JP2007299212A - Apparatus and method for evaluating porting of software - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an apparatus and a method for evaluating porting of software resources for precisely commuting a memory capacity and performance under a mass-production environment on the basis of a program size and performance measured under a development environment and for quickly conducting a selection of an appropriate target and tuning work to the target. <P>SOLUTION: An apparatus for evaluating porting of software 50 calculates processing performance of a second processing device 40 from processing performance of a first processing device 30 with an evaluated target program embedded in a processing device performing electronic control, and has a processing performance commuting means 62 for calculating the processing performance on the basis of a processing performance commuting coefficient obtained by weighting a processing performance evaluating coefficient, which is obtained from processing performance of the first and the second devices 30, 40 with a prescribed program, in accordance with a module property composing a task of the program. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a software porting evaluation apparatus composed of a plurality of tasks incorporated in a processing device of an electronic control unit mounted on a vehicle or the like.

  A program as a software resource incorporated in a microcomputer that is a processing unit of an electronic control unit mounted on a vehicle or the like is used to improve the performance and functionality of the engine to be controlled, and to comply with legal regulations such as exhaust gas regulations. Corresponding prior development is required.

  The microcomputer in which such a program is incorporated is also progressing day by day, and a development environment is being constructed so that the program can be developed in advance, targeting a next-generation microcomputer whose performance is expected to improve.

  For example, Patent Document 1 discloses a device for developing logic of an embedded microcomputer that is used by being incorporated in an electronic control unit, and includes a first central processing unit that performs logic processing and data including a logic program. A first block having at least a first memory for storing, a first interface for communicating with the outside, and a first internal bus for connecting these, and a peripheral device of the microcomputer as software A second block having at least a pseudo-microcomputer peripheral device realized by the above, a second interface that communicates with the outside, and a second internal bus that connects these, a first block, and a second block The interface bus is connected, and these first and second blocks and interface bus replace the embedded microcomputer. Logic development system of the microcomputer which is adapted to operate the logic has been proposed.

  Such a logic development device is equipped with a high-performance CPU used in a personal computer or the like that surpasses the performance of a microcomputer mounted in an electronic control unit, and can develop a highly flexible program over a considerable period of time. It is configured to be able to.

  However, when a program, which is a software resource developed in such a development environment, is ported to a target microcomputer selected as a mass production environment, it will fit in the planned memory capacity of ROM and RAM, and performance will be reduced. Since it is not certain whether it can be realized or not, it is necessary to actually transplant and check the operation, but such work has a problem that it is extremely complicated and causes a serious development loss.

  Therefore, as a method for estimating and evaluating the memory capacity required when porting to the mass production environment from the memory capacity of the ROM or RAM in the development environment of the generated evaluation target program, a model program (also called a benchmark program) is previously known. The memory capacity in the development environment and mass production environment is determined in units of files that are grouped together in functional units higher than the individual tasks that make up the model program, and the memory capacity conversion in the development environment and mass production environment obtained for each file A method has been proposed in which the average value of the coefficient is derived as a memory capacity conversion coefficient, and the memory capacity required in the mass production environment of the evaluation target program is converted from the memory capacity in the development environment based on the memory capacity conversion coefficient. It had been.

  Although it is possible to obtain the program size by converting the developed source program into an object program using a compiler or assembler that is compatible with a microcomputer in a mass-production environment, it is therefore compatible with a compiler or assembler that uses the source program. There is also a problem that the work for the conversion process becomes complicated and lacks practicality.

  Further, as a method for estimating and evaluating the performance when transplanted to the mass production environment from the measured performance value in the development environment of the evaluation target program, the model program is configured with the development environment for the model program and the processing speed in the mass production environment in advance. Measured in units of tasks, and the performance ratio of the task units was derived as a speed conversion factor for evaluation, and the performance of each task in the mass production environment of the evaluation target program was measured in the development environment based on the speed conversion factor. There has been proposed a method for converting and outputting from the performance of each task.

  On the other hand, as disclosed in Patent Document 2, in a general-purpose computer system, a computer is used as a performance prediction device that calculates the rate of change in processing performance accompanying a change in the operating status of the computer system, such as the configuration of the computer system and changes in application programs. Based on the processing performance information of the system and the application program executed on the computer system, a plurality of tasks constituting the application program are processed for each task when the task is executed on the computer system. A no-load response estimating means for estimating a certain no-load response, and an operation response for estimating an operation response that is a processing execution time for each task when a plurality of the tasks are operated in the computer system based on the processing performance information An estimation means, the no-load response and the Based on work response, performance prediction unit of a computer system and a specific rate of change detecting means for detecting a specific change rate is a change rate of the processing performance of the computer system due to the change of the load is proposed.

JP 2003-167756 A Japanese Patent Laid-Open No. 9-218807

  However, according to the method and performance prediction device for converting the memory capacity and performance based on the above-described conventional conversion coefficient, the accuracy is extremely low because all of them are evaluated in a large group of file units and task units. There was a problem that it could not be put to practical use.

