JP2990286B2 - Computer system in multi-thread environment - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a user interface of a computer system that performs multiprogramming, and more particularly, to a multithread suitable for collectively controlling mutually related processes or threads. The present invention relates to a computer system and a method for controlling a multi-thread execution environment.

[Conventional technology]

A representative example of an operating system that realizes a multi-process environment is UNIX developed by AT & T Corporation in the United States. In UNIX, the concept of a process group is provided as shown below.

a. A process called a process group leader exists in the process group.

b. When the parent process creates a child process, the child process automatically enters the same process group as the parent process.

c. All processes can create a process group with their own process as the process group leader and exit from the original process group.

d. The process group is used to terminate all processes belonging to the process group when the process group leader terminates.

Therefore, in UNIX, a variable indicating which process group the own process belongs to is provided in a process management table for managing each process. The process ID of the process group leader is set as a variable of each process, and indicates that the process belongs to the process group. The process group leader is a process in which the own process ID of the value of the variable matches.

When the process group leader is terminated, all process management tables existing in the system are checked, and if the value of the variable matches the process ID of the process group leader, the corresponding process is terminated.
The above d. Is realized.

Mach at Carnegie Mellon University in the United States is a representative example of an operating system that implements a multi-thread environment that introduces threads that execute in parallel in a process.
And Sun Microsystems SunOS.

In each case, threads can be controlled collectively in units called processes. However, functions for grouping processes and stopping or restarting processes belonging to the group for the relevant process group, and grouping threads inside processes There is no function to classify threads and a function to group threads across processes.

Examples of this kind of conventional technology include: (1) MJBACH: The Design of the UNIX Operating System
stem (PHI Publishing pp.210-213) (2) M.Accetta and others: Mach: A New kernel Foundatin for
UNIX Development (Proceedings of USENIX 1986 Summer Conference pp.
93-112) (3) Tomoaki Yamashita: Sun Lightweight Process Programming (UNIX MAGAZINE July 1988, pp. 90-115).

[Problems to be solved by the invention]

The process groups realized by the UNIX are realized by preparing variables indicating which process group the process belongs to in a management table of each process. Therefore, in order to perform the process on the process group, it is necessary to search the entire process management table, check whether each process belongs to the process group, and perform the process only when the process belongs.

Further, in UNIX, the existence significance of a process group is only for terminating other processes in the process group when the process group leader ends.

An object of the present invention is to enable various processes such as stopping / restarting a process belonging to a group to be executed at high speed without grouping processes and searching all process management tables. In addition, a thread that shares an execution environment within a process and implements thread grouping, and is a multi-processor with means for controlling a plurality of threads independently of the process. An object of the present invention is to provide a computer system in a thread environment and a method for controlling a multi-thread execution environment.

[Means for solving the problem]

In order to achieve the above object, the present invention provides a computer system in a multi-thread environment where a plurality of threads sharing an execution environment in a process operate in parallel in the process. A means for connecting a table to group a thread group independently of a process and for controlling a plurality of threads in the group collectively designates one group and waits for all threads belonging to the group. Means for making a state,
A multi-unit comprising: means for designating one group and making all threads belonging to the group executable, and means for designating one group and terminating all threads belonging to the group. We propose a threaded computer system.

The means for controlling the plurality of threads collectively includes means for creating a new group, designating one group, removing all threads belonging to the group from the group, and deleting the group. Means and 1
Means for designating one group and one thread and registering the thread in the group;
There may be provided means for designating one thread and removing the thread from the group, and means for designating one group and one thread and examining whether the thread exists in the group.

The present invention also provides a computer system in a multi-thread environment in which a plurality of threads sharing an execution environment in a process operate in parallel in the process to enable thread grouping and synchronous message communication. Means for connecting the management tables of the mutually related threads to group the thread group independently of the process, and for controlling a plurality of threads in the group collectively,
If one group is specified and all the threads belonging to the group are waiting for a message to be received from the own thread, the message is notified to all the threads in the group and any of the threads belonging to the group is If the thread is not in a state of waiting for a message from its own thread, the thread is put into a waiting state until all threads in the group issue a request for receiving a message from the own thread. We propose a multi-threaded computer system with a means for notifying a thread of a message.

