JPH0877026A - Information processing method and device - Google Patents

Abstract

(57) [Abstract] [Purpose] To provide an information processing method and apparatus for effectively using resources of a processor on another computer connected to a communication line with a minimum overhead. [Configuration] A CPU load state detection unit (250) that detects the load state of the CPU 202 of the first group (201) and a CPU (302, 303, 402) of each group (301, 401) other than the first group Other CPU load status acquisition unit (250) that acquires each load status
One group from each group other than the first group, based on the load state detected by the CPU load state detection unit and each processing load state in the CPU of each group other than the first group acquired. Group selection section (2
50), a CPU lending request unit (250) for requesting a CPU lending from the selected group, and selecting a CPU to lend based on the CPU lending request, and selecting the selected CP.
U is provided with an external CPU execution unit (350) that permits execution of a predetermined process and executes the process, and distributes the load on the CPU.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information processing method and an apparatus therefor, and in particular, a plurality of computers are connected by a high-speed communication line, and the memory (or a part thereof) arranged in each computer as a whole. The present invention relates to an information processing method and an apparatus which constitute one memory space and can operate in the form of a NUMA (Non Uniform Memory Access) type multiprocessor system as a whole.

[0002]

2. Description of the Related Art Conventionally, in the case of a computer system connected by a low-speed communication line such as a LAN, when a large number of programs are executed as processes on its own computer and the processors are crowded, an OS having a process migration function is used. In this case, by moving the running process to another computer connected by a communication line, the processor on the other computer can be effectively used.

[0003]

However, in the case of process migration, in order to move and re-execute a process, memory contents (text, data, stack) necessary for executing the process, a process management table in the OS, etc. Must be transferred to the memory on the destination computer, resulting in a large overhead.

In addition, there are some restrictions as described below for moving a process and re-executing the process. That is, it can be moved only when one process is created. 2 The parent-child relationship between the process and the moved process is broken. 3 The moved process cannot create a new child process.

Subject to restrictions such as This is because it is very difficult to keep the process management information inside the OS consistent between the migration destination OS and the migration source OS. The present invention has been made in view of the above-mentioned conventional example, and with a minimum overhead,
It is another object of the present invention to provide an information processing method and apparatus capable of effectively using the resources of a processor on another computer connected to a communication line without being subject to inconvenient restrictions in process generation, execution, and management. .

[0006]

In order to achieve the above object, an information processing method and apparatus of the present invention have the following configurations. That is,

An information processing apparatus having one or more CPUs, the CPU load status detecting means for detecting a processing load status of each of the CPUs, and the CPUs of other information processing apparatuses.
Other CPU load state acquisition means for acquiring the processing load state in the above, the processing load state detected by the CPU load state detection means, and the CPU of another information processing apparatus acquired by the other CPU load state acquisition means. An information processing device selecting means for selecting another information processing device based on the processing load state, a CPU lending requesting device for requesting the information processing device selected by the information processing device selecting means to lend a CPU, An external CPU execution permitting unit that permits the rented CPU to execute a predetermined process based on the CPU lending request of the CPU lending request unit is provided. Another invention is one or more CPUs.
An information processing apparatus including: a selection means for selecting a rentable CPU based on a CPU renting request from another information processing apparatus; and a current process of the rentable CPU selected by the selection means, or The process allocating means for allocating the process scheduled to be executed to another CPU, and the lendable CPU selected by the selecting means are loaded with a program stored in a predetermined storage means of the other information processing apparatus. Other process execution means for accessing and executing. Another invention is
An information processing method using one or more CPUs, comprising:
CPU load state detection step of detecting the CPU processing load state of the CPU of another group, another group CPU load state acquisition step of acquiring the processing load state of the CPU of another group, and the processing detected by the CPU load state detection step Another group selection step of selecting one group from another group based on the load status and the processing load status of the CPU of the other group acquired by the other group CPU load status acquisition step; and the other group selection step. A CPU lending request step of requesting a CPU lending from another group selected by the group; a lending CPU selecting step of selecting a CPU to lend based on the CPU lending request of the CPU lending request step; and a lending CPU selecting step. External that allows the selected CPU to execute a given process And PU execution permission step, the external CP
The other process execution step of causing a CPU, which is permitted to execute the predetermined process in the U execution permission step, to access and execute the program stored in the predetermined storage means of the other group. Another invention is an information processing method having one or more CPUs, which comprises a selecting step of selecting a lendable CPU based on a CPU lending request, and a lendable CP selected in the selecting step.
A process allocating step of allocating the current process of U or a process scheduled to be executed to another CPU, and the lendable CPU selected in the selecting step
And another process execution step of executing a predetermined process of another group. Further, another invention is a CPU load state detection means for detecting a processing load state of a CPU belonging to the first group, and another CPU for obtaining each processing load state of a CPU belonging to each group other than the first group. Load state acquisition means, processing load state detected by the CPU load state detection means, and each processing load state in the CPU belonging to each group other than the first group acquired by the other CPU load state acquisition means Group selection means for selecting one group from each group other than the first group based on the above, and a CP requesting a CPU lending from the group selected by the group selection means.
U lending request means and CP of the CPU lending request means
Based on a lending request from U, a lending CPU selection unit that selects a CPU to lend, an external CPU execution permission unit that allows the CPU selected by the lending CPU selection unit to execute a predetermined process, and the external CPU execution C permitted by the permitting means to execute the predetermined process
The PU is provided with other process execution means for accessing and executing the program stored in the predetermined storage means of the other group.