  For example, a program executed by a processing device of an electronic control unit mounted on a vehicle or the like must surely execute a required process within an extremely short time. It is not possible to fully evaluate the details of which parts of multiple modules need to be tuned. As a result, it is necessary to tune while checking the operation by actually transplanting the program to the mass production environment. The problem cannot be solved. For this reason, it took a very long time to determine the final target, and there was a problem that timely market introduction of new products was difficult.

  An object of the present invention is to store ROM capacity and variables necessary for storing programs and data in a mass production environment based on the program size and performance actually measured in the development environment in view of the above-mentioned conventional drawbacks. Select the appropriate target and tune the target more quickly by accurately converting the memory capacity as the RAM capacity required as a work area and the performance as the processing performance such as processing speed and response time. An object of the present invention is to provide a software resource transplantation evaluation device and an evaluation method.

  In order to achieve the above object, the first characteristic configuration of the software porting evaluation apparatus according to the present invention is the processing performance in the first processing apparatus of the evaluation object program comprising a plurality of tasks incorporated in the processing apparatus that performs electronic control. From the software porting evaluation device for calculating the processing performance in the second processing device, the processing performance evaluation coefficient obtained from the processing performance in the first processing device and the second processing device of the predetermined program, A processing performance conversion processing means for calculating based on a processing performance conversion coefficient obtained by weighting the module characteristics constituting the task of the program to be evaluated is provided.

  According to the above configuration, the processing performance evaluation coefficient for a predetermined program such as a model program obtained in advance, that is, each of the first processing device used in the development environment and the second processing device used in the mass production environment. Instead of performing general performance conversion with performance evaluation coefficients that reflect the characteristics, the processing performance conversion coefficients weighted by the module characteristics that make up the tasks of the evaluation target program are further obtained, thereby specializing in the evaluation target program. The performance can be determined accurately.

  Here, the processing performance conversion coefficient is a processing performance in units of modules constituting each task executed by the first processing device with respect to a predetermined program composed of a plurality of tasks incorporated in the processing device of the electronic control unit. Obtained by multiplying a processing performance evaluation coefficient, which is a ratio of the processing performance of each module executed by the second processing apparatus, by a weighting coefficient according to an execution load ratio of each module of the evaluation target program. It is preferable.

  That is, a plurality of module units constituting each task of a predetermined program measured in the first processing device used in the development environment and the second processing device used in the mass production environment, for example, logic, input / output processing, Since processing performance evaluation coefficients are obtained in units of modules such as integer arithmetic processing and floating point arithmetic processing, for example, performance conversion processing in units depending on the architecture of the microcomputer constituting the core part of each processing device can be performed. Based on the value obtained by multiplying the processing performance evaluation coefficient by the weighting factor based on the execution load ratio represented by the execution ratio of each module that constitutes each task of the evaluation target program, the processing performance conversion coefficient for each task is obtained. Therefore, according to such processing performance conversion coefficient, the characteristics of the evaluation target program are reflected. It's becomes possible to properly calculate the performance of the serial second processing unit.

  As described in claim 4, the second characteristic configuration is based on the memory capacity used in the first processing device of the evaluation target program consisting of a plurality of tasks incorporated in the processing device that performs electronic control. A software porting evaluation device for calculating a memory capacity used in a processing device, which is a task used in the first processing device and the second processing device of a predetermined program or for each module type constituting a task Approximation formula calculating means for calculating an approximate expression for converting the memory capacity used in the second processing device from the memory capacity, and for the evaluation target program, from the memory capacity used in the first processing device to the second The memory capacity conversion processing means for calculating the memory capacity used in the processing device based on the approximate expression is provided.

  According to the above-described configuration, the correlation between the memory capacities in the first processing device and the second processing device for the predetermined program, that is, the approximate expression, is obtained based on the memory capacity used in units of the tasks or modules constituting the tasks. Therefore, according to such an approximate expression, the memory capacity required in the second processing device can be converted with extremely high accuracy from the memory capacity of the evaluation target program obtained for the first processing device. It becomes like this.

  As described above, according to the present invention, it is possible to store ROM capacity and variables necessary for storing programs, data, and the like in a mass production environment based on the program size and performance actually measured in the development environment. The memory capacity as the RAM capacity required for the area and the performance as the processing performance such as processing speed and response time are accurately converted, and the appropriate target selection and target tuning can be performed more quickly. It has become possible to provide a software transplantation evaluation apparatus and an evaluation method that can be used.

  Hereinafter, an embodiment in which a software transplant evaluation device according to the present invention is applied to a port evaluation from a development environment to a mass production environment of an evaluation target program incorporated in a processing device of an electronic control unit that is mounted on a vehicle and performs electronic control will be described. To do.

  As shown in FIG. 1, a vehicle includes a plurality of electronic control units (ECUs) 10 that are network-connected via a CAN (controller area network) bus 20, and the plurality of electronic control units 10 cooperate with each other. It is configured to perform the desired control. The plurality of electronic control units 10 include, for example, an engine control ECU 11, an automatic transmission (AT) control ECU 12, an antilock brake system (ABS) control ECU 13, and the like.

  Each of the plurality of electronic control units 10 is provided with a microcomputer as a processing device, and the microcomputer stores a ROM for storing a program, a RAM as a working area used for data processing, and the program. The CPU is configured to be executed, and is configured to realize a predetermined function based on the program.

