from the machine language instruction sequence table 6. Based on the result, a judgment 5-2 is made as to whether or not the system function name to be inlined and the input function name match. If they do not match, the process returns to 5-1 and the machine language instruction sequence table 6 indicates that the next inline After obtaining the name of the system function for performing the above, the control is returned to 5-2.

When they match, the processing 5-3 for obtaining the condition of the argument from the machine language instruction sequence table 6 is performed, and the judgment 5-4 of whether or not it matches the condition of the input argument is performed. When the condition of the argument for expanding the instruction sequence matches, the processing is terminated while holding the information of the machine language instruction sequence. When the input argument and the condition for expanding the machine language instruction sequence do not match, during the number of conditions of the arguments of the machine language instruction sequence table 6, processing 5-3 for obtaining the condition of the argument from the machine language instruction sequence table 6 and The determination process 5-4 for determining whether the condition for expanding the machine language instruction sequence and the input argument match is repeated until they match.

[0017]

As described above, according to the machine language instruction generation method of the present invention, when a system function is executed, the processing to be performed by the system function is executed without expanding the machine language instruction sequence that calls the system function. The expansion by the word instruction sequence has an effect that an optimum machine language instruction sequence at the time of function call for the purpose of improving the execution speed can be generated.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of a machine language instruction generation system of the present invention.

FIG. 2 is a diagram showing an example of a system function name list table 3 that can be inlined.

FIG. 3 is a diagram showing an example of inline determination result additional information 4.

FIG. 4 is a diagram showing an example of a machine language instruction table 6.

FIG. 5 is a flowchart showing an example of processing of this embodiment.

FIG. 6 is a flowchart showing an example of processing of the inline determination processing means 2.

FIG. 7 is a flowchart showing an example of processing of the inline expansion processing means 5.

[Explanation of symbols]

 1 Compiler Control Means 2 Inline Judgment Processing Means 3 Inline System Function Name List Table 4 Inline Judgment Result Additional Information 5 Inline Expansion Processing Means 6 Machine Language Instruction Sequence Table

Claims (1)

[Claims] 1. In a machine language instruction generation method of a compiler that reads a source program and outputs a target program, processing to be performed by a system function without expanding a machine language instruction that calls a system function provided to a programmer is performed. In order to perform inline expansion with a machine language instruction sequence instead of function call, an inline determination processing unit that determines whether to perform inline expansion, a system function name list table that can be expanded inline, and an inline by the inline determination processing unit. Inline determination result additional information holding a determination result of whether or not to perform expansion, inline expansion processing means for expanding an inline machine language instruction sequence by the inline determination result additional information, optimal by the inline expansion processing means Machine language instruction sequence for expanding the machine language instruction sequence Machine language instructions generation method characterized by comprising the Buru.