JPH09237193A - Dynamic load distribution method for parallel computer - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent the quantity of load from being increased and to improve the efficiency of processing by allocating a new generation request task generated by a processor in a system to another processor at the time of judging that the task can not be processed in the processor concerned. SOLUTION: When a new generation request task is generated in a certain processor in a real time multi-processor system forcing each processor to execute an allocated task within prescribed frame time, whether the request task can be processed in the processor or not is previously judged, and at the time of judging the disablement of the processing, dynamic load distribution is executed so as to request task allocation to another processor. When the disablement of processing the request task by its own processor is judged and substitutive processing is requested to another processor, the determination of a receiver is executed among processors qualified for receivers. Consequently the load quantity of each processor can be prevented from being unnecessarily increased and the processing efficiency can be improved.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for performing dynamic load distribution in a real-time parallel computer so that all tasks generated by each processor can be processed within a predetermined frame time. It is a thing.

[0002]

2. Description of the Related Art FIG. 25 shows a conventional parallel computer system in which a plurality of processors 102 to 1mn are connected by a network 101. In order for all the processors in the system to operate efficiently, tasks that occur sequentially are sequentially assigned to these processors for distributed processing. Each processor generates load information from the processing time of the assigned task.

Conventionally, in such a parallel computer system, when a new task generation request is made in a state where each processor has a heavy load, the processor is searched for the lightest load by checking the load information of the other processors. Request the processor to allocate the task, or a processor with a heavy load randomly selects another processor and examines the load information of the selected processor to determine whether the load is lighter than itself. When the other processor has a lighter load, the processor is requested to be assigned the task, and the loads of all the processors are averaged to improve the efficiency of the entire system.

[0004]

In a conventional parallel computer as described above, when a processor with a heavy load requests a processor with the lightest load to request task allocation, the load information of the other processor is sequentially examined. It may take a lot of time in some cases because the load of the selected processor is heavier than one's own load, so it takes a lot of time to repeat the selection. In some cases, even if the other processor has a lighter load than itself, there is no difference, and the processing time required for assigning the task is longer, meaningless load distribution may be performed. However, there is a problem that the load for dynamically distributing the load increases and the processing efficiency of the original application task decreases.

The present invention has been made to solve the above problems, and in a real-time multiprocessor system in which each processor must process an assigned task within a predetermined frame time, When a new generation request task occurs in a certain processor, it is first judged whether or not the request task can be processed in the processor, and only when it is determined that the task cannot be processed, the processor becomes a sender and another processor. Dynamic load balancing is performed in order to request task allocation to. As a result, it is not necessary to always try to distribute the load, and thus it is possible to prevent the load of the processor from unnecessarily increasing. Further, since the receivers are determined by the processors requested by the sender to allocate the tasks, the load required for determining the receivers of the sender can be reduced. From the above, the purpose is to reduce the amount of load required for dynamic load distribution in the entire system and improve processing efficiency.

[0006]

According to a first aspect of the present invention, there is provided a dynamic load balancing method for a real-time parallel computer, wherein when a task generation request is issued to a certain processor, the processor determines whether or not the processor can execute the task. When it is determined that the task cannot be executed, the processor becomes a sender and selects a processor having the next processor number as a receiver candidate, and load information for the receiver candidate. Communication means for requesting the transmission of 1 to 1, and means for determining whether or not the receiver candidate is qualified as a receiver based on the received load information,
Repeat the selection and qualification of new receiver candidates as long as there is enough time to process tasks and request allocation until the number of qualified receivers who have passed the judgment reaches a certain number or until the end of the frame time. Means, a means for issuing a receiver decision command to all receiver qualified persons, a means for the receiver qualified persons to exchange mutual load information between qualified persons, and each based on the load information of all qualified persons. A means by which the receiver qualified person determines one of the least qualified receiver qualified persons, a means by which the processor decided by the receiver sends a decision report to the sender, and a sender which receives the decision report acts on behalf of the receiver by the receiver. It is achieved by providing a means for requesting.

In the real-time parallel computer dynamic load balancing method according to the second aspect of the present invention, when a task generation request is issued to a processor, the processor determines whether or not the processor can execute the task, and each processor periodically. And a means for broadcasting its own load information to another processor, and when the result of load judgment indicates that the task cannot be executed, the processor becomes a sender and is acquired in advance by the broadcasting means. A means to judge the qualification as a receiver of each processor based on the load information of each processor, a means to issue a receiver decision command to all receiver qualified people who have passed the judgment result, and a receiver qualified person Based on the means of exchanging load information between qualified persons using the broadcasting means and the information of all qualified persons, each receiver qualified person receives the least load. By providing a means for determining one qualified server, a processor for the receiver to report to the sender, and a sender for receiving the decision to request the receiver to process the task. To be achieved.

A dynamic load balancing method for a real-time parallel computer according to the third aspect of the present invention is a method for determining whether or not a processor can execute a task when a task generation request is issued to the processor, and a shared memory for each processor. And a means for acquiring load information of each processor by reading the shared memory when the processor cannot execute the task as a result of the load determination. , Means for performing qualification determination as a receiver of each processor based on the load information of each processor, and means for issuing a receiver determination command to all receiver qualified persons who have passed the determination result,
A means for the qualified receivers to exchange load information with each other using the shared memory, and based on the information of all qualified users, each qualified receiver determines the qualified receiver with the lightest load. This is achieved by providing a means for making one decision, a means for the processor decided for the receiver to make a decision report to the sender, and a means for the sender receiving the decision report asking the receiver to act on behalf of the task.

Further, in the dynamic load balancing method for a real-time parallel computer according to the fourth, fifth and sixth inventions, if the receiver cannot accept the task generation request after the receiver is determined, the receiver sends a reject signal to the sender. Means and means for newly issuing a receiver determination command to the remaining qualified receivers when the sender has sufficient time to process tasks and request allocations by the end of the frame time.

The dynamic load balancing method for a real-time parallel computer according to the seventh, eighth and ninth aspects of the invention is a means for each qualified receiver to create a list of qualified receivers and a receiver. If the receiver cannot accept the task creation request after the decision, the receiver will be newly added from the processors registered in the qualified person list as long as there is enough time to process the task and request the allocation by the end of the frame time. This is achieved by providing means for determining the receiver.

A dynamic load balancing method in a real-time parallel computer according to the tenth, eleventh and twelfth inventions is as follows:
A means for previously securing resources such as a memory and a processing time within a frame time required for the receiver candidate to generate a task for which a generation request is made when the determination result is a pass when the receiver candidate is qualified. It is achieved by providing.