  For example, the engine control ECU 11 includes an intake air temperature sensor 110, a water temperature sensor 111, a sensor 112 for detecting a target valve timing in a variable valve timing mechanism, an accelerator pedal operation amount sensor 113, a battery voltage sensor 114, and the like. Based on the input signal from the sensor and communication data from other related electronic control units such as the ECU 12 for automatic transmission control, the engine is controlled to realize a predetermined engine function, or the control state of the other is changed. Send data to other electronic control units.

  Further, in other electronic control units such as the automatic transmission control ECU 12 and the ABS control ECU 13, the accelerator pedal operation amount sensor 113, the battery voltage sensor 114, and the like that have been input to the engine control ECU 11, etc. By controlling the input signals from the sensors 120, 121, 130, 131, etc. that are used independently, the plurality of electronic control units 10 can be configured to cooperate to implement control as a system. Yes.

  As shown in FIG. 2, the development environment of the software built in each electronic control unit described above is the first process as a microcomputer of the electronic control unit 10, that is, a preceding development ECU used for development of a mass production ECU to be described later. The apparatus 30, the second processing device 40 as a microcomputer of an electronic control unit (mass production ECU) to be actually mounted on the vehicle, and the first measurement for measuring the processing performance and memory capacity of the first processing device 30 From the processing performance and memory capacity of the device 31, the second measuring device 41 that measures the processing performance and memory capacity of the second processing device 40, and the software processing performance and memory capacity of the first processing device 30, A transplantation evaluation apparatus 50 for calculating processing performance and memory capacity is provided.

  The memory capacity means the capacity of ROM and variables necessary for storing programs, data, etc., and the capacity of RAM necessary as a work area. The processing performance is processing performance such as processing speed and processing time, that is, performance, and the transplantation evaluation apparatus 50 according to the present embodiment mainly calculates the processing speed out of the processing performance. explain.

  The preceding development ECU realizes the same function as that of the plurality of electronic control units 10 in a pseudo manner, and includes, for example, a logic development device as proposed in Japanese Patent Laid-Open No. 2003-167756. ing. The logic development device is a device for developing logic of an embedded microcomputer (microcomputer in the present invention) used by being incorporated in an electronic control unit, and includes a first central processing unit that performs logic processing, A first block including at least a first memory for storing data including a program, a first interface for communicating with the outside, and a first internal bus for connecting these, and a peripheral device of the microcomputer A second block having at least a pseudo microcomputer peripheral device that realizes the above by software, a second interface that communicates with the outside, and a second internal bus that connects these, An interface bus for connecting the second block, and the first and second blocks and the interface bus. Therefore, a device which is adapted to operate the logic replaces the built-in microcomputer.

  Note that the prior-developed ECU is not limited to the above-described configuration, and may be constructed based on an ECU currently used.

  The first measuring device 31 is configured by a PC or a workstation incorporating a measurement program for measuring the processing speed and memory capacity of the first processing device 30, and the second measuring device 41 is configured by the second processing. It is composed of a PC and a workstation incorporating a measurement program for measuring the processing speed and memory capacity of the device 40. In addition, said 1st measuring device 31 and the 2nd measuring device 41 are with respect to the 1st measuring device 31 used with respect to said 1st processing device 30, and said 2nd processing device 40, as shown in FIG. In addition to the configuration including the second measurement device 41 to be used, the configuration includes a common measurement device that measures the processing speed and memory capacity of both the first processing device 30 and the second processing device 40. Also good.

  When the porting evaluation apparatus 50 transfers the software from the first processing apparatus 30 to the second processing apparatus 40, the porting evaluation apparatus 50 determines whether the second processing apparatus 40 uses the software processing speed or the memory capacity of the first processing apparatus 30. A device that calculates the processing speed and memory capacity of software, and includes a performance conversion means 60 that calculates processing speed and a capacity conversion means 70 that calculates memory capacity.

  The performance conversion means 60 converts the processing speed in the second processing device 40 of the evaluation target program from the processing speed in the first processing device 30 of the evaluation target program consisting of a plurality of tasks incorporated in the processing device of the electronic control unit. A processing performance conversion coefficient calculating means 61 for calculating a processing performance conversion coefficient to be performed; a processing performance conversion processing means 62 for calculating a processing speed in the second processing device 40 of the evaluation target program based on the calculated processing performance conversion coefficient; The performance evaluation means 63 for determining whether or not the processing speed calculated by the processing performance conversion processing means 62 satisfies the allowable processing speed, and the performance evaluation means 63 evaluated that the allowable processing speed is not satisfied. An evaluation content changing means 64 for changing the processing performance conversion coefficient is sometimes provided.

  The program is configured with a plurality of tasks, and the tasks are configured with a plurality of functions and classes. Each of the tasks includes, as a main task, a base processing routine that repeatedly executes a predetermined calculation process such as a fuel injection amount calculation and a fuel ignition timing calculation, an engine speed calculation execution signal, a vehicle speed calculation execution signal, etc. Analog data such as temperature and intake air amount is converted into digital data for use in interrupt processing routines that perform arithmetic processing as specific tasks based on interrupt signals from, and in the base processing routines and interrupt processing routines. This represents processing such as an A / D conversion processing routine to be converted.

