JP2000003268A - Method for deciding optimization of memory map - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a method for deciding memory map optimization for quickening the executing speed of a program on a system on which plural kinds of memories whose access speed is different are loaded, and for easily constructing an optimized memory map. SOLUTION: This method comprises priority order deciding steps (ST1 and ST2) for deciding a priority order for programs to be arranged in the order of a memory whose access speed is fast and memory map optimizing steps (ST3-ST5) for deciding a memory and the area sufficient for arranging each program in the order of the memory whose access speed is fast based on the priority order decided in the priority order deciding steps by a linker, and for deciding the memory map of the program operating on the system on which plural kinds of memories whose access speed is different.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a memory map which is suitable for developing a program which operates at high speed on a system equipped with a plurality of types of memories having different access speeds and which is optimal for high-speed operation of the program. The present invention relates to a memory map optimization determination method that can be easily determined.

[0002]

2. Description of the Related Art In a system equipped with a plurality of types of memories, access speeds to the respective memories often differ greatly. Therefore, the execution speed of the program greatly differs depending on which part of the program is arranged in which memory. Therefore, in order to determine a memory map for increasing the execution speed of a program, it is necessary to consider that a frequently accessed part is arranged in a memory as fast as possible.
For this reason, the arrangement in the memory is determined by specifying the start address of each section to the linker by a command line option or a script file.

[0003]

Since the conventional memory map optimization determining method is configured as described above, the size of each section and the capacity of each memory become problems. In other words, it is necessary to consider whether or not a high-priority section can fit in the fast memory, so check the size of each section in advance, and then determine the optimal memory map while considering both size and access frequency. There were issues that needed to be determined.

Further, even if the program is slightly modified, the size of each section changes, so that it is necessary to reconsider the memory map each time, and this work has a problem that it takes time and effort.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and has an optimized memory map capable of increasing the execution speed of a program on a system equipped with a plurality of types of memories having different access speeds. It is an object of the present invention to obtain a memory map optimization determining method that can easily construct a memory map.

[0006]

A memory map optimization determining method according to the present invention includes a priority determining step of determining a priority for a program to be allocated in the order of a memory having a higher access speed, and a priority determining step. Based on the priority determined in the above, the linker determines a memory and its area sufficient for arranging the respective programs in order from the memory having the fastest access speed, and on a system equipped with a plurality of types of memories having different access speeds. And a memory map optimizing step of determining a memory map of an operating program.

A memory map optimization determining method according to the present invention includes a priority determining step of determining a priority of each program based on a priority of each program described in a script file. It is.

The memory map optimization determining method according to the present invention includes a priority determining step of determining a priority of each program based on a priority specified as an option from a command line when the linker is started. Things.

A memory map optimization determining method according to the present invention determines a priority in each object file in common with other object files by determining a priority for each section classified for each function, each variable, and each constant. It is provided with steps.

The memory map optimization determining method according to the present invention includes a priority determining step of determining a priority for each function, variable, and constant of each object file.

A memory map optimization determining method according to the present invention includes a priority determining step of determining a priority for each object file.

The memory map optimization determining method according to the present invention is characterized in that each object file has a section classified by function, variable, and constant in common with other object files, The method includes a priority determination step of determining a priority in any combination of each function, each variable, each constant, and each object file.

In the memory map optimization determining method according to the present invention, the linker counts locations where all functions, variables, constants, and the like are accessed for all object files, and based on the count value, a priority order is determined. Is automatically determined.

A memory map optimization determining method according to the present invention includes a priority determining step of determining a priority based on a result of profiling by executing a program.

[0015]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below. Embodiment 1 FIG. FIG. 1 is an explanatory diagram showing an example of a configuration of a sample program in the memory map optimization determining method according to the first embodiment. In the example shown in FIG. 1, the program is composed of three object files F1, F2 and F3. In the object file F1, a function func1 having a size of 4 and a variable x having a size of 1
1, a constant c1 having a size of 2 is defined. In the object file F2, a function f having a size of 7
unc2, a variable x2 having a size of 1, and a constant c2 having a size of 1 are defined. In the object file F3, a function func3 having a size of 3, a variable x3 having a size of 2, a variable x4 having a size of 1, and a variable x5 having a size of 1 are defined. In each object file, a function is classified as a P section, a variable is classified as a D section, and a constant is classified as a C section.