The present invention further provides a multi-threaded computer system in which a plurality of threads sharing an execution environment in a process operate in parallel in the process to enable grouping of threads and synchronous message communication. Connecting the management tables of each of the interrelated threads to group the threads independently of the process,
The means for collectively controlling a plurality of threads in the group designates one group, and if all threads belonging to the group are in a state of waiting for a message to be received from the own thread, all the threads in the group are Message to the other thread, and if any thread belonging to the group is not in a state of waiting for a message from its own thread, until all threads in the group issue a request to receive a message from its own thread A means for notifying a thread to the own thread, and then notifying a message to all threads in the group; and specifying one group and allowing any thread belonging to the group to receive a message from the own thread. If so, send a message to the thread and wish to belong to the group If message reception waiting state for a record thread may own threads until one thread in the group issues a request message received from the own thread, a state waiting for its own thread, then,
A means for notifying a message to a thread that has issued a message reception request, and a group is designated, and if any thread belonging to the group is in a state of waiting for a message to be received by its own thread, a message is transmitted from the thread. If the thread is received and none of the threads belonging to the group are in a state of waiting for a message to be sent to the own thread, the thread waits until the thread in the group issues a request to send a message to the own thread. And a means for receiving a message from the thread that has issued the message transmission request, and then a multi-thread environment computer system is proposed.

In order to achieve the above object, the present invention provides a computer system in a multi-thread environment where a plurality of threads sharing an execution environment in a process operate in parallel in the process. A method of controlling a multi-thread execution environment in which a table is connected to group a thread group independently of a process, and a plurality of threads in the group are collectively controlled.
Specifying one group and placing all the threads belonging to the group in a waiting state; specifying one group and placing all the threads belonging to the group in an executable state; Specifying a group and terminating all threads belonging to the group.

[Action]

The next process / group in the group is displayed in the group link area.
Set the address of the thread management table. The program that performs group management can search all processes / threads belonging to the group by tracing the group link area.

In the group table, the addresses of the management tables of all processes / threads belonging to the group are registered as elements. The group management program can search all the processes / threads belonging to the group by checking the elements of the group table.

The group management table manages the group itself existing in the system. That is, it is a table for managing what groups exist in the system. An entry in the group management table is associated with each group. Each entry of the group management table has the following role.

e. When the group link area is used, the address is set to the address of the leading management table in the management tables of the processes / threads belonging to the group, so that all the processes / threads belonging to the group can be traced.

f. When the group table is used, the address is set to the address of the corresponding group table, so that all the processes / threads belonging to the group can be searched.

〔Example〕

Next, an embodiment of the present invention will be described with reference to the drawings.
Hereinafter, in the embodiment, only the grouping of processes will be referred to, but the grouping of threads will have a similar configuration only by replacing processes with threads.

FIG. 1 shows an example in which a storage location (group link area) for pointing to the management table itself is provided in the process management table, and processes are grouped. In the figure, reference numeral 10 denotes a group management table, each entry
11 corresponds to each group existing in the system.
The entry in the group management table points to the first management table in the process management tables belonging to the group. 12a, 12b and 12c are process management tables, respectively. Except for the group link area described below, the contents of the process management table can be considered to be all the same as in the related art. Specifically, a process ID, a CPU usage time, and the like are stored.

A group link area 13 is newly established in each of the process management tables 12a, 12b, and 12c. The group link area connects the processes in the group to each other. Specifically, the address of the next management table among the process management tables existing in the group is set. By tracing the group link area from each entry of the group management table, all the processes existing in the group can be searched. In the example of FIG. 1, the process indicated by 12a forms the first group, and the process indicated by 12b forms the second group by setting the group link area. Reference numeral 12c denotes a process that is not pointed to by any entry in the group management table and does not belong to the group.

FIG. 2 shows the process management table 12c in FIG.
5 is an example showing a method for newly adding the process indicated by the process group into the second process group. By reconnecting the group link areas of the process management table 12b, it becomes possible to put the corresponding process into the second group.
In this example, it is placed at the last position of the group. However, if the connection of the group link area is changed, it can be placed at the head position or the center of the group.