  The evaluation target program is development software developed for the purpose of being mounted and operated on a microcomputer of an electronic control unit (mass production ECU) that is actually mounted on a vehicle. It is a configuration specialized for.

  The processing performance conversion coefficient calculation means 61 calculates a processing performance conversion coefficient for converting the processing speed of the evaluation target program in the second processing apparatus 40 from the processing speed of the evaluation target program in the first processing apparatus 30. The processing performance evaluation coefficient calculating means 610, the weight coefficient calculating means 611, and the weight calculating means 612 are provided.

  The processing performance evaluation coefficient calculation means 610 constitutes each task executed by the first processing device 30 with respect to a model program as a predetermined program composed of a plurality of tasks incorporated in the processing device of the electronic control unit. A processing performance evaluation coefficient that is a ratio between the processing speed in module units and the processing speed in module units executed by the second processing device 40 is calculated.

  The model program is benchmark software for measuring the processing speed of the first processing device 30 and the second processing device 40, and the evaluation target program is specialized for each model of the electronic control unit. On the other hand, the model program is composed of essential processes of the electronic control unit, and is composed of a collection of almost common tasks executed by each electronic control unit.

  The processing performance evaluation coefficient calculation means 610 will be described in detail. As shown in FIG. 3, each of the first processing device 30 and the second processing device 40 executes a task T1 in the same model program, The processing speed is calculated in units of modules (T21 to T2n) constituting the task T1.

  As shown in FIG. 3, the modules (T21 to T2n) are processes indicated in alphabetical order from the process A, and each process content in the task T1 (that is, n types of process contents in FIG. 3). Each processing content includes logic processing such as assignment, branching, and repetition of a value to a variable, input / output processing (I / O) to the outside, and arithmetic operations such as integer arithmetic and floating point arithmetic. It represents processing and other processing.

  The first processing result (T31 to T3n) in the first processing device 30 and the second processing result (T41 to T4n) in the second processing device 40 expressed as the calculated processing speed in units of modules are shown in FIG. , Expressed in alphabetical order from the processing A result, the first processing result divided by the second processing result for the same module as the first processing result, for example, A value obtained by dividing the first processing result T31 by the second processing result T41 is calculated as the processing performance evaluation coefficient.

  The weighting factor calculation means 611 is configured to calculate a weighting factor based on the execution load ratio of each module of the evaluation target program as weighting based on the ratio of each module type of the evaluation target program. More specifically, as shown in FIG. 4, the task T5 in the evaluation target program is processed for each module type, for example, input / output processing T21, arithmetic processing T22, and processing other than the input / output processing T21 and arithmetic processing T22. And the ratio of the number of execution times of each module to the number of execution times of the entire task T1 is calculated as weighting coefficients r1, r2, and r3. For example, when the total number of executions of the task T5 is 1000 and the number of executions of the input / output process T21 is 300, the weight coefficient r1 is 0.3. Note that the unit of the weight coefficient may be a percentage or the like.

  As another example, when calculating the weighting factor of a task shown in C language, as shown in FIG. 5 (a), after excluding a portion where processing is not executed in each process of task T5, (Process P) As shown in FIG. 5B, the processes are classified according to module types. In FIG. 5B, the process P1 is classified as input / output processing, the process P2 is classified as integer arithmetic processing, the process 4 is classified as logic processing, the process 5 is classified as floating point arithmetic processing, and the process 6 is classified as logic processing. The process P3 is deleted in the process P. Then, as shown in FIG. 5C, the ratio of the number of execution times of each module to the number of execution times of the entire process of the task T5 is calculated as a weighting factor r1 to rn.

  The weight calculation means 612 calculates the processing performance conversion coefficient based on a value obtained by multiplying the processing performance evaluation coefficient by the weight coefficient calculated by the weight coefficient calculation means 611. In other words, the processing performance conversion coefficient is a coefficient obtained by weighting the processing performance evaluation coefficient for the model program obtained in advance by the module characteristics constituting the task of the evaluation target program, that is, the module execution load ratio.

  More specifically, as shown in FIG. 3, the processing performance evaluation coefficient (for example, a value obtained by dividing the first processing result T31 by the second processing result T41) and a weighting coefficient (for example, the processing A) for the same module (for example, processing A). For example, a value obtained by multiplying a value obtained by multiplying r1) for all modules included in the task is calculated as the processing performance conversion coefficient.

  The processing performance conversion processing means 62 is configured to calculate the processing speed in the second processing device 40 based on the calculated processing performance conversion coefficient, specifically, the first processing device 30. A value obtained by multiplying the processing speed when the evaluation target program is incorporated into the processing performance conversion coefficient is calculated as the processing speed of the second processing device 40.

  The performance evaluation means 63 is configured to determine whether or not the calculated processing speed satisfies an allowable processing speed that is a processing speed required when the processing speed is actually mounted on the vehicle. More specifically, the maximum speed at which the second processing device 40 can operate is registered in the performance evaluation unit 63 as an allowable processing speed, and the processing speed calculated in the processing performance conversion processing unit 62 is registered. When the calculated processing speed is lower than the allowable processing speed compared with the processing speed, the second processing device 40 cannot operate the evaluation target program at a speed that satisfies the allowable processing speed. It is comprised so that the result of being may be output.

