JPH04178731A - Reuse system for method cache information in object oriented processing system - Google Patents

Abstract

PURPOSE:To always process an application by the highest and stable performance when the application on an object oriented processing system is executed for plural times by providing a cache information storing function and a storage cache information read-in function, and effectively using cache information generated once even in the next session. CONSTITUTION:When the application program 30 of a user is executed on the object oriented processing system 10, the cache information is accumulated one after another in a cache table 11b. The cache table 11b consists of a class address table 11b1 and a method address table 11b2, and is formed in such structure that a correspondent method can be taken out setting a class address as a key. Address information with respect to class and method in a generated cache table 11b, etc., is stored in an external memory device 20 via a cache information input/output mechanism 11c. Thereby, the cache information generated once can be reused, and performance in which the effect of method cache can be displayed to a maximum can be obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a method processing method in a language processing system having an object-oriented function.

[Conventional technology]

In an object-oriented processing system, a method or a combination of methods that matches the class (type) of a given actual argument is executed from a defined group of methods having the same name. The method or combination of methods that is executed is called an effective method. The process of generating an effective method according to the type of the given actual argument is called method dispatch. The effective method is dynamically determined by the given actual arguments, and especially in processing systems that allow combinations of methods called method combinations as effective methods, the combination must be calculated dynamically.
The overhead of method dispatch is large. Since this overhead directly affects the performance of the object-oriented processing system, how to reduce this overhead is the most important issue in terms of the performance of the object-oriented processing system.

In recent years, a technique called method caching has become popular as a means of speeding up method dispatch. This uses a "cache table" in which effective methods are registered using the class (type) of the given actual argument as a key. If it exists, the registered effective method is used.

(b) If the class of the given actual argument does not exist in the cache table, calculate the effective method and register it in the cache table. The calculated method is used as the effective method.

Therefore, if you use the method cache method, immediately after the processing system starts up and until a certain amount of information is set in the cache table, it will be the same as when no countermeasures are taken, but after that, the performance will be significantly improved. expected. (
Regarding only the dispatch section, the information registered in the cache table will be simply referred to as cache information.

Note that related devices of this type include, for example, CO
P SLA '86 Proceeding
s page 1-8 ” Object -Orient
edProgrammingWithFla
vors”, etc.

[Problem to be solved by the invention]

The above conventional technology does not consider the overhead of method dispatch until the cache information is set, and in particular, even when the same application on the processing system is executed multiple times, the cache information is sent each time. The problem was that the best performance could not be achieved until the

An object of the present invention is to ensure that when an application on an object-oriented processing system is executed multiple times, the application is always processed with the highest and stable performance (more precisely, from the second time onwards).

[Means to solve the problem]

In order to achieve the above objectives, we added a cache information input/output mechanism with a "cache information storage function" and a "stored cache information read function" to the conventional method cache unit, making it possible to reuse cache information once generated. This is what I did.

[Effect]

According to the user's designation, the cache information input/output mechanism stores all information in the cache table in the designated external storage device. Further, after restarting the processing system, the cache information input/output mechanism reads cache information stored from an external storage device and sets the information in the cache table according to a user's designation. By combining these two operations, the present method cache system allows the user to repeatedly reuse cache information that has been generated once.

〔Example〕

An embodiment of the present invention will be described below with reference to the drawings.

First, the overall configuration will be explained with reference to FIG.

This processing system consists of an object-oriented processing system main body 10 and an external storage device 20 that stores cache information.

In this example, the external storage device i20 is a direct access storage device@
(DASD), but it can also be changed to a magnetic tape device or the like according to the user's specifications. The method cache unit 11 executes method caching. Conventionally, the method cache unit 11 consists of a cache control mechanism 11a that executes method dispatch based on cache information and a cache table Ilb that stores cache information, but in the present invention, a cache information input/output mechanism (ILc) is further added. do.

When the user's application program 30 is executed on the object-oriented processing thread IO, cache information is gradually accumulated in the cache table llb. Cache table llb is class address table llb
It consists of a method address table 11b2, and has a structure in which a corresponding method can be retrieved using a class address as a key. Based on this information, the cache controller @Lla.

The method cache method performs method dispatch. In the present invention, after the user's application program 30 is executed, address information regarding classes and methods in the cache table 11b generated up to that point is transferred to the external storage device 20 via the cache information input/output mechanism 11c according to the user's specifications. △ Can be stored. Also.

When using the same application after restarting the object-oriented processing system, the cache information input/output mechanism lie reads the stored cache information from the external storage device 20 according to the user's specifications, and stores it in the cache table ll.
Set to b. Thereby, the cache table llb at the time of the previous application program depth termination can be restored as is. Note that the ``71-response information'' of certain data used here is stored in an external storage device,
In the next session, the algorithm for reading the stored address information etc. from the external storage device and restoring the original data is the same as the algorithm for the suspend/resume function generally known in the LISP language and the like.

FIG. 2 shows a flowchart of how the present invention is used by a user. Figure 2(a) shows the flow of processing when the application program is executed for the first time, and Figure 2(b)
shows the flow of processing when the same program is executed for the second and subsequent times. After the application program is executed for the first time, storage of cache information is designated and the information of cache table llb is stored in the external storage device 20. From now on, when this application program is executed, after the program is loaded, the stored cache information is restored to the cache table Llb according to the cache information restoration designation, and then the program is executed. This allows the cache information created the first time to be effectively reused, improving program execution performance.

The expected performance improvement effect is shown graphically in Figure 3. The vertical axis is program execution performance, and the horizontal axis is .

This is the elapsed time since the program started. By reusing cache information, the present invention can obtain maximum and stable performance by method caching immediately after program startup.

〔Effect of the invention〕

According to the present invention, since cache information that has been created once can be reused, performance that takes full advantage of the method cache effect can be obtained immediately after the user starts the application program (second time or later).

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a processing system for realizing an optical function in an embodiment of the present invention, FIG. 2 is a flowchart showing a procedure for use by a user who makes the present invention effective, and FIG. 3 is a diagram of the present invention. FIG. 3 is a diagram showing the expected effects as a graph. 10. Object-oriented processing system, 11. Method cache unit.

Claims (1)

[Claims] 1. A method cache realized on a volatile storage device of an object-oriented processing system, which is characterized by providing a cache information storage function and a storage cache information read function, so that once generated cache information can be effectively used in the next session. method.