In the first embodiment of the present invention, FIGS. 1 and 2 show the group link area as if it were only a single link (one-way connection). The method is changed to a method in which a reverse connection is added), a cyclic link (a method in which the group link area of the process management table at the last position points to the process management table at the first position), a combination of a double link and a cyclic link, and the like. It is easy. In any of these methods, the group link area only needs to prepare one or two variables (in the case of a double link) whose value is the address of the process management table.

Next, the contents of the entry 11 of the group management table 10 are shown in FIG. Each entry has a group link area 50 indicating the address of the process management table at the head of the group. In addition, a field indicating the number of processes in the group (group counter)
51, a field for setting group information such as a field (group link tail) 52 indicating the address of the process management table at the last position of the group may be provided.
Fields 51 and 52 are optional.

FIG. 4 shows an example in which an array is provided for each process group, and a value indicating a management table of processes belonging to the group is set as an element of the array, thereby realizing grouping. Reference numeral 20 denotes a group management table as in FIG. 1. Each entry 21 indicates the address of the corresponding group table. 22a and 22b are group tables. Each entry of the group tables 22a and 22b is
It is a variable that indicates the address of the process management table. In this entry, there is no information to be set other than the address of the process management table.

By pointing the address of the process management table from the entry of each group table, it is possible to group processes. In FIG. 4, the group tables 22a and 22a
2b makes the process management table the first group
12a and a second group 12b. Since the process management table 12c is not pointed to from the group table, it is a table for managing processes that do not belong to any group.

FIG. 5 shows a method in which a process 12c that does not belong to any of the groups is put into a second group. When a new entry in the group table is prepared and the entry is set to point to the address of the new process management table, the process can be put into the group.

In the second embodiment of the present invention, two or more group table entries are set so as to point to the same process management table.
It can be in more than one group.

FIG. 6 shows the contents of the entries in the group management table 20. In the entry, a group table pointer to point to the group table corresponding to each group
There are 60. In addition, similarly to FIG. 3, a group counter 51 may be provided.

A user interface constructed using the grouping means will be described. An example of a processing procedure of a function (subroutine) for realizing a group management method for simultaneously processing a plurality of processes using the concept of grouping will be described in order. In the following example, a group realized using the group link area (FIG. 1) will be described. However, when the group is realized using the group table (FIG. 4), the link connection is changed to the array connection. Other than that, the procedure is the same.

(1) Generation of a plurality of processes This is a function for generating a plurality of processes and registering them in a group. For example, a plurality of processes that perform the same or similar processing are generated at the same time, and the processing for these process groups is used as a group. FIG. 7 shows an example of the algorithm of the multiple process generation process.

Within the function, if the specified group does not exist, create the group itself. Thereafter, a specified number of processes are generated, and each process is registered in a group. The registration of the process in the group is the same as in FIG. 2 or FIG. Also, the process generation process may be considered to be the same as the conventional method.

(2) Termination of Intra-Group Process FIG. 8 shows an example of an algorithm of a function for terminating the in-group process. This function terminates all processes in the group except the calling process.

All the process management tables in the group are traced from the group link area 50 of the group management table corresponding to the specified group, and if the process management table is not for the own process, the process is terminated. If all the processes belonging to the group are traced, the processing of the function is completed. In the flowchart of FIG. 8, reference numeral 80 denotes processing for tracing the group link area and terminating the loop when the process at the final position of the group is completed. The process termination process is the same as the conventional process. However, because of the termination waiting process shown in the following (3) and (4), it is necessary to add a process for restarting the process in the termination waiting state. .

(3) Waiting for end of process in group FIG. 9 is an example of an algorithm of a function for waiting for the end of any one process in the group.

In the processing of this function, all processing is a check processing except for the processing in the waiting state indicating that the process in the group is waiting for completion. The basic operation for entering a wait state in a normal operating system is “processing of designating one virtual address and entering a wait state for that address”. As an example of this address, the address of the entry of the corresponding group in the group management table may be selected. The waiting process is registered in the system queue. In the process termination process,
As shown in (2), the process for restarting the process in the end waiting state is added.