In the first embodiment, an example in which these object files F1, F2, and F3 are linked and mapped on a system equipped with a fast memory A (size 10) and a slow memory B (size 20). A description is given below.

FIG. 2 is a flowchart showing the configuration of the memory map optimization determining method according to the first embodiment. The operation will be described below with reference to this flowchart. The flowchart shown in FIG. 2 shows the processing operation of designating the priority order for allocating in the fast memory by describing the script file (file) in section units and determining the optimized memory map of the program of FIG. . In this case, the priority is P section, C section
Section and D section shall be specified in this order.

First, the priorities to be arranged in the fast memory described in the script file in sections are read (step ST1, priority determining step). Next, the priorities among the sections are determined from the read priorities (step ST2, priority determining step). In this case, the priority order is P section, C section, and D section.

Therefore, each of the object files F1, F2, F3 included in the P section having the highest priority
It is checked whether each function can fit in the fast memory A. As a result, func1 of the object file F1 has a size of 4 and fits in a memory A of size 10, so that it is allocated in the memory A. Next, fu of the object file F2
Since nc2 is a size 7 and does not fit in the remaining size 6 of the memory A, it is arranged in the memory B. Further, func3 of the object file F3 has a size of 3 and fits in the remaining size 6 of the memory A, so that it is arranged in the memory A (step ST3, memory map optimization step).

Next, it is checked whether each constant of each of the object files F1, F2, and F3 included in the C section having a higher priority can be stored in the fast memory A. As a result,
Since the constant c1 of the object file F1 is size 2 and fits in the remaining size 3 of the memory A, it is arranged in the memory A. Next, the constant c2 of the object file F2 has a size of 1 and fits in the remaining size 1 of the memory A.
(Step ST4, memory map optimization step).

At this point, since the area of the memory A has no area which can be allocated and allocated, the remaining programs of the object files F1, F2 and F3 are all stored in the memory B.
To place. As a result, variables x3, x4 and x5 of the object file F3 are arranged in the remaining area of the memory B (step ST5, memory map optimization step).

FIG. 3 shows that the fast memory A and the slow memory B are designed so that the execution speed of the program can be increased based on the priority determined for each section of each object file. 3 shows an optimized and constructed memory map.

As described above, according to the first embodiment, the section-based program, which is required to operate at a high speed, is stored in the memory A, which has a high access speed, with the priority determined for each section of the object file. The memory allocation optimization method can be easily allocated and arranged in the order of the slower memory B, and the load on the construction work of the optimized memory map that can increase the execution speed of the program can be reduced. is there.

Embodiment 2 FIG. FIG. 4 shows the second embodiment.
4 is a flowchart showing a configuration of a memory map optimization determination method of FIG. In the second embodiment, the priority order for allocating in the fast memory A is specified by describing it in a script file in units of variables or modules. In the second embodiment, the priorities are in the order of the function func1, the variable x3, the function func3, the constant c1, and the variable x1, and the other priorities are at a lower level than the variable x1.

Referring to the flowchart of FIG. 4, first, the script file reads the priority described in units of variables or modules (step ST11, priority determination step), and sets the priority to the function func1 and variable x3. , Function func3, constant c1,
It is determined that the order is the order of the variable x1 (step ST1).
2, priority determination step).

Next, it is determined whether the function func1 enters the memory A or not. Since the function func1 is of size 4 and fits in the memory A of size 10, it is arranged in the memory A (step ST13, memory map optimization step).

Next, it is determined whether or not the variable x3 enters the memory A. Since the variable x3 has a size of 2 and fits in the remaining size 6 of the memory A, it is arranged in the memory A (step S
T14, memory map optimization step).

Next, it is determined whether the function func3 enters the memory A or not. Since the function func3 has a size of 3 and fits in the remaining size 4 of the memory A, it is arranged in the memory A (step ST15, memory map optimizing step).

