JPS58194188A - Accessing mechanism of random file - Google Patents

Abstract

PURPOSE:To decrease the time of execution of a program, by collecting and storing access information to a random file, and reducing the number of accessing to the random file with an optimum access pattern decided on the basis of the access information at accessing. CONSTITUTION:If the content of an accessed page does not exist in a page memory 2, the information of FAIT5 is received from an observer 6, the content of a page having the lowest probability of use in future among pages existing in the page memory 2 is returned to the random file, the content of a record number is transferred from the random file to the area from which the content is returned and also to a buffer. Since the access information is fed back to a file control program 7 via the observer 6, the accessing is done according to the access pattern in matching with the program. The observer 6 is operated independently without any relation to the file control program 7, collects the access information successively and processes the information and stores it in the FAIT5. The access information stored in the FAIT5 is stored just after the use of file and loaded to the FAIT5 at the start of use.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention This invention relates to a random file access mechanism in a computer system.

Figure 1 is an explanatory diagram showing the conventional random file access mechanism. In the figure, (1) is the main memory, (2) is the page memory, and (3) is the page memory management system. A management table for (41 is a random file).

Random file (4: Divide based on the concept of all pages,
The main memory is allocated an area capable of storing lt/c several pages to several tens of pages as a page memo 21, and has a management table (3) for managing all of this page memory (2). The management table 131 is
The page number of the page stored in page memory (2), the page memory address, the last accessed page,
Contains information about the next page you will access.

Operation of the prior art When a random file (4) is accessed in the form of a record number, a page number is calculated from the record number and the management table (41 is searched.
2) If the required page exists, the page memory (
2) Transfer the contents of the record number to the buffer. If it does not exist in the page memory (2), it searches for an area that has not been written to, transfers the contents of the record number from the random file (4) to the area of the page memory (2) that has not been written, and then buffers it. Transfer to. The search for xIIa that has not been written to is performed in the order of the oldest accessed pages. If page memory demon 2
: If there is no page that has not written to
The contents of the oldest accessed page are returned to the random file, and the contents of the record number are transferred from the random file (41) to the resulting blank area of the page memory (2).

Disadvantages of the prior art Since the conventional undam file access mechanism is configured as described above, the main memory (II's page memory (
2) When a page that does not exist is accessed, the content of the oldest accessed page is returned to the random file (41) and then the new page content is loaded. The disadvantage is that the number of transfers increases.

Purpose of the present invention This invention was made in order to eliminate the drawbacks of the conventional ones as described above.Access information to a random file is collected and stored, and the access information is determined based on the access information at the time of access. The purpose of this invention is to provide a random file access mechanism that can minimize the number of accesses to a random file using a well-defined access pattern.

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will now be described with reference to the drawings. Second
The figure is an explanatory diagram showing one embodiment of the present invention.
) is a file access information table (hereinafter abbreviated as FAIT) that stores the collected access information V&, 16
+ is an observer that collects access information and stores it in the FAIT 151 and provides necessary information to the file control program (7), and (7) is the file control program.

When the random file 141 of the present invention is accessed in the form of a record number, a page number is calculated from the record number and the management table +41 is searched. If the necessary page exists in the page memory (2), it is the same as before.

If the contents of the accessed page do not exist in the page memory (21, #'i, observer 16) FAIT1
51 information, return the contents of the page that has the lowest probability of being used in the future among the pages existing in page memory (2) to the random file, transfer the contents of the record number from the random file to the subsequent area, and transfer the contents of the record number to the buffer. Transfer to.

Since the access information is fed back to the file control program (7) via the observer (6), access is performed according to an access pattern suitable for the program.

The observer (61) operates independently without being related to the file control program (7) so as not to affect the system overhead, and constantly collects access information, processes it, and stores it in the FAITt51. The access information stored in this FAITI51 is stored at the beginning after the file is used, and is stored in the FAITI51 at the beginning of use.
Loaded into AIT15+.

Other Embodiments of the Invention In the above embodiments, the case of accessing a large-capacity file has been described, but the same effect can be obtained even if it is applied to the case of accessing the main memory. In this case, it is sufficient to replace the main memory 1211 with a cache memory, the file control program (71) with the main memory control program, and the random file (4; with the main memory).

Effects of the Present Invention As described above, according to the present invention, the number of accesses to random files can be reduced by using an access pattern that matches the user program, so the execution time of the program can be shortened, and the overhead of the entire system can be reduced. This has the effect of reducing system throughput and improving system throughput.

[Brief explanation of the drawing]

FIG. 1 is an explanatory diagram showing a conventional random file access mechanism, and FIG. 2 is an explanatory diagram showing an embodiment of the present invention. In the figure, (1) is main memory, (2) is page memory, (
31 is a management table, (41 is a random file, (5
) is a file access information table, ;6) is an observer, and (7) is a file control program. Note that the same reference numerals in each figure indicate the same or corresponding parts. Agent Shin Kuzuno −

Claims (1)

[Claims] An area of the main memory that can store the entire contents of several to several dozen pages of each page of a random file divided according to the page concept, and an area of the main memory that is used to manage this area. A management table containing the page number and page memory address of the page stored in the above area, the last accessed page and the next accessed page, and the access information to the above random file are collected and stored. A random file access mechanism comprising means for determining an optimal access pattern based on the above access information at the time of frequent access.