  The evaluation content changing means 64 is configured to change the processing performance conversion coefficient when the calculated processing speed is evaluated by the performance evaluation means 63 as not satisfying an allowable processing speed. More specifically, a new processing performance conversion coefficient is derived from a weighting factor that is changed based on a pre-set rule, for example, a rule that increases or decreases each weighting factor by a certain value to satisfy an allowable processing speed. is there.

  Furthermore, the evaluation content changing means 64 changes individual weighting factors by automatically changing a processing speed when the evaluation target program is incorporated into the first processing device or a part of the evaluation target program. You may comprise as follows.

  As an example, the evaluation content changing means 64 is changed so as to reduce the precision of the floating point arithmetic processing included in the evaluation target program (such as reducing the number of bytes), and the evaluation target program associated with this change is processed in the first process. A case where the weighting coefficient of each module in the evaluation target program is changed by changing the processing speed when incorporated in the apparatus is considered.

  Thereafter, the operator (program evaluator) seems to satisfy the conditions such as the evaluated processing performance conversion coefficient and the weighting coefficient for the evaluation target program determined to satisfy the allowable processing speed as a result of the change. While paying attention to the above, work such as concrete correction and review of performance (specifications) of the mass production ECU is performed.

  Note that the change of the processing performance conversion coefficient or the like in the evaluation content changing means 64 may be changed by manual input by an operator (program evaluator) in addition to the above-described automatic change configuration. . More specifically, the evaluation content changing unit 64 includes an input unit composed of a keyboard and a mouse that can change the weighting coefficient, the processing performance conversion coefficient, and the like by an operator, and the weight input by the input unit. A configuration including a display unit that displays a coefficient, the evaluation target program, the processing performance conversion coefficient, and the like may be used.

  In this case, the operator (program evaluator) examines how the program to be evaluated can be modified, changes the processing performance conversion coefficient, weighting coefficient, etc. according to the contents of the modification, or actually After the evaluation object program is corrected, the evaluation object program satisfying the allowable speed can be created by evaluating the performance evaluation means 63 again.

  The transplantation evaluation apparatus 50 is configured to execute the processing performance conversion processing unit 62 and the evaluation unit 63 based on the processing performance conversion coefficient changed by the evaluation content changing unit 64, and the processing performance The above operation is repeated until the evaluation means 63 or the operator determines that the conversion coefficient is the optimum value.

  In the above description, the transplantation evaluation apparatus 50 has been described with respect to a configuration in which only the processing speed is calculated as the processing performance. However, the processing performance is not limited to the processing speed or in addition to the processing speed. The structure which calculates an element may be sufficient. For example, the element may be configured to calculate a processing time that is an execution time of a module or a program.

  Further, in the above description, the transplantation evaluation apparatus 50 has been described with respect to a configuration for calculating a processing performance evaluation coefficient using a model program as a predetermined program. However, a program other than the model program is used as the predetermined program. The structure to be used may be used. For example, it may be a program other than a program specially prepared for benchmarking such as the model program, and may be an evaluation target program used in the past. However, the programs that are operated by the first processing device 30 and the second processing device 40 are basically configured to use the same program even if they are slightly different due to a compiler or the like.

  Furthermore, in the above description, the configuration for calculating the weighting coefficient by the execution load ratio of each module of the evaluation target program as the weighting by the ratio of each module type of the evaluation target program has been described. A configuration using a ratio other than the execution load ratio may be used as the weighting. For example, a configuration in which the weighting factor is calculated based on a ratio based on the number of each module existing in the evaluation target program, or a configuration in which the weighting factor is calculated based on a ratio based on the total number of lines and characters in the program of each module. It may be.

  As described above, with the configuration including the performance evaluation unit 63 and the evaluation content changing unit 64, the processing performance conversion coefficient is set to an optimum value by changing the parameters and the program through trial and error. It is possible to approach.

  The capacity conversion means 70 calculates the memory capacity in the second processing device 40 from the memory capacity in the first processing device 30 when the software is transferred from the first processing device 30 to the second processing device 40. A model program comprising a plurality of tasks incorporated in the processing unit of the electronic control unit, and a memory capacity in units of modules constituting the task or task executed in the first processing unit, and the second Memory executed by the second processing device 40 from memory capacity executed by the first processing device 30 of the model program based on a task executed by the processing device or a memory capacity of a module unit constituting the task An approximate expression calculating means 71 for calculating an approximate expression for converting the capacity, and a processing device of the electronic control unit. Memory capacity conversion processing means 72 for calculating the memory capacity executed by the second processing device 40 based on the approximate expression from the memory capacity executed by the first processing device 30 for the evaluation target program; Capacity evaluation means 73 for determining whether or not the calculated memory capacity satisfies the allowable memory capacity is provided.

  The approximate expression calculating means 71 is configured to calculate an approximate expression for converting the memory capacity executed by the second processing device 40 from the memory capacity executed by the first processing device 30 of the model program. Has been. The approximate expression is configured to be calculated in three steps as shown below.