[0012]

BEST MODE FOR CARRYING OUT THE INVENTION

Embodiment 1. 1 and 2 are process flow charts showing the first embodiment of the present invention. FIG. 1 is a process flow diagram when a task generation request is issued to a certain processor, and FIG. 2 is a process flow diagram of a processor which is selected as a receiver candidate by a sender and to which a receiver determination instruction is transmitted. In FIG. 1, when a task generation request occurs in a certain processor, the processor obtains the task scheduling time schedule until the end of the frame time, load information such as memory usage status, and the required memory amount of the task. Executability determination 1 of the task is performed based on the task information regarding the required processing time and the like. The processing time required for the task is longer than the time from the current time to the end time of the frame time, or another task is already assigned from the current time to the end time of the frame time, so there is room to process the task. If not, it is determined to be unexecutable, and if not, it is determined to be executable. If it is determined to be executable, task generation 2 is performed in the processor and execution is started. If it is determined to be inexecutable, the processor becomes a sender, and the sender selects the processor with the processor number next to it as a receiver candidate 3 and requests the receiver candidate to transmit load information 4 To do. In FIG. 2A, when the receiver candidate receives the load information transmission request from the sender, it performs load measurement 51 from its current task schedule status,
The obtained load information is returned 52 to the sender.

In FIG. 1, the sender receives the load information 5
Then, based on the load information and the task information of the task, it is determined 6 whether or not the receiver candidate is qualified to process the task. Since another task has already been assigned from the current time to the end time of the frame time, if there is no room to process the task, it is determined to be unqualified, and otherwise, it is determined to be qualified. As a result of the determination, when the number of qualified receivers who have been qualified has reached a certain number 7, the receiver determination instruction is transmitted 8 to each qualified receiver. In FIG. 2B, each receiver qualified person who has received the receiver determination instruction acquires the load information 53 of the receiver qualified persons other than himself and selects the receiver qualified person with the lightest load based on the load information. Determine 54 as one receiver. The processor 55, which is determined to be the receiver,
This is reported to the sender 56. In FIG. 1, when the sender receives 9 the receiver decision report, it issues a task allocation request 10 to the receiver and ends the dynamic load distribution. If the number of qualified persons does not reach a certain number, 7: If there is enough time to process a task or request allocation by the end of the frame time, 11: Re-execute from the receiver candidate selection 3; If there is not 11, if there is at least one receiver qualified person at that time even if the number of qualified persons does not reach a certain number, then 12 and a receiver determination instruction is transmitted 8. If there is no receiver qualified person, the request to generate the task is rejected.

Embodiment 2 FIG. 3 and 4 are process flow charts showing the second embodiment of the present invention. FIG. 3 is a processing flow chart when a task generation request is issued to a certain processor, and FIG. 4 is a processing flow chart of the processor which is selected as a receiver candidate by the sender and to which the receiver determination instruction is transmitted.
In FIG. 3, when a task generation request is generated in a certain processor, the processor acquires load information such as a time schedule of task allocation until the end of the frame time obtained as a result of task scheduling and a memory usage state,
Execution permission / inhibition determination 1 of the task is performed based on the task information regarding the required memory amount and the required processing time of the task. The processing time required for the task is longer than the time from the current time to the end time of the frame time, or another task is already assigned from the current time to the end time of the frame time, so there is room to process the task. If not, it is determined to be unexecutable, and if not, it is determined to be executable. If the result of determination is that execution is possible, task generation 2 is performed in this processor, execution is started, and the load information of its own processor that has fluctuated due to the generation of this task is updated 13.

In FIG. 4A, the load information of each processor is regularly broadcast 57 to all other processors. In FIG. 3, when it is determined that the execution is impossible as a result of execution determination, the processor becomes a sender, and based on the load information and the task information of the task acquired in advance by the broadcasting function 57, each processor It is judged 14 whether the receiver candidate is qualified to process the task. Since another task has already been assigned from the current time to the end time of the frame time, if there is no room to process the task, it is determined to be unqualified, and otherwise, it is determined to be qualified. As a result of the determination, when the number of qualified persons determined to be qualified reaches a certain number 7, the receiver determination instruction is transmitted 8 to each receiver qualified person. In FIG. 4B, each receiver qualified person who has received the receiver determination instruction broadcasts the load information of his own processor to each of the other receiver qualified persons to thereby obtain the load information of the receiver qualified persons other than himself. Acquisition 53 is performed, and one receiver qualified person with the lightest load is determined as one receiver 54 based on this information. If the user decides to be the receiver 55, it reports 56 to the sender accordingly. In FIG. 3, when the sender receives 9 the receiver decision report, it makes a task allocation request 10 to the receiver and ends the dynamic load distribution. If the number of qualified persons does not reach a certain number, 7: If there is enough time to process a task or request allocation by the end of the frame time, 11: Re-execute from the receiver candidate qualification judgment 14 If there is not, 11 even if the number of qualified persons does not reach a certain number, if there is at least one receiver qualified person at that time, 12, then a receiver determination instruction is transmitted 8. If there is no receiver qualified person, the request to generate the task is rejected.

Embodiment 3 5 and 6 are process flow charts showing the third embodiment of the present invention. FIG. 5 is a process flow diagram when a task generation request is issued to a certain processor, and FIG. 6 is a process flow diagram of the processor which is selected as a receiver candidate by the sender and to which the receiver determination instruction is transmitted.
In FIG. 5, when a task generation request is generated in a certain processor, the processor obtains the task scheduling time schedule until the end of the frame time obtained as a result of the task scheduling, and load information such as memory usage status,
Execution permission / inhibition determination 1 of the task is performed based on the task information regarding the required memory amount and the required processing time of the task. The processing time required for the task is longer than the time from the current time to the end time of the frame time, or another task is already assigned from the current time to the end time of the frame time, so there is room to process the task. If not, it is determined to be unexecutable, and if not, it is determined to be executable. If the result of determination is that execution is possible, task generation 2 is performed in this processor, execution is started, and the load information of its own processor that has fluctuated due to the generation of this task is updated 13.

In FIG. 6A, the load information of each processor is periodically written 59 in a shared memory accessible by all the processors. In FIG. 5, if the result of the execution permission / inhibition judgment indicates that the execution is impossible, the processor acts as a sender and reads the load information written in advance in the shared memory 15, and this information and the task of the task are read. Based on the information, it is judged 6 whether each receiver candidate is qualified to process the task. Since another task has already been assigned from the current time to the end time of the frame time, if there is no room to process the task, it is determined to be unqualified, and otherwise, it is determined to be qualified. As a result of the determination, when the number of qualified receiver qualified persons reaches a certain number 7, the receiver determination instruction is transmitted 8 to each qualified receiver receiver. In FIG. 6B, each receiver qualified person who has received the receiver determination instruction acquires the load information of the receiver qualified persons other than himself from the shared memory by writing 60 the load information of the own processor to the shared memory. Then, based on this information, one receiver qualified person with the lightest load is determined as one receiver 54. If the user decides to be the receiver 55, it reports 56 to the sender accordingly. In FIG. 5, when the sender receives the receiver decision report 9, the sender receives a task allocation request 1
0 is performed, and the dynamic load distribution ends. When the number of qualified persons does not reach a certain number, 7 indicates that there is enough time to process tasks and request allocation by the end of the frame time. 1
1 is re-executed from reading load information 15, and when there is no time, 11 is receiver 12 even if the number of qualified persons does not reach a certain number. The decision instruction is transmitted 8. If there is no receiver qualified person, the request to generate the task is rejected.