Next, it is determined whether or not the constant c1 enters the memory A. Since the constant c1 is of size 2 and does not fit in the remaining size 1 of the memory A, it is arranged in the memory B (step ST16, memory map optimization step).

Next, it is determined whether or not the variable x1 enters the memory A. Since the variable x1 is size 1 and fits in the remaining size 1 of the memory A, it is arranged in the memory A (step S
T17, memory map optimization step).

At this time, since there is no more area in the memory A which can be allocated, the rest is allocated in the memory B (step ST18, memory map optimizing step).

FIG. 5 shows the function func1, the variable x3, and the function func of each object file.
3, based on the priority determined in the order of the constant c1, the variable x1, etc., a memory map optimized and constructed for the fast memory A and the slow memory B so that the execution speed of the program can be increased. Show.

As described above, according to the second embodiment, with the priority determined for each variable or module,
An optimized memory map that can easily allocate and arrange the variables or the module-based programs requiring high-speed operation in the order of the memory A with the fastest access and the memory B with the slowest access, thereby increasing the execution speed of the program There is an effect that a memory map optimization determination method that can reduce the load on the construction work of the memory map can be obtained.

Embodiment 3 FIG. 6 shows the third embodiment.
4 is a flowchart showing a configuration of a memory map optimization determination method of FIG. In the third embodiment, the priority order for allocating in the fast memory A is specified by describing the script file in object file units. Note that in the third embodiment, the priority order is the object file F
It is assumed that the order is 1, object file F3, and object file F2.

Referring to the flow chart of FIG. 6, first, the priority described in the script file in object file units is read (step ST21, priority determination step). It is determined that the file is in the order of the file F3 and the object file F2 (step ST22, priority determination step).

Next, it is checked whether the function func1, variable x1, and constant c1 included in the object file F1 can be stored in the memory A (step ST23, memory map optimizing step). In this case, the function func1 is of size 4 and fits in the memory A of size 10, so it is arranged in the memory A. Next, since the variable x1 of the object file F1 has a size of 1 and fits in the remaining size 6 of the memory A, it is arranged in the memory A. Further, the constant c1 of the object file F1 has a size of 2 and fits in the remaining size 5 of the memory A, so that it is arranged in the memory A.

Subsequently, it is checked whether the function func3, the variable x3, the variable x4, and the variable x5 included in the object file F3 can be stored in the memory A (step ST24,
Memory map optimization step). In this case, the function fun
Since c3 is size 3 and fits in the remaining size 3 of memory A, it is arranged in memory A. At this point, since the memory A has no area that can be allocated and allocated, the rest is allocated to the memory B (step ST25, memory map optimization step).

FIG. 7 shows an example of the memory A based on the priority determined for each object file.
4 shows a memory map optimized and constructed so that the execution speed of the program can be increased for the slow memory B.

As described above, according to the third embodiment, the program of the object file requiring the high-speed operation is stored in the memory A with the fast access and the memory B with the slow access at the priority determined for each object file unit. The memory map optimization determination method can be obtained in which the load on the work of constructing an optimized memory map that can increase the execution speed of the program can be reduced easily.

Embodiment 4 FIG. FIG. 8 shows the fourth embodiment.
4 is a flowchart showing a configuration of a memory map optimization determination method of FIG. In the fourth embodiment, the priority order for allocating in the fast memory A is set in units of sections, modules,
Specify by mixing object file units and writing them in the script file. In the fourth embodiment, it is assumed that the priority is in the order of the function func1, the D section and the C section in the object file F3.

Referring to the flowchart of FIG. 8, first, the function func is added to the script file.
1. The priority described in the order of the D section and the C section in the object file F3 is read (step ST31, priority determination step), and the priority is set to the function func1, and the D in the object file F3 is read.
It is determined that the section and the C section are in order (step ST32, priority determination step).