  As a first step, the memory capacity when the model program is incorporated into the first processing device 30 and the memory capacity when the model program is incorporated into the second processing device 40 are measured in units of tasks or modules. The memory capacity in this case is the capacity of the model program stored in the ROM and the capacity of the data by the model program stored in the RAM generated by being expanded in the RAM when the model program is executed. Including both.

  As a second step, a map of the measured memory capacity of each task or module is created. The map is shown in FIG. 6 with the memory capacity when the model program is incorporated into the first processing device 30 and the memory capacity when the model program is incorporated into the second processing device 40 as coordinate axes. It is created on such a two-dimensional coordinate plane. Then, a map is created by plotting the memory capacity of the task or module in the second processing device 40 against the memory capacity of the task or module in the first processing device 30 on the two-dimensional coordinate plane. As shown in FIG. 6, the map is created so that each point indicated by a black dot (one of which is indicated by B) corresponds to the memory capacity of each task or each module.

As a third step, an approximate expression f is derived from the distribution of the black spots on the map. More specifically, in the model program composed of i types of tasks or modules, the memory capacity of the tasks or modules in the first processing device 30 is x _n (n: an integer from 1 to i), and the second processing device. A plurality of data (x _n , y _n ) in which the memory capacity of the task or module in 40 is y _n (n: an integer from 1 to i) is _expressed by a linear expression of y = a + bx or y = ax + bx ² by the least square method. A known method such as a method of deriving the coefficients a and b for approximating the quadratic equation or the like is used.

  Based on the calculated approximate expression, the memory capacity conversion processing means 72 performs the first calculation from the memory capacity executed by the first processing device 30 on the evaluation target program incorporated in the processing device of the electronic control unit. The memory capacity executed by the two processing devices 40 is calculated. More specifically, the memory capacity when the evaluation target program is incorporated into the first processing device 30 is expressed by the variable of the approximate expression ( By substituting for x or y), the memory capacity (y or x) when the evaluation object program is incorporated into the second processing device 40 is calculated. That is, in the approximate expression f drawn on the two-dimensional coordinate plane in FIG. 6, when the memory capacity when the evaluation target program is incorporated in the first processing device 30 is applied as the value of the vertical axis, A value on the horizontal axis corresponding to the value of the axis is calculated as a memory capacity when the evaluation target program is incorporated into the second processing device 40.

  The capacity evaluation unit 73 is configured to determine whether or not the calculated memory capacity satisfies an allowable memory capacity. More specifically, the maximum memory capacity that can be mounted on the second processing device 40 is registered in the capacity evaluation means 73 as an allowable memory capacity, and the memory capacity calculated in the memory capacity conversion processing means 72 is registered. When the calculated memory capacity is larger than the allowable memory capacity compared to the allowable memory capacity, it is impossible to operate the second processing device 40 with the calculated memory capacity. It is configured to output.

  The evaluation content changing unit 74 is configured to change the evaluation target program when the capacity evaluation unit 73 evaluates that the allowable memory capacity is not satisfied. More specifically, for the program to be evaluated, a predetermined rule, for example, a specific module is deleted in order to satisfy an allowable memory capacity, or an address range for accessing the RAM is substantially limited. The memory size is reduced, and the capacity evaluation means 73 is configured to automatically change the contents so that the memory capacity determined to satisfy the allowable memory capacity is calculated. The operator (program evaluator) performs operations such as further modification of the evaluation target program and review of performance (specifications) of the mass production ECU based on the changed evaluation target program.

  In addition, the change of the evaluation object program in the evaluation content changing means 74 may be changed by manual input by an operator (program evaluator) in addition to the above-described automatic change configuration. More specifically, the evaluation content changing means 74 includes an input unit composed of a keyboard and a mouse that can change the evaluation target program by an operator, the evaluation target program input by the input unit, and the evaluation A configuration may be provided that includes a display unit that displays a calculated memory capacity or the like based on the target program.

  In this case, the operator (program evaluator) examines how the evaluation target program can be corrected, changes the configuration of the microcomputer according to the correction contents, or actually evaluates the program to be evaluated. After correcting the above, by evaluating again with the capacity evaluation means 73, a program satisfying the allowable memory capacity can be created.

  As described above, by adopting the configuration including the capacity evaluation means 73, the memory capacity can be appropriately within the allowable range by changing the configuration of the microcomputer and the evaluation target program through trial and error. It becomes possible.

  Hereinafter, the evaluation of the processing speed and the memory capacity in the transplantation evaluation apparatus 50 will be described based on the flowcharts shown in FIGS.

  First, before evaluating a target evaluation target program, an approximate expression used for evaluation and a processing performance evaluation coefficient are calculated.

  The memory capacity of the module unit of the model program as benchmark software incorporated in the first processing device 30 is measured by the first measuring device 31 (SA1), and the module unit of the model program incorporated in the second processing device 40 Is measured by the second measuring device 41 (SA2), and the approximate expression calculating means 71 calculates an approximate expression based on the measured memory capacities (SA3).

  The module unit processing speed of the model program incorporated in the first processor 30 is measured by the first measuring device 31 (SB1), and the model program module unit incorporated in the second processor 40 is measured in module units. The processing speed is measured by the second measuring device 41 (SB2), and the processing performance evaluation coefficient calculation means 610 calculates a processing performance evaluation coefficient based on the measured processing speed (SB3).