Embodiment 4 7 and 8 are process flow charts showing the fourth embodiment of the present invention. FIG. 7 is a process flow diagram when a task generation request is issued to a certain processor, and FIG. 8 is a process flow diagram of the processor that is selected as a receiver candidate by the sender and has received a receiver determination instruction.
In FIG. 7, when a task generation request is generated in a certain processor, the processor acquires load information such as a time schedule of task allocation until the end of the frame time obtained as a result of task scheduling and a memory usage state,
Execution permission / inhibition determination 1 of the task is performed based on the task information regarding the required memory amount and the required processing time of the task. The processing time required for the task is longer than the time from the current time to the end time of the frame time, or another task is already assigned from the current time to the end time of the frame time, so there is room to process the task. If not, it is determined to be unexecutable, and if not, it is determined to be executable. If the result of determination is that execution is possible, task generation 2 is performed in this processor and execution is started. If the result of determination is that execution is not possible, the processor becomes a sender, and the sender selects 3 as the receiver candidate for the processor with the next processor number, and sends the load information to the receiver candidate. Request 4 to. FIG.
In (a), when the receiver candidate receives the load information transmission request from the sender, the receiver candidate measures the load 51 from the current task schedule state of itself, and returns 52 the obtained load information to the sender. In FIG. 7, when the sender receives 5 the load information, it determines 6 based on this information and the task information of the task whether the receiver candidate is qualified to process the task. If another task is already assigned from the current time to the end time of the frame time and there is no room to process the task, it is determined that there is no qualification,
If not, it is determined to be qualified. As a result of the determination, when the number of qualified receivers who have been qualified has reached a certain number 7, the receiver determination instruction is transmitted 8 to each qualified receiver.

Each receiver qualified person who has received the receiver determination instruction in FIG. 8B acquires 53 the load information of the receiver qualified persons other than himself, and selects the processor with the lightest load based on the load information. One receiver 54 is determined. If the user decides to be the receiver 55, it reports 56 to the sender accordingly. In FIG. 7, when the sender receives 9 the receiver decision report, it issues a task allocation request 10 to the receiver. In FIG. 8C, the receiver that has received the task allocation request determines whether or not the task can be executed based on its own load information and the task information of the task. Run 63,
An accept signal is transmitted 64 to the sender to report that the task allocation request has been accepted. Execution is impossible because the receiver receives a task allocation request due to a delay in the network, etc., and the receiver accepts another task allocation request from another processor during that time. If it is determined that the reject signal is transmitted 65 to the sender. In FIG. 7, the sender waits for a response 16 from the receiver after making a task allocation request 10, ends load balancing when an accept signal is returned, and ends the frame time when a reject signal is returned. If there is enough time to process the task or request allocation, the receiver determination instruction is retransmitted to the remaining receiver qualified person if 17 and if there is no time, the request to generate the task is issued 17 Reject. When the number of qualified people has not reached a certain number, 7,
If there is enough time, 11 is re-executed from receiver candidate selection 3. If there is not enough time, 11 is one receiver qualified person at that time even when the number of qualified persons does not reach a certain number. If there is more than 12, 12 the receiver determination instruction is transmitted. If there is no receiver qualified person, the request to generate the task is rejected.

Embodiment 5. 9 and 10 are process flow charts showing the fifth embodiment of the present invention. FIG. 9 is a process flow diagram when a task generation request is issued to a certain processor,
FIG. 10 is a process flow diagram of the processor that is selected as a receiver candidate by the sender and has received the receiver determination instruction. In FIG. 9, when a task generation request is made in a certain processor, the processor obtains the task scheduling time schedule until the end of the frame time, the load information such as the memory usage status, and the required memory amount of the task. Executability determination 1 of the task is performed based on the task information regarding the required processing time and the like. The processing time required for the task is longer than the time from the current time to the end time of the frame time, or another task is already assigned from the current time to the end time of the frame time, so there is room to process the task. If not, it is determined to be unexecutable, and if not, it is determined to be executable. If the result of determination is that execution is possible, task generation 2 is performed in this processor, execution is started, and the load information of its own processor that has fluctuated due to the generation of this task is updated 13.

In FIG. 10A, the load information of each processor is periodically broadcast 57 to all other processors.
I do. In FIG. 9, if the result of the execution propriety determination is that execution is impossible, the processor becomes a sender,
Based on the load information and the task information of the task acquired in advance by the broadcasting function 57, it is determined 14 whether each receiver candidate is qualified to process the task. Since another task has already been assigned from the current time to the end time of the frame time, if there is no room to process the task, it is determined to be unqualified, and otherwise, it is determined to be qualified. As a result of the determination, if the number of qualified receivers who have been determined to be qualified reaches a certain number 7, the receiver determination instruction is transmitted 8 to each qualified receiver. FIG.
In 0 (b), each receiver qualified person who receives the receiver determination instruction broadcasts the load information of the own processor to each of the other receiver qualified persons to acquire the load information of other receiver qualified persons. 53, and based on this information, the qualified person with the lightest load is determined as one receiver 5
4. If the user decides to be the receiver 55, it reports 56 to the sender accordingly. When the sender 9 receives the receiver decision report, it sends a task allocation request 10 to the receiver.
I do.

In FIG. 10C, the receiver that has received the task allocation request determines whether or not the task can be executed based on its own load information and the task information of the task.
Then, if possible, task generation is executed 63, an accept signal is transmitted 64 to the sender, and it is reported that the task allocation request has been accepted. Execution is impossible because the receiver receives a task allocation request due to a delay in the network, etc., and the receiver accepts another task allocation request from another processor during that time. If it is determined that the reject signal is transmitted 65 to the sender. In FIG. 9, the sender waits for a response from the receiver 16 after making a task allocation request 10, ends load balancing when an accept signal is returned, and when a reject signal is returned,
If there is enough time to process a task or request an allocation by the end of the frame time, the receiver determination instruction is retransmitted to the remaining receiver qualified person if 17 and if there is no time, Reject the task creation request. When the number of qualified persons does not reach a certain number, 7 when there is enough time, 11 when it is qualified for each receiver candidate 14
If there is not enough time 11, then if there is one or more receiver qualified persons at that time even if the number of qualified persons does not reach a certain number, 12, then send 8 receiver decision instruction 8 . If there is no receiver qualified person, the request to generate the task is rejected.