Next, it is determined whether the function func1 can be arranged in the memory A. Since the function func1 is of size 4 and fits in the memory A of size 10, it is arranged in the memory A (step ST33, memory map optimization step).
Subsequently, the memory A stores the variables x3, x4, and x5 of the D section included in the object file F3.
Find out if it fits. Variable x3 is size 2 and memory A
Are placed in the memory A because they fit in the remaining size 6. The variable x4 has a size of 1 and the remaining size of the memory A is 4
, It is arranged in the memory A. Also, since the variable x5 has a size of 1 and fits in the remaining size 3 of the memory A, it is arranged in the memory A (step ST34, memory map optimization step).

Subsequently, constants c1 and c, which are C sections,
2 is stored in the memory A (step ST
35, memory map optimization step). Since the constant c1 has a size of 2 and fits in the remaining size 2 of the memory A, it is arranged in the memory A.

At this point, since the memory A has no area that can be allocated and allocated, the rest is allocated to the memory B (step ST36, memory map optimizing step).

FIG. 9 shows the function func1,
Based on the priority determined in the order of the D section and the C section in the object file F3,
4 shows a memory map optimized and constructed so that the execution speed of the program can be increased for the slow memory B.

As described above, according to the fourth embodiment, the section unit, the module unit, and the high-speed operation are required in the order of priority determined over the section unit, the module unit, the object file unit, and the like.
A program such as an object file unit can be easily allocated and arranged in the order of the memory A with the fastest access and the memory B with the slowest access. There is an effect that a memory map optimization decision method that can be reduced can be obtained.

Embodiment 5 In each of the embodiments described above, the specification of the priority is described in the script file. However, the priority may be specified by a command line option.

Embodiment 6 FIG. In the sixth embodiment, the designation of the priority order is not described in the script file, but the linker counts the places where the functions, variables, constants, and the like are accessed for all the object files, and the count value is used as the value. The priority is automatically determined based on the priority. FIG. 10 is a flowchart showing the configuration of the memory map optimization determining method according to the sixth embodiment. First,
The linker counts locations where all functions, variables, constants, and the like are accessed for all object files, and determines a priority based on the count value (step ST41, priority determination step). In this case, as a result of the search for the access location, the function func1, the variable x3,
Function func3, constant c1, variable x1, function func
2, priority is automatically determined in order of variable x2, variable x4, variable x5, and constant c2 (step ST4).
2, priority determination step).

Next, it is determined whether the function func1 enters the memory A or not. Since the function func1 is of size 4 and fits in the memory A of size 10, it is arranged in the memory A (step ST43, memory map optimization step).

Next, it is determined whether or not the variable x3 enters the memory A. Since the variable x3 has a size of 2 and fits in the remaining size 6 of the memory A, it is arranged in the memory A (step S
T44, memory map optimization step).

Next, it is determined whether the function func3 enters the memory A or not. Since the function func3 has a size of 3 and fits in the remaining size 4 of the memory A, it is arranged in the memory A (step ST45, memory map optimization step).

Next, it is determined whether or not the constant c1 enters the memory A. Since the constant c1 is of size 2 and does not fit in the remaining size 1 of the memory A, it is arranged in the memory B (step ST46, memory map optimization step).

Next, it is determined whether or not the variable x1 enters the memory A. Since the variable x1 has a size of 1 and fits in the remaining size 1 of the memory A, it is arranged in the memory A (step S
T47, memory map optimization step).

At this point, since the memory A has no area that can be allocated and allocated, the rest is allocated to the memory B (step ST48, memory map optimizing step).

FIG. 11 is a diagram showing a configuration in which the fast memory A and the slow memory B are optimized so that the program execution speed can be increased based on the priority automatically determined by the linker. FIG.

As described above, according to the sixth embodiment, the linker automatically determines all the object files based on the count value of the location where each function, variable, constant, etc. is accessed. Based on the priorities, the programs such as the functions, variables, and constants requiring high-speed operation can be easily allocated and arranged in the order of the memory A with the fastest access and the memory B with the slowest access. This has the effect of providing a memory map optimization determination method that can reduce the load on the work of constructing an optimized memory map that can achieve high speed.