In the above configuration, the information processing apparatus having one or more CPUs, wherein the CPU load state detecting means is
The processing load state of each of the CPUs is detected, and the other CPUs are detected.
The load state acquisition means acquires the processing load state in the CPU of another information processing device, and the information processing device selection means sets the C
Based on the processing load state detected by the PU load state detection means and the processing load state of the CPU of the other information processing apparatus acquired by the other CPU load status acquisition means, another information processing apparatus is selected, The CPU lending request unit requests the information processing device selected by the information processing device selection unit to lend the CPU, and the external CPU execution permission unit lends out based on the CPU lending request of the CPU lending request unit. Allow the CPU to execute a predetermined process. Another invention is an information processing apparatus having one or more CPUs, wherein the selecting means selects a rentable CPU based on a CPU lending request from another information processing apparatus, and the process allocating means , The lendable CP selected by the selection means
The current process of U or the process scheduled to be executed is assigned to another CPU, and the other process execution means selects the lendable CPU selected by the selection means.
To access and execute a program stored in a predetermined storage means of the other information processing apparatus. Also,
Another invention is an information processing method using one or more CPUs, which detects a processing load state of a CPU of a first group, acquires a processing load state of a CPU of another group, and detects the processing load state. Based on the obtained processing load state and the obtained processing load state in the CPU of the other group, one group is selected from the other groups, and the lending of the CPU from the selected other group is requested. A CPU to be rented out is selected based on a CPU lending request, the selected CPU is allowed to execute a predetermined process, and the CPU permitted to execute the predetermined process is assigned a predetermined CPU of the other group. To access and execute the program stored in the storage means. Another invention is an information processing method having one or more CPUs,
A lendable CPU is selected based on a CPU lending request, and the current process of the selected lendable CPU or a process scheduled to be executed is assigned to another CPU and the selected lendable CPU is available. CP
Cause U to execute a predetermined process of another group. Another aspect of the present invention is that the CPU load state detection means detects the processing load state of the CPUs belonging to the first group, and the other CPs are detected.
The U load status acquisition means acquires each processing load status in the CPUs belonging to each group other than the first group, and the group selection means detects the processing load status detected by the CPU load status detection means, On the basis of each processing load state in the CPU belonging to each group other than the first group acquired by the CPU load state acquisition means, the first
1 group is selected from each group other than the above group, the CPU lending request means requests the CPU lending from the group selected by the group selection means, and the lending CPU selection means is the CPU lending request means C.
Select the CPU to rent based on the PU lending request,
The external CPU execution permission unit permits the CPU selected by the lending CPU selection unit to execute a predetermined process, and the other process execution unit permits the external CPU execution permission unit to execute the predetermined process. CP
Let U access and execute the program stored in the predetermined storage means of the other group.

The technical background relating to the embodiments of the present invention will be briefly described below.

In recent years, with the advent of a communication medium such as an optical cable, which is reliable and high-speed, and has a performance close to an internal bus of a computer, it is becoming possible to directly use the physical memory from another computer. As a result, a system has been proposed in which the memory (or part of it) allocated to each computer can be directly accessed from any computer as a single memory space.