Embodiment 6 FIG. 11 and 12 are process flow charts showing the sixth embodiment of the present invention. FIG. 11 is a process flow diagram when a task generation request is issued to a certain processor, and FIG. 12 is a process flow diagram of the processor which is selected as a receiver candidate by the sender and to which a receiver determination instruction is transmitted. In FIG. 11, when a task generation request occurs in a certain processor, the processor obtains the task scheduling time schedule until the end of the frame time, load information such as the memory usage status, and the required memory amount of the task. Executability determination 1 of the task is performed based on the task information regarding the required processing time and the like. The processing time required for the task is longer than the time from the current time to the end time of the frame time, or another task is already assigned from the current time to the end time of the frame time, so there is room to process the task. If not, it is determined to be unexecutable, and if not, it is determined to be executable. If the result of determination is that execution is possible, task generation 2 is performed in this processor, execution is started, and the load information of its own processor that has fluctuated due to the generation of this task is updated 13.

In FIG. 12A, the load information of each processor is periodically written 59 in a shared memory accessible by all the processors. In FIG. 11, when it is determined that the execution is impossible as a result of the execution permission / inhibition determination, the processor serves as a sender and reads the load information written in advance in the shared memory 15, and the information and the task of the task are read. Based on the information, it is judged 6 whether each receiver candidate is qualified to process the task. As a result of the judgment, if the number of qualified receivers who are judged to be qualified reaches a certain number, 7.
Send receiver decision instruction to each receiver qualified person 8
I do. In FIG. 12B, each receiver qualified person who has received the receiver determination instruction acquires the load information of the receiver qualified persons other than himself from the shared memory by writing 60 the load information of the own processor to the shared memory. 61
Then, based on this information, the qualified person with the lightest load is determined as one receiver 54. If the user decides to be the receiver 55, it reports 56 to the sender accordingly. In FIG. 11, when the sender receives 9 the receiver decision report, it issues a task allocation request 10 to the receiver. FIG.
In (c), the receiver that has received the task allocation request determines whether or not the task can be executed based on its own load information and the task information of the task, and executes the task generation 63 if possible. Then, an accept signal is transmitted 64 to the sender to report that the task allocation request has been accepted. Execution is impossible because the receiver receives a task allocation request due to a delay in the network, etc., and the receiver accepts another task allocation request from another processor during that time. If it is determined that the reject signal is transmitted 65 to the sender.

In FIG. 11, the sender is the task allocation request 1
After performing 0, wait for a response from the receiver 16, terminate load balancing when an accept signal is returned, and perform task processing or allocation request by the end of the frame time when a reject signal is returned If there is enough time, the receiver determination instruction is retransmitted 8 to the remaining qualified receivers, and if there is no time, the task generation request is rejected. When the number of qualified persons does not reach a certain number, 7: If there is enough time, re-execute from load information reading 15 of the shared memory; if there is not enough time, 11: The number of qualified persons is constant. If there is one or more receiver qualified persons at that time even if the number has not reached 12, the receiver determination instruction is transmitted 8 if it is 12. If there is no receiver qualified person, the request to generate the task is rejected.

Embodiment 7 13 and 14 are process flow charts showing the seventh embodiment of the present invention. FIG. 13 is a process flow diagram when a task generation request is issued to a certain processor, and FIG. 14 is a process flow diagram of a processor that is selected as a receiver candidate by a sender and has received a receiver determination instruction. In FIG. 13, when a task generation request is made in a certain processor, the processor obtains the load information such as the time schedule of task allocation and the memory usage status until the end of the frame time obtained as a result of the task scheduling, and the required memory amount of the task. Executability determination 1 of the task is performed based on the task information regarding the required processing time and the like. The processing time required for the task is longer than the time from the current time to the end time of the frame time, or another task is already assigned from the current time to the end time of the frame time, so there is room to process the task. If not, it is determined to be unexecutable, and if not, it is determined to be executable. If the result of determination is that execution is possible, task generation 2 is performed in this processor and execution is started. If the result of determination is that execution is impossible, the processor becomes a sender, and the sender selects 3 as the receiver candidate the processor with the processor number next to it,
Request 4 to transmit load information to the receiver candidate. In FIG. 14A, when the receiver candidate receives the load information transmission request from the sender, it performs load measurement 51 from the current task schedule status of itself, and returns 52 the load information obtained to the sender. In FIG. 13, when the sender receives 5 the load information, the sender determines 6 based on this information and the task information of the task whether the receiver candidate is qualified to process the task. Since another task has already been assigned from the current time to the end time of the frame time, if there is no room to process the task, it is determined to be unqualified, and otherwise, it is determined to be qualified. As a result of the determination, when the number of qualified receivers who have been qualified has reached a certain number 7, the receiver determination instruction is transmitted 8 to each qualified receiver.

In FIG. 14B, each receiver qualified person who has received the receiver determination instruction acquires the load information 53 of the receiver qualified persons other than himself and creates a list of the receiver qualified persons based on this information. Create 66, and at the same time determine 54 one qualified person with the least load as a receiver. If the user decides to be the receiver 55, it reports 56 to the sender accordingly. In FIG. 13, when the sender receives 9 the receiver decision report, it issues a task allocation request 10 to the receiver. In FIG. 14C, the receiver that has received the task allocation request determines whether or not the task can be executed based on its own load information and the task information of the task.
2. If possible, perform task generation 63 if possible, send 64 an accept signal to the sender, and report that the task allocation request has been accepted. Execution is impossible because the receiver receives a task allocation request due to a delay in the network, etc., and the receiver accepts another task allocation request from another processor during that time. If it is determined that there is enough time to process a task or request allocation by the end of the frame time, 67 searches the receiver qualified person list 68, and when there is a qualified person, 69 searches the list. One receiver 54 is determined from the inside, and a new task allocation request 70 is issued to the receiver. If you do not have time 67
If there is no qualified person or 69, a reject signal is transmitted 65 to the sender.

In FIG. 13, the sender waits for a response 16 from the receiver after making a task allocation request 10, ends load balancing when an accept signal is returned, and terminates the task when a reject signal is returned. Reject the request to generate. If the number of qualified persons has not reached a certain number, 7 re-execute from the receiver candidate selection 3 if there is enough time, and if there is not enough time, 11 the number of qualified persons reaches a certain number. Even if not, if there is one or more qualified receivers at that point in time 12, the receiver determination instruction is transmitted 8. If there is no receiver qualified person, the request to generate the task is rejected.

Embodiment 8 15 and 16 are process flow charts showing the eighth embodiment of the present invention. FIG. 15 is a process flow diagram when a task generation request is issued to a certain processor, and FIG. 16 is a process flow diagram of the processor that is selected as a receiver candidate by the sender and has received a receiver determination instruction. In FIG. 15, when a task generation request is generated in a certain processor, the processor obtains load information such as the time schedule of task allocation and the memory usage status until the end of the frame time obtained as a result of task scheduling, and the required memory amount of the task. Executability determination 1 of the task is performed based on the task information regarding the required processing time and the like. The processing time required for the task is longer than the time from the current time to the end time of the frame time, or another task is already assigned from the current time to the end time of the frame time, so there is room to process the task. If not, it is determined to be unexecutable, and if not, it is determined to be executable. If the result of determination is that execution is possible, task generation 2 is performed in the processor itself, execution is started, and the load information of the processor that has fluctuated due to the generation of the task is updated 13.