(4) Waiting for the end of all processes in the group FIG. 10 is an example of an algorithm of a function for waiting for the end of all processes belonging to the group.

The operation is basically the same as that of FIG. This process
This process is repeated until there are no more processes in the group or only the own process exists in the group. Whether a process exists in a group depends on whether the group link area 50 of the group management table does not point to any process management table or whether the value of the group counter 51 is 0.
Can be investigated by checking whether

(5) Stopping the process in the group FIG. 11 shows an example of the algorithm of the function for stopping the process in the group. This function stops all processes in the group except the calling process.

By tracing all the process management tables in the group from the group link area 50 of the group management table corresponding to the specified group, if the process management table is not the process of the own process and the process is executable, the process is Stop. When all the processes belonging to the group are traced, the processing of the function ends. The process of stopping a process is the same as the conventional process of stopping a process (“designating one virtual address and putting the process in a wait state for that address”).

(6) Resumption of Intra-Group Process FIG. 12 shows an example of an algorithm of a function for resuming the in-group process.

All the process management tables in the group are traced from the group link area 50 of the group management table corresponding to the specified group, and if the process is in a waiting state, the process is restarted. When all the processes belonging to the group are traced, the processing of the function ends. In this function, it is not checked whether the processes in the group match the own process, because the own process executing the function is not in a waiting state. The process of resuming a process is a process of taking out a process that is in a waiting state for a specified virtual address from a queue and making it executable.

(7) Creation of group To create a group, register a new entry in the group management table and initialize the entry (setting so that the group link area does not point to any process management table, group counter, etc.). Resetting to 0). FIG. 13 shows an example of an algorithm for creating a group.

(8) Deletion of a group In the deletion of a group, before deleting an entry in the group management table, a process belonging to the group is deleted.
It is necessary to set again so that it does not belong to any group. FIG. 14 shows an example of the algorithm of this processing.

The management table of the process belonging to the group is traced from the group link area 50 of the group management table, and the group link area of each management table is reset.

(9) Registering a process in a group FIG. 15 shows a method of registering one process in a group.

The illustrated method is a method in which, when grouping is performed using a group link area, a process to be registered is placed at the last position of the group. In this method, the process is set so that the process management table at the final position of the group is reached by tracing the group link area, and the group link area of the process management table at the final position indicates the address of the management table of the newly registered process. Note that this is a method of changing from the state of FIG. 1 to the state of FIG.

(10) Removal of a process from a group FIG. 16 shows an algorithm for removing one process from a group.

The processes belonging to the group are traced from the group link area 50 of the group management table, and when the specified process is found, the process management tables located before and after the specified process are connected to remove the management table of the specified process from the group. Thereafter, the group link area of the process removed from the group needs to be reset.

(11) Process existence check FIG. 17 shows a method for checking whether or not a process belongs to a specified group.

This means that the process management table belonging to the group is traced using the group link area, and if there is a management table of the designated process, a value indicating YES only needs to be returned.

(12) Message Broadcast to Group FIG. 18 shows an algorithm of a function for notifying a message to all processes in the group. For example, processes that perform the same processing are grouped and put together, and used in a form such as "Simultaneously notify a message to a group and simultaneously start processes belonging to the group".

In the process of FIG. 18, in a loop 180, a process of waiting until all processes of the group (other processes than the own process belong to the group) waits for a message to be received from the own process. When all processes in the group are waiting to receive a message, a loop
The control shifts to 181 to transfer a message to each process in the group and restart the process waiting for reception.

(13) Message transmission to group FIG. 19 shows an algorithm of a function for notifying any process in the group of a message.
For example, processes that perform the same processing are grouped and put together, and "Message is notified to the group,
Start processes that currently have no work among the processes belonging to the group. "

In the process shown in Fig. 19, it checks whether any of the processes in the group is waiting to receive a message from its own process. I do. If all the processes in the group are not in the message waiting state, they themselves enter the message transmission waiting state. In this case, when any of the processes in the group receives a message, the state is changed from the transmission waiting state to the executable state.