In such a system, by moving a process as in the process migration as described above, a processor on another computer is not directly used but a processor on another computer is directly used. By moving to the management of the above OS, there is a possibility that it can be used in the same form as the processor on the own computer, that is, in the same environment as the number of processors of the own computer has increased.

In view of such a situation, the information processing apparatus according to the embodiment of the present invention has the following configuration. That is, a processor load detection unit that detects a load state of a processor under the control of a computer, another computer information collection unit that collects load information of another computer connected by a communication line, and a computer that is a lending source of the processor. The lending source computer selection unit that selects the number of processors and the processor lending request unit that requests lending of the number of processors determined by the lending source computer selection unit to the computer selected by the lending source computer selection unit, and the request A processor lending unit that rents out the number of processors that follow, and a processor receiving unit that receives the rented-out processor and puts it under the management of the computer,
The processor return unit is provided for returning the borrowed processor to the lending source computer.

Embodiments according to the present invention will be described in detail below with reference to the accompanying drawings.

FIG. 1 shows an outline of the information processing system configuration of the present invention. Reference numeral 101 is a high speed communication line such as an optical network, and 201, 301 and 401 are communication lines 1.
It is a computer connected by 01. 202, 302, 30
3, 402 are MMU (Memory Management Unit) built-in processors respectively provided in the computers 201, 301, 401, 210, 410 are display & keyboards as user interfaces provided in the computers 201, 401, and 220, 320 are computers 201. , 301 is a secondary storage device such as a hard disk device.

The display and keyboard and the secondary storage device may or may not be installed in the computer. For example, the computer 301 that is not equipped with a display and keyboard can be used as a server machine or a processor pool.

Reference numerals 230, 330 and 430 are physical memories 240, 3 which constitute one single memory space as a whole.
Reference numerals 40 and 440 are local memories for each computer.

Although this local memory is mainly required for the processing of the OS and the like, the physical memories 230 and 33 are used.
Since part of 0,430 can be substituted, it is not necessary.

Each processor 202, 302, 303, 4
02 can access the physical memories 230, 330, and 430.

Numerals 250, 350 and 450 are computers 20
It is each OS which operates on 1,301,401. These OSs (Operating Systems) have almost the same functions, and each computer 201, 301, 401
Each of them operates on the above, and is configured to work cooperatively with each other. As a result, the independence of each computer can be maintained and the flexibility of the entire system can be increased.

FIG. 5 shows an outline of an element processing unit of each main OS (for example, OS 250) in this embodiment. A processor load detecting unit, another computer information collecting unit, and a lending source computer selecting unit. , A processor lending request unit and the like.

Even if each OS operates on the local memory 240, 340, 440 on each computer, the physical memory 2
It may operate on 30, 330, 430. However, O
A part of the data area of S has a physical memory 23 in order to exchange the load information of the processor managed by each OS.
It is placed on 0, 330, 430. 260,36
0,460 are these information exchange areas. Also, 10
Reference numerals 2, 302 and 402 are inter-computer interrupt processing lines for performing asynchronous communication between OSs operating on the respective computers.

2, 3 and 4 show the flow of processing of one embodiment according to the present invention. FIG. 2 shows a processing flow of the processor on the computer which requests the lending of the processor, and FIG.
FIG. 4 shows a lending process flow of a processor that is a lending-requested processor and is a lending target, and FIG. 4 shows a process flow when the lending processor is finished and the lending processor is returned. Indicates.

The execution program corresponding to this flow chart is installed in advance in each OS 250, 350, 450 and is executed by the CPU of each computer.

The computer 20 will be described below according to these flows.
A case where the processor 202 on the first computer requests the computer 301 to lend and the processor 303 on the computer 301 is lent will be described as an example.

In the system as shown in FIG. 1, the OS 250 executed by the processor 202 manages only the processor (processor 202) of its own computer at the time of start-up, and uses the scheduling method such as time slice to set the executable state. A processor is assigned to a process and executed (step S1).