In FIG. 16A, the load information of each processor is periodically broadcast 57 to all other processors.
I do. In FIG. 15, if the result of the execution permission / inhibition determination indicates that the processor cannot be executed, the processor becomes a sender, and based on the load information and the task information of the task acquired in advance by the broadcasting mechanism 57, each processor It is judged 4 whether the receiver candidate is qualified to process the task. Since another task has already been assigned from the current time to the end time of the frame time, if there is no room to process the task, it is determined to be unqualified, and otherwise, it is determined to be qualified. As a result of the determination, if the number of qualified receivers who have been determined to be qualified reaches a certain number 7, the receiver determination instruction is transmitted 8 to each qualified receiver. In FIG. 16B, each receiver qualified person who has received the receiver determination instruction broadcasts the load information of the own processor to each of the other receiver qualified persons to thereby obtain the load information of other receiver qualified persons. Acquiring 53, a receiver qualified person list is created 66 based on this information, and at the same time, one lightest qualified person is determined as a receiver 54. If the user decides to be the receiver 55, it reports 56 to the sender accordingly. In FIG. 15, when the sender receives 9 the receiver decision report, it makes a task allocation request 10 to the receiver.

In FIG. 16C, the receiver that has received the task allocation request determines whether or not the task can be executed based on its own load information and the task information of the task.
If possible, the task is generated 63, the load information of the own processor changed by the generation of the task is updated 71, and the accept signal is transmitted to the sender 64.
Then, it reports that the task allocation request has been accepted. Execution is impossible because the receiver receives a task allocation request due to a delay in the network, etc., and the receiver accepts another task allocation request from another processor during that time. If it is determined that there is enough time to process a task or request allocation by the end of the frame time, 67 searches the receiver qualified person list 68, and when there is a qualified person, 69 searches the list. One qualified person with the lightest load is decided as the receiver, except for those already decided as the receiver 5
4 and a new task allocation request 70 is issued to the receiver.
I do. When there is not enough time 67 or when there is no qualified person 69, a reject signal is transmitted 65 to the sender. In FIG. 15, the sender waits for a response 16 from the receiver after making a task allocation request 10, ends load balancing when an accept signal is returned, and requests the generation of the task when a reject signal is returned. Reject. When the number of qualified persons has not reached a certain number, 7 when there is enough time, 11 when the qualification of each receiver candidate is 1
Re-execute from 4 and if there is no time, 11 If there is one or more receiver qualified persons at that time even if the number of qualified persons has not reached a certain number, 12 Send receiver decision instruction 8 . If there is no receiver qualified person, the request to generate the task is rejected.

Embodiment 9 17 and 18 are process flow charts showing the ninth embodiment of the present invention. FIG. 17 is a process flow diagram when a task generation request is issued to a certain processor, and FIG. 18 is a process flow diagram of the processor which is selected as a receiver candidate by the sender and to which the receiver determination instruction is transmitted. In FIG. 17, when a task generation request is generated in a certain processor, the processor determines load information such as the time schedule of task allocation and the memory usage status until the end of the frame time obtained as a result of task scheduling, and the required memory amount of the task. Executability determination 1 of the task is performed based on the task information regarding the required processing time and the like. The processing time required for the task is longer than the time from the current time to the end time of the frame time, or another task is already assigned from the current time to the end time of the frame time, so there is room to process the task. If not, it is determined to be unexecutable, and if not, it is determined to be executable. If the result of determination is that execution is possible, task generation 2 is performed in the processor itself, execution is started, and the load information of the processor that has fluctuated due to the generation of the task is updated 13.

In FIG. 18A, the load information of each processor is periodically written 59 in a shared memory accessible by all the processors. In FIG. 17, when the result of the execution permission / inhibition determination indicates that the execution is impossible, the processor acts as a sender to read the load information written in advance in the shared memory 15, and this information and the task of the task. Based on the information, it is judged 6 whether each receiver candidate is qualified to process the task. Since another task has already been assigned from the current time to the end time of the frame time, if there is no room to process the task, it is determined to be unqualified, and otherwise, it is determined to be qualified. As a result of the determination, when the number of qualified receiver qualified persons reaches a certain number 7, the receiver determination instruction is transmitted 8 to each qualified receiver receiver. In FIG. 18B, each receiver qualified person who has received the receiver determination instruction acquires the load information of the receiver qualified persons other than himself from the shared memory by writing 60 the load information of the own processor to the shared memory. 61, a list of receiver qualified persons is created 66 based on this information, and at the same time, the lightest qualified person is determined 54 as one receiver. If the user decides to be the receiver 55, it reports 56 to the sender accordingly.

In FIG. 17, when the sender receives the receiver decision report 9, it makes a task allocation request 10 to the receiver. In FIG. 18C, the receiver that has received the task allocation request determines whether or not the task can be executed based on its own load information and the task information of the task.
If possible, the task is generated 63, the load information of the own processor changed by the generation of the task is updated 71, and the accept signal is transmitted to the sender 64.
Then, it reports that the task allocation request has been accepted. Execution is impossible because the receiver receives a task allocation request due to a delay in the network, etc., and the receiver accepts another task allocation request from another processor during that time. If it is determined that there is enough time to process a task or request allocation by the end of the frame time, 67 searches the receiver qualified person list 68, and when there is a qualified person, 69 searches the list. One qualified person with the lightest load is decided as the receiver, except for those already decided as the receiver 5
4 and a new task allocation request 70 is issued to the receiver.
I do. When there is not enough time 67 or when there is no qualified person 69, a reject signal is transmitted 65 to the sender. In FIG. 17, the sender waits for a response from the receiver 16 after making a task allocation request 10, ends load balancing when an accept signal is returned, and requests the generation of the task when a reject signal is returned. Reject. When the number of qualified persons does not reach a certain number, 7 re-executes from the load information reading 15 of the shared memory if there is enough time, and if there is not enough time, the number of qualified persons reaches 11 Even if not, if there is one or more qualified receivers at that point in time 12, the receiver determination instruction is transmitted 8. If there is no receiver qualified person, the request to generate the task is rejected.