Embodiment 7 FIG. In the seventh embodiment, the priority is determined based on the result of profiling the program once. In the seventh embodiment, as a result of the profile, the function func1, the variable x3, and the function func
c3, constant c1, variable x1, function func2, variable x
It is assumed that the priorities are determined in the order of 2, variable x4, variable x5, and constant c2. FIG. 12 is a flowchart showing the configuration of the memory map optimization determining method according to the seventh embodiment. Note that in the flowchart shown in FIG.
Steps that are the same as or correspond to 0 are denoted by the same reference numerals, and description thereof is omitted. First, the priority is determined based on the result of actually executing the program or profiling using a simulator or the like (step ST51, priority determination step). In this case, function func1, variable x3, function func3, constant c1, variable x1, and function f
The priority is determined in the order of unc2, variable x2, variable x4, variable x5, and constant c2 (step ST5).
2, priority determination step). The operation of each of the steps after the next step ST43 is the same as the operation described in the flowchart of FIG.

FIG. 13 shows the execution speed of the program for the early memory A and the slow memory B according to the priority determined based on the result of actually executing the program or profiling by using a simulator or the like. 3 shows a memory map optimized and constructed so as to speed up the operation.

As described above, according to the seventh embodiment, a function requiring high-speed operation is determined by the priority determined based on the result of actually executing a program or profiling using a simulator or the like. , Variables, constants, etc., can be easily allocated and arranged in order from the memory A with the fastest access to the memory B with the slowest access. There is an effect that a memory map optimization determination method that can reduce the problem can be obtained.

[0060]

As described above, according to the present invention, the priorities for the programs to be arranged in the order of the memory having the faster access speed are determined, and based on the determined priorities,
The linker determines a sufficient memory and its area for arranging the programs in order from the memory having the fastest access speed, and is optimal for increasing the execution speed of the program on a system equipped with a plurality of types of memories having different access speeds. Optimized memory map that can increase program execution speed on systems equipped with multiple types of memories with different access speeds There is an effect that can be easily constructed.

According to the present invention, since the priorities of the programs are determined based on the priorities of the programs described in the script file, a plurality of types of memories having different access speeds are mounted. This makes it possible to easily construct an optimized memory map that can increase the execution speed of a program on a system that has been set up, without any trouble, by prioritizing each program described in the script file.

According to the present invention, since the priority of each program is determined based on the priority specified as an option from the command line when the linker is started, a plurality of types of memories having different access speeds can be stored. There is an effect that an optimized memory map that can increase the execution speed of a program on a mounted system can be easily constructed without any trouble by the priority specified as an option from the command line.

According to the present invention, in each object file, the priority is determined for each section classified by function, variable, and constant in common with other object files. Optimized memory maps that can increase the execution speed of programs on systems equipped with different types of memory are shared by all functions, variables, and constants in all object files. There is an effect that it can be easily constructed without any trouble according to the priority order.

According to the present invention, the priorities are determined for each function, variable, and constant of each object file, so that the program can be executed on a system equipped with a plurality of types of memories having different access speeds. There is an effect that an optimized memory map that can increase the speed can be easily constructed without any trouble by priority of each function, variable, and constant of each object file.

According to the present invention, the priorities are determined for each object file. Therefore, the optimization can be performed to increase the execution speed of the program on a system equipped with a plurality of types of memories having different access speeds. This makes it possible to easily construct the memory map according to the priority of each object file without any trouble.

According to the present invention, each object file has a section classified by function, variable, and constant in common with other object files.
For each function, variable, and constant of each object file,
Optimized memory that can increase the program execution speed on a system equipped with multiple types of memories with different access speeds because the priority is determined by any combination of the object files There is an effect that the map can be easily constructed without trouble by the sections, the functions, variables and constants of the object files, and the priorities determined by including the object files.

According to the present invention, the linker counts locations where all functions, variables, constants, etc. are accessed for all object files, and automatically determines priorities based on the count values. Because it was configured as
Priority determined based on the access frequency of each of the above functions, variables, constants, etc., with an optimized memory map that can increase the execution speed of the program on a system equipped with a plurality of types of memories having different access speeds Thus, there is an effect that the construction can be easily performed without trouble.