Next, in step S2, the processor load detection unit (FIG. 5) in the OS 250 detects the load status of the processor under the control of its own computer, and in step S11,
Determine whether the load is high, low, or normal. The load status is the ready status (Re
It can be found by the ratio of the number of processes in the ady state) to the number of processors. If this ratio exceeds a certain threshold value, for example, the number of executable processes exceeds four times the number of processors, it is determined that the load is high, and if it is smaller than another threshold value, for example, If the number is less than or equal to the number of processors, it can be determined that the load is low, and if it is between the two thresholds, it can be determined that neither. If the number of processes in the executable state increases, the utilization rate of the processor increases, and if it is determined that the load is high, the process proceeds to step S3. If it is determined that the state is neither, the process returns to step S2. The case where the load is determined to be low will be described later.

The load detection process of this processor is to be executed periodically at an empirically determined interval such as once every 10 mSec.

In step S3, the processor load information of another computer connected by the communication line is collected by the other computer information collecting unit (FIG. 5) of the OS 250. This information collection can be realized by data exchange between OSs using the information exchange areas 260, 360, 460 and the inter-computer interrupt lines 102, 103, 104. At this time, the information of all the computers connected by the communication line may be collected, or some of them may be selected and collected. Further, the number of collected information may be limited in consideration of the time required to collect information from one computer. Further, the computers to be collected at this time may be randomly or automatically selected in a round-robin manner, or may be preset by the user.

In step S4, the lending source computer selection unit (FIG. 5) of the OS 250 selects a computer to rent a processor. As a selection method, a computer with a lighter load than the computer itself and the lightest load is selected. Also,
For example, a selection method may be used in which a load that is at least twice as light as that of the own computer is selected. Also, in this step, the number of processors to borrow is determined. In this determination method, for example, the maximum number of processors in which the load of the computer selected by the above selection method does not exceed the load of the own computer is determined as the number of borrowed processors. For example, if the load value of the selected computer is 2 (for example, the number of processes is 8 / the number of processors is 4), and the load value of the own computer is 8 (for example, the number of processes is 16 / the number of processors is 2), the number of borrowed processors is It becomes 2.

In the above example, one computer is selected and then a plurality of processors are borrowed. However, a plurality of computers may be selected and a processor may be borrowed from each of them.

If the lending source computer is not selected, that is, if there is no computer having a lighter load than its own computer, the high load determination threshold value of the processor load detection unit of the OS 250 is set to the previous value. The setting is made higher (step S7), and the process returns to step S2.

Although the setting was made higher, it is checked again whether the load of the other computer is lighter than that of the time when it was checked before when the operation proceeds to step S03 again.
You may make it return the threshold value to the original value. Such a case can be easily realized by providing the other computer information collecting unit with a mechanism for holding the load information of each computer collected previously.

Next, the case where the computer 301 is selected as the lending source computer in step S4 and it is decided to rent one processor will be described as an example.

In step S5, the processor lending request unit (FIG. 5) of the OS 250 requests the computer (for example, the computer 301) selected by the lending source computer selection unit to
Request the lending of the determined number of processors (for example, one). This is a request from the OS 250 on the own computer to the OS (for example, OS 350) on the lending source computer, which also sets the information exchange areas 260, 360, 460 and the inter-computer interrupt lines 102, 103, 104. It can be realized by communication between the OSs used.

On the other hand, the processing of the lending source computer that has received the lending request will be described below with reference to the flowchart of FIG.

With reference to FIG. 3, a processor (here, either processor 302 or processor 303, which is the processor 303, which executes the OS (OS 350 in this case) on the lending source computer that received the request is the processor 303. In this case, the process for renting out the number of processors requested by the processor lending unit is started in step S101. This lending is performed as follows.

First, in step S102, the processor 303 that executes the requested OS 350 selects a processor to rent. For example, if the configuration method of the multiprocessor system is a symmetric method and the scheduling method is a method in which the loads of all the processors are evenly distributed, whichever processor is selected, the same may be selected randomly. Here, symmetry means that the possessed processors have the same configuration.

If the method of constructing the multiprocessor system is the master-slave system, the master processor cannot be lent out, and the slave processor is selected from the slave processors.

In the following, for example, the case where the processor 303 is selected as the processor to be rented will be described as an example.