(14) Message reception from group FIG. 20 shows an algorithm of a function for receiving a message from any process in the group. For example, processes that perform the same processing are grouped together and used in a form such as "Wait for a message from a group, and start processing from the message of the process that has completed the earliest of the processes belonging to the group." .

In the process of FIG. 20, it is checked whether any of the processes in the group is waiting for a message to be sent to the own process.
Put the process in an executable state and wait for a message notification. During this time, the receiving process is in a reception waiting state, and waits for the end of the message transfer. If all the processes in the group are not in the message transmission waiting state, they themselves enter the message reception waiting state. In any case, when any of the processes in the group has finished transferring the message, the state is changed from the waiting state to the executable state.

By using the function (subroutine) that supports the process grouping process described above, the user can group processes that perform the same process / similar process, processes that handle the same control target, and the like as a group.
If control is performed on a group basis without controlling individual processes, process control can be easily realized.

Hereinafter, how to use the basic control function of the grouping will be described with reference to the configuration of the graphic system as an example.

FIG. 21 shows an example of the configuration of an ordinary graphic system. In this example, a process is selected as a unit for executing a system program. However, it is also possible to replace the process with a thread to increase the speed.
Threads share the execution environment inside one process and operate in parallel, so there is no need to switch the execution environment,
This is because the thread switching time is shorter than the process switching time.

When a user gives a drawing command from an input device such as the keyboard 210, the command processing process 21
1 interprets the command and activates the graphics system. In the graphic system, first, the graphic data conversion processes 212a, 212b, 212c are started. Each graphic data conversion process converts the data required for drawing into a graphic database.
213 and converted into a rendering primitive 214 which is data that can be easily processed by a computer. In the configuration example of FIG. 21, each graphic data conversion process is associated with graphic data such as a sphere, a cube, and a cone. The drawing primitives are, for example, line data (positions of start point / end point), plane data (positions of vertices in the case of triangle approximation or quadrangle approximation), and the like. This exists in various ways depending on the type of drawing algorithm.

Image primitives are coordinate transformation processes 215a, 215b, 215c
Is converted into coordinate data 216. X coordinate, Y coordinate,
Since the Z coordinate can be calculated independently, it is calculated by a separate process. If the computer system that operates the graphics system is a multiprocessor computer and can execute a plurality of processes at the same time, the speed can be increased. Finally, the brightness calculation process 21
7 performs a brightness calculation or the like on the coordinate data, and generates display data 219 that can be displayed on the CRT display 218.

Next, how the process grouping control is used for the above graphic system will be described.
FIG. 22 shows an example of a configuration when the grouping control of the present invention is used in a graphic system.

In this embodiment, three process groups are introduced. The first process group is a graphic data conversion group 220, which collectively controls graphic data conversion processes 212a, 212b, and 212c corresponding to various data. Each graphic data conversion process is individually associated with graphic data, but can be started from the command processing process 211 by a request process for one group (the request may be made by message broadcasting to the group).

The second process group is a coordinate conversion group 221. For each drawing primitive, be sure to set the X coordinate, Y
Three transformations of coordinates and Z coordinates must be performed. By grouping the processes for performing these coordinate conversions as a group, the entire group can be regarded as a processing unit for performing three-dimensional coordinate conversion without being aware of the existence of individual processes. The drawing primitive can be notified by message broadcasting to the group, and the coordinate data can be received by message reception from the group.

The third process group is a luminance calculation group 222. In the configuration example shown in FIG. 21, there is only one luminance calculation process. However, the brightness calculation (including the boundary line smoothing process, the shading process, and the like) has a large load, and when this process is performed by one process, this process may be a bottleneck in the system. For this reason, in implementing a graphic system with a multiprocessor computer, a method of generating a plurality of luminance calculation processes and operating in parallel using a plurality of processors is often adopted. At this time,
When a process is requested to the luminance calculation process, a means of checking the state of each process and requesting a process that is not executing the process to perform a new process is used. If the luminance calculation processes for performing the same processing are grouped by using the group control, this trouble can be omitted. More specifically, if a message transmission to a group is used, a library function for performing group control or a process that is not busy on the system call side is selected, and a process request is notified to the process.