According to the present invention, the priority is determined based on the result of profiling by executing the program. Therefore, the execution speed of the program on a system equipped with a plurality of types of memories having different access speeds is determined. There is an effect that an optimized memory map which can achieve high speed can be easily constructed without any trouble by the priority determined based on the profiled result.

[Brief description of the drawings]

FIG. 1 is an explanatory diagram showing an example of a configuration of a sample program in a memory map optimization determining method according to a first embodiment of the present invention.

FIG. 2 is a flowchart illustrating a configuration of a memory map optimization determining method according to the first embodiment of the present invention;

FIG. 3 is a memory map diagram optimized and constructed by the memory map optimization determining method according to the first embodiment of the present invention;

FIG. 4 is a flowchart illustrating a configuration of a memory map optimization determining method according to Embodiment 2 of the present invention;

FIG. 5 is a memory map diagram optimized and constructed by a memory map optimization determining method according to the second embodiment of the present invention;

FIG. 6 is a flowchart illustrating a configuration of a memory map optimization determining method according to Embodiment 3 of the present invention;

FIG. 7 is a memory map diagram optimized and constructed by a memory map optimization determining method according to a third embodiment of the present invention;

FIG. 8 is a flowchart illustrating a configuration of a memory map optimization determining method according to Embodiment 4 of the present invention;

FIG. 9 is a memory map diagram optimized and constructed by the memory map optimization determining method according to the fourth embodiment of the present invention;

FIG. 10 is a flowchart showing a configuration of a memory map optimization determining method according to a sixth embodiment of the present invention.

FIG. 11 is a memory map diagram optimized and constructed by a memory map optimization determining method according to a sixth embodiment of the present invention.

FIG. 12 is a flowchart showing a configuration of a memory map optimization determining method according to a seventh embodiment of the present invention.

FIG. 13 is a memory map diagram optimized and constructed by the memory map optimization determining method according to the seventh embodiment of the present invention.

[Explanation of symbols]

F1, F2, F3 object file, step S
T1, step ST2, step ST11, step S
T12, Step ST21, Step ST22, Step ST31, Step ST32, Step ST41, Step ST42, Step ST51, Step ST52
Priority determination step, step ST3 to step S
T5, step ST13 to step ST18, step ST23 to step ST25, step ST33 to step ST36, step ST43 to step ST48
Memory map optimization step.

Claims (9)

[Claims] A priority determining step for determining a priority for a program to be allocated in the order of a memory having a higher access speed, and a priority determined based on the priority determined in the priority determining step.
A memory that determines a memory and an area thereof sufficient for arranging the programs in order from the memory having the fastest access speed, and a memory that determines a memory map of a program operating on a system equipped with a plurality of types of memories having different access speeds. A memory map optimization determining method including a map optimization step. 2. The memory map optimization determining method according to claim 1, wherein, in the priority determining step, the priority of each program is determined based on the priority of each program described in the script file. 3. The memory map optimization determining method according to claim 1, wherein in the priority determining step, the priority of each program is determined based on the priority specified as an option from a command line when the linker is started. 4. In the priority determining step, priority is determined for each section classified into each function, each variable, and each constant in each object file in common with other object files. The memory map optimization determination method according to any one of claims 1 to 3. 5. The priority determining step according to claim 1, wherein the priority is determined for each function, variable, and constant of each object file. Memory map optimization decision method. 6. The memory map optimization determining method according to claim 1, wherein in the priority order determining step, the priority order is determined for each object file. 7. In the priority order determining step, in each object file, each section classified into each function, each variable, and each constant, each function in each object file, 4. The memory map optimization determining method according to claim 1, wherein the priority is determined based on a combination of each of the followings, each of the constants, and each of the object files. . 8. In the priority determining step, the linker counts locations where all functions, variables, constants, and the like are accessed for all object files, and automatically determines the priority based on the count value. The method according to claim 1, wherein the memory map optimization is determined. 9. The memory map optimization determining method according to claim 1, wherein in the priority order determining step, the priority order is determined based on a result of profiling by executing the program.