Next, in step S103, the process running in the processor 303 and the process scheduled to run are transferred to another processor. In the case of this example, since the processor itself that is currently performing this processing is the lending target processor, the processor 303 is
It is running 0 and the process is never running. However, if processor 302 is chosen as the lending processor, then processor 302
May be running the process. Therefore, a method of transferring a process in operation and a process scheduled to operate to another processor will be described below.

Normally, a process in the ready state (ready state) and waiting for processor allocation for execution is connected to a queue called a ready queue for management. In the current system, this Ready queue is prepared by preparing multiple queues for each process execution priority (such as UNIX), or sorted in order of the process execution priority and queued in one Ready queue. There are methods. When the scheduling method is the time slice method, when the time quantum of the process that is currently assigned and the processor (called the Running state) expires or the event wait state (called the Suspend state) is entered, the processor Switches the execution process. Suspen to return a process that was running until now to the Ready queue (when the time quantum has expired, etc.) or to wait for an event
d Put in the queue, take out one process from the Ready queue instead, put that process in the Running state, and run it. When selecting from the Ready queue, it is common to select from the highest priority.

In the case of a multiprocessor system, this Re
There are several ways to use the ady queue. Each processor has its own Ready queue, and each processor selects an execution processor only from its own Ready queue. Also, prepare one Ready queue that is shared by multiple processors. (In the case of a method that has a queue for each rank, prepare only the number of priorities),
A method in which each processor selects an execution process from the shared Ready queue can be considered.

To transfer a process in the case of a system having a unique Ready queue for each processor, first, the execution of the process being executed by the transfer source processor is interrupted and put in the Ready queue of the transfer destination processor, and then the transfer source processor. of
This can be achieved by removing all the processes queued in the Ready queue from the queue and switching to the Ready queue of the transfer destination processor. In the case of the method of sharing the Ready queue, the process transfer can be realized by interrupting the execution of the process being executed by the transfer source processor and putting it in the shared Ready queue.

When the transfer of all processes is completed and the processor 303 becomes idle, the processor 30
3 is moved from the management of the OS 350 to the management of the OS 250 to move the management source.

First, the processor 303 makes the physical memory area managed by the OS 250 accessible in step S104 so that the program code of the OS 250 can be executed. To realize this, the processor 303 switches its execution mode from the virtual storage mode to the real storage mode. As a result, the destination O
The memory area managed in S250 can be accessed.

Then, in step S105, the OS 250
Jumps to the entry prepared in advance for the processor lending process.

After finishing the jump process, the processor 303 which has started to execute the program code of the OS 250 sets the value of the page table managed by the OS 250 in the base register of the page table (not shown) in step S106. The virtual memory mode is switched to enable operation in the virtual memory space of the OS 250.

Next, in step S107, the base register of the exception vector table (not shown) is also set in the OS 25.
It is set to the position of an exception vector table (not shown) managed by 0. In the case of a processor in which the exception vector table cannot be relocated due to the base register setting, the same processing can be performed by rewriting the jump address of the table.

Through the above processing, the processor 303 is set to O
Since it has been moved to the control of S250, step S1
At 08, the fact that the processor 303 is a processor borrowed from the computer 301 is recorded in the OS 250. Then, in step S109, just like the processor 202 under the control of the normal OS 250, a process in an executable state on the OS 250 is allocated and the process is executed. This is the end of the processor lending process.

In this way, the processor is borrowed from another computer to reduce the load and continue the operation.
Of these, some running processes may be terminated and the load may be lighter. That is, in step S2,
When the processor load detection unit detects that the load is lower than a certain threshold value and determines that the load is low, the process proceeds to step S6. This determination is made based on whether or not the load value is smaller than 1 (the number of processors is the same as the number of processors in the executable state).

After renting the processor 303 as in this example, the processing of FIG. 2 may be executed by the processor 202 or the processor 303. In the case of this example, the case in which the processor 202 executes is taken as an example, but if the processor 303 executes the processing in FIG. 2 and proceeds to steps S1, S2, and S6, the step S201 in FIG. The return processing of the Nissumi processor is started. In the case of this example, the processor 202 proceeds to steps S1, S2, and S6, requests the processor 303, and returns it to the processor 303 (step S2
01) is started. This request may be a means such as an interrupt, or may be made by polling using the process switching timing of the processor 303.