Embodiment 10 FIG. 19 and 20 are process flow charts showing the tenth embodiment of the present invention. FIG. 19 is a process flow diagram when a task generation request is issued to a certain processor, and FIG. 20 is a process flow diagram of the processor that is selected as a receiver candidate by the sender and has received a receiver determination instruction. In FIG. 19, when a task generation request is generated in a certain processor, the processor determines load information such as the time schedule of task allocation and the memory usage status until the end of the frame time obtained as a result of task scheduling, and the required memory amount of the task. Executability determination 1 of the task is performed based on the task information regarding the required processing time and the like. The processing time required for the task is longer than the time from the current time to the end time of the frame time, or another task is already assigned from the current time to the end time of the frame time, so there is room to process the task. If not, it is determined to be unexecutable, and if not, it is determined to be executable. If the result of determination is that execution is possible, task generation 2 is performed in its own processor and execution is started. If the result of determination is that execution is not possible, the processor becomes a sender, and the sender selects 3 as the receiver candidate for the processor with the next processor number, and sends the load information to the receiver candidate. Request 4 to. In FIG. 20A, when the receiver candidate receives a load information transmission request from the sender, it performs load measurement 51 from the current task schedule status of itself, and returns 52 the obtained load information to the sender. In FIG. 19, when the sender receives the load information 5, it determines 6 based on this information and the task information of the task whether the receiver candidate is qualified to process the task. Since another task has already been assigned from the current time to the end time of the frame time, if there is no room to process the task, it is determined to be unqualified, and otherwise, it is determined to be qualified. When it is determined that the candidate is qualified, the receiver candidate is instructed to secure resources such as a memory and a processing time within a frame time required for task generation / execution 1
8

In FIG. 20B, when the receiver candidate receives the resource reservation instruction, it secures the resource 72 and ensures that the task can be executed when a task allocation request is made. In FIG. 19, the qualification judgment 6
As a result, if the number of qualified receivers who have been determined to be qualified has reached a certain number, a receiver determination instruction is transmitted 8 to each qualified receiver. In FIG. 20C, each receiver qualified person who has received the receiver determination instruction acquires the load information of the receiver qualified persons other than himself 53.
Then, based on this information, the qualified person with the lightest load is determined as one receiver 54. If the user decides to be the receiver 55, it reports 56 to the sender accordingly. When the receiver itself is not selected as the receiver, the reservation of the resource reserved by the resource reservation 72 is released 73 to avoid unnecessary consumption of the resource. In FIG. 19, when the sender receives 9 the receiver decision report, it makes a task allocation request 10 to the receiver. As a result of the qualification determination 6, when it is determined that there is no qualification, or when the number of qualified persons does not reach a certain number, 7 means that there is a time margin to process a task or request an allocation by the end of the frame time. Re-execute 11 from the selection 3 of receiver candidates, and if there is no time, then if there is at least one receiver qualified person at that time even if the number of qualified persons has not reached a certain number. 12, the receiver determination instruction is transmitted 8. If there is no receiver qualified person, the request to generate the task is rejected.

Embodiment 11 FIG. 21 and 22 are processing flow charts showing an embodiment 11 of the present invention. FIG. 21 is a processing flow chart when a task generation request is issued to a certain processor, and FIG. 22 is a processing flow chart of the processor which is selected as a receiver candidate by the sender and to which the receiver determination instruction is transmitted. In FIG. 21, when a task generation request is made in a certain processor, the processor obtains the task scheduling time schedule until the end of the frame time, load information such as memory usage status, and the required memory amount of the task. Executability determination 1 of the task is performed based on the task information regarding the required processing time and the like. The processing time required for the task is longer than the time from the current time to the end time of the frame time, or another task is already assigned from the current time to the end time of the frame time, so there is room to process the task. If not, it is determined to be unexecutable, and if not, it is determined to be executable. If the result of determination is that execution is possible, task generation 2 is performed in the processor itself, execution is started, and the load information of the processor that has fluctuated due to the generation of the task is updated 13.

In FIG. 22A, the load information of each processor is periodically broadcast 57 to all other processors. In FIG. 21, as a result of the execution feasibility determination,
If the execution is impossible, the processor becomes a sender, and each receiver candidate is qualified to process the task based on the load information and the task information of the task acquired in advance by the broadcasting function 57. It is determined 14 whether there is any. Since another task has already been assigned from the current time to the end time of the frame time, if there is no room to process the task, it is determined to be unqualified, and otherwise, it is determined to be qualified. If you are qualified,
The receiver candidate is instructed 18 to secure resources such as a memory and a processing time within a frame time necessary for generating and executing a task. In FIG. 22B, when the receiver qualified person receives the resource reservation instruction, it secures the resource 72 and ensures that the task can be executed when a task allocation request is made. In FIG. 21, as a result of the qualification determination 14, when the number of qualified receivers who are determined to be qualified reaches a certain number 7, a receiver determination instruction is transmitted 8 to each qualified receiver. In FIG. 22C, each receiver qualified person who has received the receiver determination instruction broadcasts the load information of his own processor to each of the other receiver qualified persons to thereby obtain the load information of the receiver qualified persons other than himself. Acquisition 53 is performed, and one qualified person with the lightest load is determined 54 as a receiver based on this information. If the user decides to be the receiver 55, it reports 56 to the sender accordingly. When the receiver itself is not selected as the receiver, the reservation of the resource reserved by the resource reservation 72 is released 73 to avoid unnecessary consumption of the resource.

In FIG. 21, when the sender receives the receiver decision report 9, it makes a task allocation request 10 to the receiver. As a result of the qualification determination 14, when it is determined that there is no qualification, or when the number of qualified persons does not reach a certain number, 7 means that there is a time margin to process a task or request an allocation by the end of the frame time. Re-execute from the qualification judgment 14 of each receiver candidate in 11, and when there is no time, in 11 even if the number of qualified people does not reach a certain number, at least one receiver qualified person at that time If there is 12, then the receiver determination instruction is transmitted 8. If there is no receiver qualified person, the request to generate the task is rejected.

Twelfth Embodiment 23 and 24 are processing flow charts showing Embodiment 12 of the present invention. FIG. 23 is a processing flow chart when a task generation request is issued to a certain processor, and FIG. 24 is a processing flow chart of the processor which is selected as a receiver candidate by the sender and to which the receiver determination instruction is transmitted. In FIG. 23, when a task generation request occurs in a certain processor, the processor obtains the task scheduling time schedule until the end of the frame time obtained as a result of task scheduling, load information such as memory usage, and the required memory amount of the task. Executability determination 1 of the task is performed based on the task information regarding the required processing time and the like. The processing time required for the task is longer than the time from the current time to the end time of the frame time, or another task is already assigned from the current time to the end time of the frame time, so there is room to process the task. If not, it is determined to be unexecutable, and if not, it is determined to be executable. If the result of determination is that execution is possible, task generation 2 is performed in the processor itself, execution is started, and the load information of the processor that has fluctuated due to the generation of the task is updated 13.

In FIG. 24A, the load information of each processor is periodically written 59 in a shared memory accessible by all the processors. In FIG. 23, as a result of the execution permission / inhibition determination, if the execution is impossible, the processor acts as a sender and reads the load information written in advance in the shared memory 15. This information and the task of the task are read. Based on the information, it is judged 6 whether each receiver candidate is qualified to process the task. Since another task has already been assigned from the current time to the end time of the frame time, if there is no room to process the task, it is determined to be unqualified, and otherwise, it is determined to be qualified. When it is determined that the receiver candidate is qualified, an instruction 18 is given to the receiver candidate to secure resources such as a memory and a processing time within a frame time necessary for generating and executing a task.