As described above, by using the process group, it is not necessary to describe a step for searching for an empty process, and the like, so that it is possible to save trouble such as program creation and debugging. The process grouping method is
In addition, it can be used in various application systems.

Although the embodiment has been described in detail for the process grouping control, the same applies to the thread grouping control. However, when using thread grouping control, grouping as shown in FIG. 23 is also possible. That is, threads can be grouped across processes. As a result, a method of grouping threads as a process and a method of grouping threads by group coexist, and finer control becomes possible. Examples of grouping threads across processes include grouping threads that communicate with each other between processes,
Alternatively, there is a method of performing exclusive control by grouping threads that handle the same control target, such as a thread that outputs an error message.

〔The invention's effect〕

According to the grouping means of the present invention, it is possible to control a plurality of processes / threads that perform the same processing or similar processing, or a plurality of processes / threads that handle the same control target. By controlling a plurality of processes / threads as one group, parallel processing and exclusive control can be easily realized. Therefore, the user interface is improved by the grouping process / thread management method of the present invention.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration of a first embodiment of a process grouping means according to the present invention, FIG. 2 is a diagram showing a method of registering a process in a group in the first embodiment, and FIG. 3 is a first embodiment. FIG. 4 is a diagram showing a configuration of a second embodiment of process grouping of the present invention, and FIG. 5 is a diagram showing a method of registering a process in a group in the second embodiment. FIG. 6 is a diagram showing the configuration of a group management table in the second embodiment, FIG. 7 is a diagram showing an algorithm for generating a plurality of processes and registering them in a group, and FIG. FIG. 9, FIG. 9 is a diagram showing an algorithm for waiting for the termination of one process in the group, FIG. 10 is a diagram showing an algorithm for waiting for termination of all processes in the group, and FIG. FIG. 12 is a diagram showing an algorithm of a process of stopping a process in a group, FIG. 12 is a diagram showing an algorithm of a process of restarting a process in a group, FIG. 13 is a diagram showing an algorithm of a process of generating a new group, and FIG. FIG. 15 is a diagram showing an algorithm for performing deletion, FIG. 15 is a diagram showing an algorithm for performing a process of adding a process to a group, FIG.
FIG. 16 is a diagram showing an algorithm for removing a process from a group, FIG. 17 is a diagram showing an algorithm for checking whether a process exists in a group, and FIG. 18 is a message to all processes in the group. FIG. 19 is a diagram showing an algorithm for transmitting a message to one process in a group, FIG. 20 is a diagram showing an algorithm for receiving a message from one process in a group, FIG. FIG. 1 is a diagram showing an example of a configuration of a conventional graphic system, FIG. 22 is a diagram showing an example of a configuration in which a process grouping unit of the present invention is introduced into a graphic system, and FIG. 23 is a diagram showing a concept of thread grouping. It is. 10, 20 ... group management table, 11, 21 ... one entry of group management table, 12a, 12b, 12c ... process management table, 13,50 ... group link area, 22a, 22b ... group table, 51 …… Group counter, 52 …… Group link tail, 60 …… Group table pointer, 210 …… Keyboard, 211 …… Command processing, 212 …… Graph data conversion, 213 …… Graph database, 214 …… Drawing primitive, 215 …… Coordinate transformation, 216 ……
Coordinate data, 217: luminance calculation, 218: CRT, 219: display data, 220: graphic data conversion group, 221: coordinate conversion group, 222: luminance calculation group.