In step S201, the processor return unit returns the own processor (processor 303) to the original computer. The processor return processing can be realized by the same processing as the processor lending processing. First, in step S202, all processes scheduled to be executed by the processor 303 are transferred to another processor. When the transfer of all the processes is completed and the processor 303 enters the idle state, the processor 303 is moved from the management of the OS 250 to the management of the OS 350. Therefore, the execution mode is changed in step S203. The virtual storage mode is switched to the real storage mode, and in step S204, O
S350 jumps to the entry prepared in advance for the processor return process. After the jump process, OS
Processor 3 starting to execute 350 program code
Next, in step S205, the 03 sets the value of the page table managed by the OS 350 in the base register of the page table, switches to the virtual memory mode, and enables operation in the virtual memory space of the OS 350. Next in step S206
Then, the pace register of the exception vector table is also S
It is set to the position of the exception vector table managed by O350.

Through the above processing, the processor 303 is
Now that the system is back under the control of OS350, step S2
At 07, the process in the executable state on the OS 350 is allocated and the process is executed in the same manner as before the lending. This is the end of the processor return process.

The present invention may be applied to a system composed of a plurality of devices or an apparatus composed of a single device. Further, it goes without saying that the present invention can be applied to the case where it is achieved by supplying a program to a system or an apparatus.

As described above, according to this embodiment,
It detects the load status of the processor managed by its own computer, and if it finds that it is in a high load status, it checks the load status of other computers connected by a communication line, selects a computer in a low load status, and determines the number of processors to borrow. It is possible to determine and move the number of processors from the management of the lending source computer to the management of the lending destination computer.

As a result, it is possible to effectively use the processor on another computer connected to the communication line with a minimum overhead and without being subjected to any extra restriction in process generation, execution, and management. An information processing device can be obtained.

[0055]

As described above, according to the present invention,

It is possible to effectively use the resources of the processor on another computer connected to the communication line with a minimum overhead and without being subjected to inconvenient restrictions in the generation, execution, and management of the process.

[Brief description of drawings]

FIG. 1 is a configuration diagram of an information processing apparatus according to an embodiment.

FIG. 2 is a flowchart of processing of a processor on an information processing device that requests lending of a processor.

FIG. 3 is a flowchart of processing of a processor on an information processing device that rents out a processor.

FIG. 4 is a flowchart of a return process of a rented processor.

FIG. 5 is a diagram showing an outline of an element processing unit of each main OS in this embodiment.

[Explanation of symbols]

101 High-speed communication line 102, 103, 104 Interrupt line between CPUs 201, 301, 401 Multiple computers connected by communication line 101 202, 302, 303, 402 MMU built-in processor 210 provided for each computer 210 , 410 Display & keyboard 220, 320 Secondary storage device such as hard disk device 230, 330, 430 Physical memory 240, 340, 440 Local memory on each computer 250, 350, 450 Each computer OS 260, 360 460 information exchange area running on

Claims (17)