In FIG. 24B, when the receiver qualified person receives the resource reservation instruction, the resource reservation 7
2 is performed to ensure that the task can be executed when there is a task allocation request. In FIG. 23, as a result of the qualification determination 6, when the number of qualified receivers who are determined to be qualified reaches a certain number 7, the receiver determination instruction is transmitted 8 to each qualified receiver. In FIG. 24C, each receiver qualified person who has received the receiver determination instruction acquires the load information of the receiver qualified persons other than himself from the shared memory by writing 60 the load information of the own processor to the shared memory. Based on this information, one qualified person with the lightest load is determined as one receiver 54. If the user decides to be the receiver 55, it reports 56 to the sender accordingly. When the receiver itself is not selected as the receiver, the reservation of the resource reserved by the resource reservation 72 is released 73 to avoid unnecessary consumption of the resource. In FIG. 23, when the sender receives 9 the receiver decision report, it makes a task allocation request 10 to the receiver. As a result of the qualification determination 6, when it is determined that there is no qualification, or when the number of qualified persons does not reach a certain number, 7 means that there is a time margin to process a task or request an allocation by the end of the frame time. Re-execute from load information reading 15 of the shared memory in 11 and if there is no time, in 11 there is at least one receiver qualified person even when the number of qualified persons does not reach a certain number. 12, the receiver determination instruction is transmitted 8. If there is no receiver qualified person, the request to generate the task is rejected.

[0043]

As described above, according to the present invention, in a real-time multiprocessor system in which each processor must process a task assigned within a predetermined frame time, a new generation request is made by a certain processor in the system. When a task occurs, it is first determined whether the requested task can be processed in that processor, and when it is determined that it cannot be processed, dynamic load balancing is requested to another processor for task allocation. By doing so, it is possible to prevent an unnecessary increase in the load on the processor and improve the processing efficiency.

Further, when the new generation request task is determined to be unprocessable by the own processor and the other processor is requested to perform the processing substitution, the receiver qualified personnel make decisions to reduce the load. It is possible to prevent the sender, which intends to perform the dynamic load balancing, from increasing the load for the dynamic load balancing process.

[Brief description of drawings]

FIG. 1 is a flowchart showing a process of a sender according to a first embodiment of a dynamic load distribution method according to the present invention.

FIG. 2 is a flowchart showing processing of a receiver candidate according to the first embodiment of the dynamic load distribution method according to the present invention.

FIG. 3 is a flowchart showing processing of a sender according to the second embodiment of the dynamic load balancing method of the present invention.

FIG. 4 is a flowchart showing processing of a receiver candidate according to the second embodiment of the dynamic load distribution method according to the present invention.

FIG. 5 is a flowchart showing processing of a sender according to the third embodiment of the dynamic load balancing method of the present invention.

FIG. 6 is a flowchart showing processing of a receiver candidate according to the third embodiment of the dynamic load distribution method of the present invention.

FIG. 7 is a flowchart showing processing of a sender according to the fourth embodiment of the dynamic load balancing method of the present invention.

FIG. 8 is a flowchart showing processing of a receiver candidate according to the fourth embodiment of the dynamic load distribution method of the present invention.

FIG. 9 is a flowchart showing processing of a sender according to the fifth embodiment of the dynamic load distribution method of the present invention.

FIG. 10 is a flowchart showing processing of receiver candidates according to the fifth embodiment of the dynamic load distribution method of the present invention.

FIG. 11 is a flowchart showing processing of a sender according to the sixth embodiment of the dynamic load distribution method of the present invention.

FIG. 12 is a flowchart showing processing of a receiver candidate according to the sixth embodiment of the dynamic load distribution method of the present invention.

FIG. 13 is a flowchart showing the processing of the sender according to the seventh embodiment of the dynamic load balancing method of the present invention.

FIG. 14 is a flowchart showing processing of receiver candidates according to the seventh embodiment of the dynamic load distribution method of the present invention.

FIG. 15 is a flowchart showing the processing of the sender according to the eighth embodiment of the dynamic load balancing method of the present invention.

FIG. 16 is a flowchart showing processing of receiver candidates according to the eighth embodiment of the dynamic load balancing method of the present invention.

FIG. 17 is a flowchart showing processing of a sender according to the ninth embodiment of the dynamic load balancing method of the present invention.

FIG. 18 is a flowchart showing processing of receiver candidates according to the ninth embodiment of the dynamic load balancing method of the present invention.

FIG. 19 is a flowchart showing the processing of the sender according to the tenth embodiment of the dynamic load balancing method of the present invention.

FIG. 20 is a flowchart showing processing of receiver candidates according to the tenth embodiment of the dynamic load balancing method of the present invention.

FIG. 21 is a flowchart showing the processing of the sender according to the eleventh embodiment of the dynamic load balancing method of the present invention.

[Fig. 22] Fig. 22 is a flowchart showing processing of a receiver candidate according to the eleventh embodiment of the dynamic load distribution method of the present invention.

FIG. 23 is a flowchart showing the processing of the sender according to the twelfth embodiment of the dynamic load balancing method of the present invention.

[Fig. 24] Fig. 24 is a flowchart showing processing of a receiver candidate according to the twelfth embodiment of the dynamic load distribution method of the present invention.

FIG. 25 is a schematic diagram showing a network of a conventional device and a parallel computer device according to the present invention.

[Explanation of symbols]

101 network, 102 processor, 103
Processor, 1 mn processor.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Hiromi Yamazaki 2-3-3 Marunouchi, Chiyoda-ku, Tokyo Sanryo Electric Co., Ltd. (72) Hideaki Yoshioka 2-3-2 Marunouchi, Chiyoda-ku, Tokyo Ryoden Co., Ltd. (72) Inventor Hiroyuki Takano 2-3-3 Marunouchi, Chiyoda-ku, Tokyo Sanryo Denki Co., Ltd. (72) Inventor Toru Hanazawa 2-3-2 Marunouchi, Chiyoda-ku, Tokyo Sanryo Denki Co., Ltd. (72) Inventor ▲ Taka ▼ Masato Hashi 2-3-3 Marunouchi, Chiyoda-ku, Tokyo Sanryo Electric Co., Ltd.