  7A and 7B, the benchmark program prepared in advance is actually operated on the first processor and the second processor before the evaluation target program is evaluated. In this case, the approximate expression and the processing performance evaluation coefficient are calculated from the processing speed and the memory capacity at that time. Note that the program used at this time may not be a program specially prepared for benchmarking, but may be an evaluation target program evaluated in the past. However, the programs to be operated by the first processing apparatus and the second processing apparatus should basically use the same program even if they are slightly different due to the compiler or the like.

  When advance preparation for evaluation is completed, the evaluation target program is evaluated next.

  An evaluation target program as development software is incorporated into the first processing device 30, and the processing speed and memory capacity of the evaluation target program in the first processing device 30 are measured by the first measuring device 31 (SC1).

  The capacity conversion will be described below. The memory capacity conversion processing means 72 calculates the memory capacity when the evaluation target program is incorporated and executed in the second processing device 40 by applying the approximate expression to the memory capacity measured in step SC1. (SC2).

  The capacity evaluation unit 73 determines whether or not the calculated memory capacity satisfies a pre-registered allowable memory capacity prescribed value (SC3). If the prescribed value of the allowable memory capacity is satisfied, porting of the evaluation target program to the second processing device 40 is executed (SC10). On the other hand, when the prescribed value of the allowable memory capacity is not satisfied, the operator adjusts the evaluation target program by changing the content of the evaluation target program (SC4). Then, the memory capacity is calculated again using the adjusted evaluation target program. This process is repeated until the capacity evaluation unit 73 determines that the calculated memory capacity satisfies the specified value of the allowable memory capacity.

  The speed conversion will be described below. The weighting factor calculating unit 611 calculates a weighting factor based on the execution load ratio of each module of the evaluation target program (SC5), and the weighting calculating unit 612 performs the processing on the weighting factor calculated by the weighting factor calculating unit 611. A processing performance conversion coefficient is calculated based on the value multiplied by the performance evaluation coefficient (SC6). Then, the processing performance conversion processing means 62 applies the processing performance conversion coefficient to the processing speed measured in step SC1, thereby executing the evaluation target program in the second processing device 40 when it is executed. The processing speed is calculated (SC7).

  The performance evaluation unit 63 determines whether or not the calculated processing speed satisfies a pre-registered prescribed value of the allowable processing speed (SC8). If the prescribed value of the allowable processing speed is satisfied, the program to be evaluated is transplanted to the second processing device 40 (SC10). On the other hand, when the specified value of the processing speed is not satisfied, the evaluation content changing unit 64 automatically changes the processing performance conversion coefficient or the like, or the operator manually changes the processing performance conversion coefficient or the like. Then, the evaluation object program is adjusted (SC9). Then, the processing speed is calculated again using the adjusted evaluation target program. This process is repeated until the performance evaluation unit 63 determines that the calculated processing speed satisfies the prescribed value of the allowable processing speed.

  Hereinafter, another embodiment will be described. In the above-described embodiment, the processing performance conversion coefficient calculation means 61 calculates the weighting coefficient based on the execution load ratio of each module in the evaluation target program, and applies the calculated weighting coefficient to each task, thereby converting the processing performance. Although the configuration in which the coefficient is calculated for each task has been described, the processing performance conversion coefficient may be calculated for each evaluation target program other than the task, for example.

  More specifically, the processing performance conversion coefficient calculating means 61 is configured to calculate a second weighting coefficient based on the execution load ratio of the task constituting the evaluation target program in addition to the weighting coefficient based on the module execution load ratio. And the structure which calculates the said processing performance conversion coefficient for every said evaluation object program by applying the calculated 2nd weighting coefficient to the said evaluation object program may be sufficient.

  In the above-described embodiment, the processing performance evaluation coefficient calculation unit 610 and the approximate expression calculation unit 71 have been described as being included in the transplantation evaluation apparatus 50. However, the processing performance evaluation coefficient calculation unit 610 and the approximate expression are described. The calculation means 71 may be configured not included in the transplantation evaluation apparatus 50.

  For example, the processing performance evaluation coefficient calculation unit 610 and the approximate expression calculation unit 71 may be included in the first measurement device 31 and the second measurement device 41.

  In the above-described embodiment, the configuration has been described in which the weighting factor calculation unit 611 calculates the ratio of the number of execution times of each module to the number of executions of the entire task processing as a weighting factor. It may be a configuration. By adopting such a configuration, it is possible to calculate a weighting coefficient by placing a weight on a module that has a small execution time but a long execution time.

  The transplantation evaluation apparatus 50 according to the present invention described above has the processing performance conversion coefficient calculation means 61 as a processing performance conversion coefficient calculation step, the processing performance conversion processing means 62 as a processing performance conversion processing step, and the performance evaluation means 63 as a speed evaluation step. As the evaluation content changing means 64, the approximate expression calculating means 71 as the approximate expression calculating step, the memory capacity conversion processing means 72 as the memory capacity conversion processing step, and the capacity evaluating means 73 as the capacity evaluation step. It is also realized as a transplant evaluation method.