[Claims] 1. An information processing apparatus having one or more CPUs, comprising CPU load status detecting means for detecting a processing load status of each of the CPUs, and processing load status of a CPU of another information processing apparatus. Other CPU load status acquisition means, a processing load status detected by the CPU load status detection means, and a processing load status in the CPU of another information processing apparatus acquired by the other CPU load status acquisition means. An information processing apparatus selecting unit that selects another information processing apparatus based on the information processing apparatus; a CPU lending requesting unit that requests the information processing apparatus selected by the information processing apparatus selecting unit to lend a CPU; and a CPU lending requesting unit. And an external CPU execution permission unit that allows the rented CPU to execute a predetermined process based on the CPU lending request. The information processing apparatus characterized by. 2. The information processing device selecting means, if the processing load state detected by the CPU load state detecting means is high, the CPU of another information processing device obtained by the other CPU load state obtaining means. The information processing apparatus according to claim 1, wherein another information processing apparatus is selected based on a processing load state. 3. The information processing device selecting means further includes: if the processing load state detected by the CPU load state detecting means is low and there is a rented CPU, the information processing device including the rented CPU. The information processing apparatus according to claim 1, further comprising means for returning the control to the controlled state. 4. The information processing apparatus according to claim 1, wherein the CPU of the information processing apparatus and the CPU of the other information processing apparatus can interpret and execute the same program code. 5. An information processing apparatus having one or more CPUs, comprising: selecting means for selecting a rentable CPU based on a CPU lending request from another information processing apparatus; and the selecting means. The current process of the CPU that can be lent out, or the process that is scheduled to be executed,
Process allocating means for allocating to another CPU, and other process executing means for causing the lendable CPU selected by the selecting means to access and execute a program stored in a predetermined storage means of the other information processing apparatus. An information processing device comprising: 6. The other process execution means performs the process execution by the lendable CPU selected by the selection means in parallel with the predetermined process execution by the predetermined CPU of the other information processing apparatus. The information processing apparatus according to claim 5, wherein: 7. An information processing method using one or more CPUs, the CP detecting a processing load state of CPUs in a first group.
U load state detection step, other group CPU load state acquisition step for obtaining CPU load state of another group, processing load state detected by the CPU load state detection step, and other group CPU load state acquisition step Another group selection step of selecting one group from another group based on the processing load state of the CPU of the other group acquired by the step, and lending of the CPU from the other group selected by the other group selection step Requesting a CPU lending request, a lending CPU selecting step of selecting a CPU to lend out based on the CPU lending request of the CPU lending requesting step, and a CPU selected in the lending CPU selecting step executes a predetermined process. An external CPU execution permitting step for permitting Information processing, comprising: another process execution step of causing a CPU, which is permitted to execute the predetermined process in a step, to access and execute a program stored in a predetermined storage means of the other group. Method. 8. The other group selection step, if the processing load state detected by the CPU load state detection step is higher than a first predetermined value, the other group C
The information processing method according to claim 7, wherein one group is selected from other groups whose processing load state acquired by the PU load state acquisition step is lower than a second predetermined value. 9. The other CPU, if the processing load state detected by the CPU load state detecting step is lower than a predetermined value and there is a CPU lent out from another group, the other group selecting step further comprises: The information processing method according to claim 7, further comprising the step of returning the control to the control of the other group. 10. The information processing method according to claim 7, wherein the CPUs of the first group and the CPUs of the other group interpret and execute the same program code. 11. The lending CPU selection step further includes selecting the C selected when the CPU to be rented is selected based on the lending request of the CPU in the CPU lending request step.
The information processing method according to claim 7, further comprising a process allocation step of allocating the process to another CPU when the PU has a currently executing process or a process scheduled to be executed. 12. An information processing method having one or more CPUs, comprising: a selecting step of selecting a rentable CPU based on a CPU lending request; and a current lending CPU selected in the selecting step. A process, or a process scheduled to run,
An information processing method comprising: a process allocation step of allocating to another CPU; and another process execution step of causing the lendable CPU selected in the selection step to execute a predetermined process of another group. 13. The other process execution step performs the process execution by the lendable CPU selected in the selection step in parallel with the predetermined process execution by the predetermined CPU of the other group. The information processing method according to claim 12. 14. The group comprises one or more Cs.
The information processing method according to claim 12, which is a management unit that manages a PU. 15. A CPU load state detecting means for detecting a processing load state of a CPU belonging to the first group, and another CPU load state for obtaining each processing load state of a CPU belonging to each group other than the first group. Based on the acquisition means, the processing load status detected by the CPU load status detection means, and the processing load status of the CPU belonging to each group other than the first group acquired by the other CPU load status acquisition means. A group selection means for selecting one group from each group other than the first group, and a CPU selected from the groups selected by the group selection means.
CPU lending request means for requesting lending, a lending CPU selecting means for selecting a CPU to lend based on the CPU lending request of the CPU lending request means, and a CPU selected by the lending CPU selecting means performs a predetermined process. An external CPU execution permission unit that permits execution, and a CPU that is permitted to execute the predetermined process by the external CPU execution permission unit access a program stored in a predetermined storage unit of the other group. An information processing apparatus, comprising: another process executing means for executing the operation. 16. The group comprises one or more Cs.
The information processing apparatus according to claim 15, which is a management unit that manages a PU. 17. The CP selected when the lending CPU selecting unit further selects a CPU to lend based on a lending request from the CPU of the CPU lending requesting unit.
The information processing apparatus according to claim 15, further comprising a process allocating unit that allocates the process to another CPU when U has a currently executing process or a process scheduled to be executed.