Claims (12)

[Claims] 1. A plurality of processors having equivalent functions are provided, and a task in each processor must be processed within a predetermined frame time, and a task to be processed at the frame time at the start of the frame time. In a real-time parallel computer that makes a determination and performs task scheduling based on the determination result, when a task generation request occurs as a result of the task scheduling in a processor, the processor obtains a frame obtained as a result of the task scheduling. Whether or not the task can be executed is determined based on the load information such as the time schedule of task allocation and the memory usage status until the end of the time, and the task information related to the required memory amount and the required processing time of the task. The task from the time until the frame time end time If the processing execution of the task is impossible because the required processing time is longer, or another task is already assigned from the current time to the frame time end time, the processor It becomes the sender of the task, selects the processor with the processor number next to that processor among other processors as a receiver candidate, requests the receiver candidate to transmit the load information obtained as a result of the task scheduling, and receives it. Based on the load information of the receiver candidate and the task information of the task, it is judged whether or not the task can be executed in the receiver candidate, that is, whether or not the receiver is qualified as a receiver, and the result of the judgment is pass. Perform tasks and request allocation until a certain number of qualified receivers reach a certain number or until the end of the frame time As long as there is time to spare, new receiver candidates are selected and qualification is repeated, receiver decision commands are issued to all receiver qualified personnel, and the receiver qualified personnel exchange load information with each other. A parallel, characterized in that a qualified person with the lightest load is determined as a receiver, the processor determined to the receiver sends a decision report to the sender, and the sender requests the receiver to process the task after receiving the decision report. Dynamic load balancing method for computers. 2. Each processor has a function of periodically broadcasting its own load information to another processor,
When a task generation request is issued to a certain processor, the processor acquires load information such as a time schedule of task allocation until the end of the frame time obtained as a result of the task scheduling and a memory usage state, and a required memory amount and a required process of the task. Whether or not the task can be executed is determined based on the task information related to time, etc., and as a result, the processing time required for the task is longer than the time from the current time to the frame time end time, or from the current time to the frame time end time. If it becomes impossible to execute the task due to other tasks already being assigned to
The processor becomes the sender of the task, and whether or not the task can be executed by each processor based on the load information of each processor and the task information of the task acquired in advance by the broadcasting function, that is, the qualification as a receiver is satisfied. It judges whether there is, and issues a receiver decision order to all qualified receivers who have passed the judgment,
The receiver qualified person exchanges the load information of each other by using the broadcasting function among the qualified persons and determines the qualified person with the lightest load as the receiver, and the processor determined as the receiver makes a determination report to the sender, 2. The dynamic load balancing method in a parallel computer according to claim 1, wherein the sender requests the receiver to process the task after receiving the decision report. 3. When each processor is provided with a shared memory function that can be freely accessed, when a task generation request is issued to a processor, the processor allocates a task allocation time until the end of the frame time obtained as a result of the task scheduling. Whether or not the task can be executed is determined based on the load information such as the schedule and memory usage status, and the task information related to the required memory size and required processing time of the task. When it becomes impossible to execute the processing of the task because the required processing time of the task is longer or another task is already assigned from the current time to the frame time end time, The processor becomes a sender, and the load information of each processor is read by reading the shared memory. Based on the load information and the task information of the task, it is determined whether each processor can execute the task, that is, whether the processor is qualified as a receiver, and the result of the determination is that there is a receiver The receiver decision command is issued to all qualified persons, and the qualified receivers exchange load information with each other using the shared memory to determine the least loaded qualified person as a receiver. 2. The dynamic load balancing method for a parallel computer according to claim 1, wherein the processor determined in step 1) makes a decision report to the sender, and the sender requests the receiver to process the task after receiving the decision report. 4. When the task allocation request is transmitted to the receiver after the receiver is determined, the receiver delays receiving the task allocation request due to a delay in the network or the like. If the receiver cannot accept the task generation request due to the reason that it has accepted the assignment request for another task, the receiver sends a reject signal to the sender, and the sender sends the reject signal to the task by the end of the frame time. 2. A dynamic load balancing method in a parallel computer according to claim 1, wherein a new receiver decision command is issued to the remaining qualified receivers as long as there is a time margin for processing and allocation request. . 5. When the task allocation request is transmitted to the receiver after the receiver is determined, the receiver delays receiving the task allocation request due to a delay in the network or the like. If the receiver cannot accept the task generation request due to the reason that it has accepted the assignment request for another task, the receiver sends a reject signal to the sender, and the sender sends the reject signal to the task by the end of the frame time. The dynamic load balancing method in a parallel computer according to claim 2, wherein a receiver decision command is newly issued to the remaining receiver qualified persons as long as there is a time margin for processing and allocation request. . 6. When a task allocation request is sent to the receiver after the receiver is determined, the receiver delays receiving the task allocation request due to a delay in the network, etc. If the receiver cannot accept the task generation request due to the reason that it has accepted the assignment request for another task, the receiver sends a reject signal to the sender, and the sender sends the reject signal to the task by the end of the frame time. 4. A dynamic load balancing method in a parallel computer according to claim 3, wherein a receiver decision command is newly issued to the remaining receiver qualified persons as long as there is a time margin for processing and allocation request. . 7. A delay occurs in the network when a task assignment request is sent to the receiver after the receiver is determined, because each receiver qualified person creates a list of the receiver qualified persons when the receiver is determined. The receiver can accept the task creation request because, for example, the receiver has received a task allocation request late due to some reason such as the receiver has received another task allocation request from another processor during that time. If there is no receiver, as long as there is enough time to process tasks and request allocation by the end of the frame time, the receiver will newly determine the receiver from the processors registered in the qualified person list and perform the task processing agency. The dynamic load balancing method in a parallel computer according to claim 1, wherein a request is made. 8. A delay occurs in the network when a task assignment request is sent to the receiver after the receiver is determined, because each receiver qualified person creates a list of the receiver qualified persons when the receiver is determined. The receiver can accept the task creation request because, for example, the receiver has received a task allocation request late due to some reason such as the receiver has received another task allocation request from another processor during that time. If there is no receiver, as long as there is enough time to process tasks and request allocation by the end of the frame time, the receiver will newly determine the receiver from the processors registered in the qualified person list and perform the task processing agency. The dynamic load balancing method in a parallel computer according to claim 2, wherein a request is made. 9. A delay occurs in the network when a task assignment request is sent to a receiver after the receiver is determined, because each receiver qualified person creates a list of the receiver qualified persons when the receiver is determined. The receiver can accept the task creation request because, for example, the receiver has received a task allocation request late due to some reason such as the receiver has received another task allocation request from another processor during that time. If there is no receiver, as long as there is enough time to process tasks and request allocation by the end of the frame time, the receiver will newly determine the receiver from the processors registered in the qualified person list and perform the task processing agency. The dynamic load balancing method in a parallel computer according to claim 3, wherein a request is made. 10. A processing time within a frame time or the like required to generate a task for which a generation request is made to the receiver candidate when the judgment result is a pass at the time of qualifying the receiver candidate. 2. The operation in a parallel computer according to claim 1, wherein the task is surely enabled to be generated and executed when the generation request is actually received by instructing to reserve the resource in advance. Load balancing method. 11. A processing time within a frame time or the like required to generate a task requested to be generated for a receiver candidate when the result of the judgment at the time of qualifying a receiver candidate is a pass. 3. The parallel computer according to claim 2, wherein the task is surely made possible to be generated and executed when the generation request is actually received by instructing to reserve the resource in advance. Load balancing method. 12. A memory, a processing time within a frame time, and the like required to generate a task requested to be generated for a receiver candidate when the judgment result is “pass” at the time of judging the qualification of the receiver candidate. 4. The operation in the parallel computer according to claim 3, wherein the task is surely enabled to be generated / executed when the generation request is actually received by instructing to reserve the resource in advance. Load balancing method.