──────────────────────────────────────────────────続 き Continuing from the front page (72) Inventor Yoshiki Kobayashi 4026 Kuji-cho, Hitachi City, Ibaraki Prefecture Inside Hitachi, Ltd. Hitachi Research Laboratory Co., Ltd. (72) Inventor Tomoaki Nakamura 5-2-1 Omikacho, Hitachi City, Ibaraki Prefecture Co., Ltd. (56) References JP-A-1-316830 (JP, A) JP-A-58-78238 (JP, A) JP-A-63-298534 (JP, A) JP-A-62-11941 (JP) JP, A) JP-A-3-148730 (JP, A) JP-A-63-4339 (JP, A) JP-A-1-236329 (JP, A) Proc. Of the 39th Annual Conference of IPSJ (▲ II ▼), pp. 1193-1198 (Nagoya Hira) Transactions of Information Processing Society of Japan, Vol. 31, No. 2 (February 1990) pp. 275-287 bits November 1988 issue (Kyoritsu Shuppan Co., Ltd.), pp. 75-82 bit December 1988 issue (Kyoritsu Publishing Co., Ltd.), pp. 68-79 (58) Field surveyed (Int. Cl. ⁶ , DB name) G06F 9 / 46,15 / 16 JOIS (JICST file) CSDB (Japan Patent Office)

Claims (5)

(57) [Claims] 1. A multi-threaded computer system in which a plurality of threads sharing an execution environment in a process operate in parallel in the process, wherein a thread group is formed by connecting management tables of mutually related threads. Means for independently grouping a plurality of threads in the group, and means for controlling a plurality of threads in the group collectively; means for designating one group and setting all threads belonging to the group to a waiting state; Means for designating one group and making all threads belonging to the group executable, and means for designating one group and terminating all threads belonging to the group. Characteristic multi-threaded computer system. 2. The computer system according to claim 1, wherein said means for controlling said plurality of threads collectively includes means for creating a new group, and specifying one group, and Means for removing all threads belonging to the group from the group and deleting the group; means for specifying one group and one thread and registering the thread in the group; one group and one thread Means for designating a thread and removing the thread from the group; means for designating one group and one thread and examining whether the thread exists in the group. Multi-threaded computer system. 3. A computer system in a multi-thread environment in which a plurality of threads sharing an execution environment in a process operate in parallel in the process to enable thread grouping and synchronous message communication. Means for connecting the management tables of the threads related to, grouping the thread group independently of the process, and controlling a plurality of threads in the group collectively, designates one group, and assigns the group to the group. If all the belonging threads are in a state of waiting for a message from the own thread, a message is notified to all the threads in the group, and any thread belonging to the group waits for a message from the own thread. If not, all threads in the group receive messages from their own thread Until issues a request, and a state waiting for its own thread, then computer system of a multi-threaded environment, characterized in that it comprises means for notifying a message to all threads in the group. 4. A computer system in a multi-thread environment in which a plurality of threads sharing an execution environment in a process operate in parallel in the process to enable thread grouping and synchronous message communication. Means for connecting the management tables of the mutually related threads to group the thread group independently of the process, and for controlling a plurality of threads in the group collectively, designates one group, and specifies the group. If all the threads belonging to the group are in a state of waiting for a message from the own thread, the thread notifies all the threads in the group, and any thread belonging to the group receives a message from the own thread. If not waiting, all threads in the group will receive messages from their thread Until a request for communication is made, the own thread is kept in a waiting state, and then, a means for notifying a message to all threads in the group, and one group is designated, and any thread belonging to the group is assigned a self-thread. If the thread is waiting to receive a message from the thread, the thread is notified of the message,
If none of the threads belonging to the group are in a waiting state for receiving a message from the own thread, the thread is set to a waiting state until any thread in the group issues a request for receiving a message from the own thread, Then, a means for notifying the message to the thread that has issued the message reception request, and specifying one group, and if any thread belonging to the group is in a state of waiting for a message to be received by its own thread, the corresponding thread Accept the message,
If none of the threads belonging to the group are in a state of waiting for a message to be sent to the own thread, the thread is set to a waiting state until any thread in the group issues a request to send a message to the own thread, And a means for receiving a message from a thread that has issued a message transmission request. 5. In a multi-thread environment computer system in which a plurality of threads sharing an execution environment in a process operate in parallel in the process, a thread group is formed by connecting respective management tables of mutually related threads. Is a method for controlling a multi-thread execution environment in which a plurality of threads in a group are controlled independently of a process, and a plurality of threads in the group are collectively controlled. Waiting step, specifying one group, and setting all threads belonging to the group to an executable state, specifying one group, and terminating all threads belonging to the group And controlling the multi-thread execution environment.