JPS61202250A - Data area control system - Google Patents

Abstract

PURPOSE:To perform the designing and control of the answer time of a real time system with high accuracy by storing the number of used memory blocks and these memory capacities and using this information to decide whether the garbage collection should be carried out or not. CONSTITUTION:A computer system consists of a memory 1 for storage of programs and data, a processor 2, a keyboard 3, a display 4, a disk 6 and a disk controller 5. The disk 6 is used with allocation of a memory block of an optional size and then released when it is used. The numbers of these used memory blocks and their memory capacities are stored. Based on this storage information, it is decided whether the garbage collection should be carried out or not.

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention relates to data processing in computers, and in particular to a data area management method suitable for memory management of databases, high-level language compilers, list processing programs, text processing, and operating systems. Regarding.

[Background of the invention]

Conventional methods (Sasa Masako: ``Dynamic storage allocation is a law'' Kuroyo = ``Sequential Garbage Collection'' Hibino Hata: ``Parallel Garbage Collection'' Information Processing Vol.
, 24. A4. (see 1983), the number of used memory blocks and the sum of the amount of memory were not stored in memory, so Garpage collection (a process that organizes memory areas so that used memory blocks can be used) was performed. I couldn't decide when to do it. As a result, a large amount of processing time is required to perform garbage collection even when there is no used memory area, or to perform garbage collection after the used area has become enormous, making it impossible to respond to high-speed responses such as machine control. The sum of used memory for garbage collection has been optimized to support the high-speed response described above. The processing time for Garpage collection can be controlled by setting the size of the amount of used memory to =49.

Note that this rice data area management system does not specifically mention data that controls the trigger for executing garbage collection, as shown in the above-mentioned known examples.

[Purpose of the invention]

An object of the present invention is to determine the timing of garbage collection by holding the sum of memory amounts of used memory blocks and the number of blocks in memory in a heap-type memory management system, and to determine the processing time of garbage collection. The purpose is to make it possible to control the processing time by associating the memory amount with the sum of the memory amount of used memory blocks.

[Summary of the invention]

The present invention allocates and uses a memory block of an arbitrary size on a disk memory in a database system, and releases it when it is no longer used. As the disk memory is used, the number of used blocks increases, making it impossible to use the disk memory efficiently. Therefore, garbage collection is performed to collect unused blocks on the disk memory so that they can be used effectively. There was no way to determine when to perform garbage collection, and garbage collection was performed only after the process reached a point where the process would not proceed unless used memory blocks were used. As a result, the processing time for Garpage collection becomes enormous, resulting in inconveniences such as the inability to respond to high-speed response processing. By remembering the total amount and number of used memory blocks, it is now possible to find the appropriate time to perform garbage collection.

[Embodiments of the invention]

An embodiment of the present invention will be explained below with reference to FIG. 1st
The figure shows the configuration of a computer system implementing the present invention, which includes a memory 1 for storing programs and data, a processor 2 for retrieving instructions and data and executing operations, and a keyboard 3 for inputting commands to be executed. It consists of a display 4 that displays program execution results and input commands, a disk 6 that is a secondary storage device for storing and retrieving data and programs, and a disk control device 5 that controls it.

FIG. 2 shows how to use the storage area 8 in the disk 6, including a memory block 9 (M,).

Blocks of arbitrary length, such as io (Mb) and 11 (M-), are allocated as necessary. Area 12 indicates a storage area to be allocated when a request is made in the future.

FIG. 3 shows the storage area on the disk in block 13 (M,
) to block 17 (M, ) are allocated as memory blocks. Among these, block 14 (Mb) and block ts (Ma) indicate used blocks and have been released. Now, with memory blocks 13 (M,) to 17 (M,) allocated, if there is a request for further memory block allocation, the used memory block will be placed in area 1 in Figure 2.
Since there is no unused storage area (unallocated area) as shown in FIG. 2, the storage area is organized as shown in FIG. 4 so that used memory blocks can be used. Fourth
The figure shows the state after organizing the storage area, and shows the memory blocks 18 (M,) and 19 (M,) in use.
, 20 (M,) are moved to the upper part of the storage area, and the used memory blocks 14 (Mb), 16 (
An area corresponding to M4) can be used as area 21 in FIG. 4.

This process also changes the reference table 22 as shown in FIG. 5 for referencing memory blocks. The reference table 22 is composed of a part 23 indicating the beginning of the memory block and a part indicating the size of the block. When the memory block is moved on the storage area 25, the starting address of the memory block is changed to a new address. You should change it.

The process of organizing the storage area by focusing on used memory blocks, changing the memory block reference table, and creating a usable area like area 21 in FIG. 4 is called garpage collection.

Until now, the reason for performing garbage collection was
As shown in the figure, this was done when there was insufficient storage space to newly allocate, or every time a memory block was released. In the case of a large-capacity storage area, the method of performing garbage collection when the storage area is insufficient requires a large amount of time to process the garbage collection, and systems that require high-speed response such as device control may become unable to respond. On the other hand, if garbage collection is performed every time a memory block is released, the amount of memory blocks that can be used is small compared to the processing content and processing time, which is not efficient. Therefore, as shown in FIG. 6, the number of used memory blocks 26 and the amount of used memory -27 are stored and updated every time a memory block is released. FIG. 7 shows a flowchart for implementing this.

First, in step 28, the memory block M+ is released. In step 29, the number of used memory blocks 26 shown in Figure 11g6 is increased by one, and then in step 3o, the amount of used memory blocks is increased by the release amount. In step 31, the number of used memory blocks 26 and the amount of used memory 27 shown in FIG. 6 are determined, and when these exceed a certain value, garbage collection shown in step 32 is performed. By adjusting this value, you can control the frequency and processing time of garbage collection.

In step 33 of FIG. 7, when the garbage collection is completed, the number of used memory chips and the amount of used memory are set to O.

〔Effect of the invention〕

According to the present invention, since the timing of garbage collection and the time required for processing can be controlled, performance design and quantitative control of the response time of a real-time system can be carried out with high precision. Since the amount of usable memory can be calculated accurately, storage areas can be managed more accurately.

[Brief explanation of the drawing]

Figure 1 is a diagram showing the configuration of a computer in which the data area management method of the present invention is used, Figure 2 is a diagram showing the storage area on a disk and the method for managing that area, and Figure 3 is a diagram showing the allocation to the storage area. Figure 4 is a diagram showing the state of the storage area after the used memory blocks and in-use memory blocks in the storage area have been sorted out. Figure 5 is a diagram showing the memory blocks in the storage area. A diagram showing the relationship between reference tables, storage areas, and memory blocks. Figure 6 is a diagram showing the number of used memory blocks in the storage area and data for storing the amount of used memory. Figure 7 is a memo 'fI+ figure word. 2 Figure

Claims (1)

[Claims] Data that is characterized by storing the number of used memory blocks and the amount of memory in memory in a heap format memory usage method, and using this information to determine whether to perform garbage collection. Area management method.