  In the embodiment described above, the software resource transplantation evaluation apparatus has been described for evaluating software incorporated in the processing device of the electronic control unit. However, the transplantation evaluation apparatus according to the present invention is an electronic control unit mounted on a vehicle. For example, the present invention can be widely applied to software for an electronic control unit of an OA device such as a printer.

  Note that the above-described embodiment is merely an example of the present invention, and it is needless to say that the specific configuration and the like of each block can be changed and designed as appropriate within the scope of the effects of the present invention.

Block diagram of multiple electronic control units Block diagram of transplant evaluation device Illustration of processing performance conversion factor calculation Illustration of weighting factor calculation (A) shows the unexecuted process elimination process in (b), shows the classification for each module type of the process, (c) is an explanatory diagram showing the calculation of the weighting coefficient Illustration of graph for calculating approximate expression Approximation formula and processing performance evaluation coefficient calculation flowchart Flow chart for explaining operation of transplantation evaluation apparatus

Explanation of symbols

30: first processing device 31: first measuring device 40: second processing device 41: second measuring device 50: transplantation evaluation device 61: processing performance conversion coefficient calculation means 62: processing performance conversion processing means 63: performance evaluation means 64 : Evaluation content changing means 72: Memory capacity conversion processing means 73: Capacity evaluation means

Claims (7)

A software porting evaluation device for calculating the processing performance of the second processing device from the processing performance of the first processing device of the evaluation target program composed of a plurality of tasks incorporated in the processing device performing electronic control,
Based on the processing performance conversion coefficient obtained by weighting the processing performance evaluation coefficient obtained from the processing performance in the first processing apparatus and the second processing apparatus of a predetermined program by the module characteristics constituting the task of the evaluation target program A software porting evaluation apparatus comprising processing performance conversion processing means for calculating the processing performance.   The processing performance conversion processing means weights the processing performance evaluation coefficient obtained for each module type constituting the task of the predetermined program by the ratio of each module type constituting the task of the evaluation target program, The software porting evaluation apparatus according to claim 1, which calculates a processing performance conversion coefficient.   A processing performance conversion coefficient calculation that calculates a processing performance conversion coefficient for converting the processing performance in the second processing device from the processing performance in the first processing device of the evaluation target program composed of a plurality of tasks incorporated in the processing device that performs electronic control. The processing performance conversion coefficient calculating means is a processing performance determined for each module type constituting a task in the first processing device and the second processing device of a predetermined program comprising a plurality of tasks. A software porting evaluation apparatus that calculates the processing performance conversion coefficient by weighting the processing performance evaluation coefficient, which is a ratio of the above, by the ratio of each module type constituting the task of the evaluation target program. A software porting evaluation device for calculating the memory capacity used in the second processing device from the memory capacity used in the first processing device of the evaluation target program composed of a plurality of tasks incorporated in the processing device performing electronic control There,
An approximate expression for converting the memory capacity used in the second processing device from the memory capacity for each module type constituting the task or the task used in the first processing device and the second processing device of a predetermined program An approximate expression calculating means for calculating, and a memory capacity for calculating a memory capacity used in the second processing device based on the approximate expression from a memory capacity used in the first processing device for the evaluation target program A software porting evaluation apparatus comprising conversion processing means. A software porting evaluation method for calculating the processing performance in the second processing device from the processing performance in the first processing device of the evaluation target program composed of a plurality of tasks incorporated in the processing device performing electronic control,
Based on the processing performance conversion coefficient obtained by weighting the processing performance evaluation coefficient obtained from the processing performance in the first processing apparatus and the second processing apparatus of a predetermined program by the module characteristics constituting the task of the evaluation target program A processing performance conversion processing step for calculating the processing performance;
A performance evaluation step for determining whether or not the calculated processing performance satisfies the allowable processing performance, and an evaluation for changing the processing performance conversion coefficient when it is evaluated that the allowable processing performance is not satisfied in the performance evaluation step Content change step;
A software porting evaluation method for executing the processing performance conversion processing step and the evaluation step based on the changed processing performance conversion coefficient.   The processing performance conversion processing step weights the processing performance evaluation coefficient obtained for each module type constituting the task of the predetermined program by the ratio of each module type constituting the task of the evaluation target program, 6. The software porting evaluation method according to claim 5, wherein the processing performance conversion coefficient is calculated, and the processing performance conversion coefficient is changed by changing the weight. A software porting evaluation method for calculating the memory capacity used in the second processing device from the memory capacity used in the first processing device of the evaluation target program composed of a plurality of tasks incorporated in the processing device performing electronic control. There,
An approximate expression for converting the memory capacity used in the second processing device from the memory capacity for each module type constituting the task or the task used in the first processing device and the second processing device of a predetermined program A memory capacity for calculating a memory capacity used in the second processing apparatus from a memory capacity used in the first processing apparatus based on the approximate expression for an approximate expression calculating step to calculate and a program to be evaluated A conversion processing step, a capacity evaluation step for determining whether or not the calculated memory capacity satisfies an allowable memory capacity, and the evaluation object when it is evaluated that the allowable memory capacity is not satisfied in the capacity evaluation step A software porting evaluation method comprising an evaluation content changing step for changing a program.

Images