JP7040272B2 - Source code conversion program and source code conversion method - Google Patents

Description

The present invention relates to a source code conversion program and a source code conversion method.

In recent years, due to advances in circuit synthesis technology such as high-level synthesis, hardware design may be performed in a general-purpose software language such as C or C ++. Therefore, it has been proposed to offload an application running on a CPU (Central Processing Unit) to hardware such as FPGA (Field Programmable Gate Array) or GPU (Graphics Processing Unit) for speeding up. There is. However, many applications are written for the CPU, and it is difficult to bring out sufficient hardware performance with the description as it is.

Therefore, it is required to design the source code of the application by classifying it into the part corresponding to the algorithm and the part corresponding to the hardware architecture, and rewrite these into the offload description. This rewriting to the offload description is difficult to automate because there are many changes in the data structure and control structures such as function calls and repetitions, and it is common to manually perform a large amount of man-hours.

International Publication No. 2017/05/6427 Special Table 2014-509106 Gazette

On the other hand, the source code of an application tends to be updated frequently due to version upgrades and bug fixes. Source code updates are often provided with source code diffs called patches. However, it is difficult to apply a patch for source code to an offload description rewritten according to the hardware. Therefore, the rewriting of the offload description corresponding to the updated source code is performed manually, and if the application is frequently updated, the man-hours required for rewriting will increase.

One aspect is to provide a source code conversion program and a source code conversion method that can improve the update efficiency of the corresponding offload description when the application is updated.

In one embodiment, the source code conversion program is based on the application having the source code, and indicates the insertion point of the algorithm information of the application corresponding to the offload description operated by the hardware with a tag and the architecture of the hardware. Causes the computer to execute the process of acquiring the offload description template that represents. When the source code is updated, the source code conversion program causes a computer to execute a process of extracting algorithm information corresponding to the tag from the source code reflecting the update. The source code conversion program inserts the extracted algorithm information into the corresponding portion of the offload description template and causes the computer to execute a process of updating the offload description.

It is possible to improve the update efficiency of the corresponding offload description when the application is updated.

FIG. 1 is a block diagram showing an example of the configuration of the conversion device of the embodiment. FIG. 2 is a diagram showing an example of source code and offload description. FIG. 3 is a diagram showing an example of algorithm information and an offload description template. FIG. 4 is a diagram showing an example of inserting an extraction tag. FIG. 5 is a diagram showing an example of a tag insertion patch. FIG. 6 is a diagram showing an example of algorithm information. FIG. 7 is a diagram showing an example of inserting glue logic. FIG. 8 is a diagram showing an example of an offload description template. FIG. 9 is a diagram showing an example of generation of an offload description. FIG. 10 is a diagram showing an example of the output of the patch program when the offload description template needs to be updated. FIG. 11 is a diagram showing an example of the difference in the abstract syntax tree. FIG. 12 is a diagram showing an example of a case where algorithm information is extracted using an abstract syntax tree. FIG. 13 is a diagram showing an example of programming language conversion. FIG. 14 is a flowchart showing an example of the initial extraction process of the embodiment. FIG. 15 is a flowchart showing an example of the conversion process of the embodiment. FIG. 16 is a diagram showing an example of a computer that executes a source code conversion program.

Hereinafter, examples of the source code conversion program and the source code conversion method disclosed in the present application will be described in detail with reference to the drawings. The disclosed technology is not limited by the present embodiment. In addition, the following examples may be appropriately combined as long as they do not contradict each other.

FIG. 1 is a block diagram showing an example of the configuration of the conversion device of the embodiment. The conversion device 100 shown in FIG. 1 is an example of an information processing device that performs a process of converting a source code of an application (hereinafter, also referred to as an application) into a hardware description. The conversion device 100 acquires an offload description template corresponding to the offload description to be operated by the hardware based on the application having the source code. The offload description template indicates the insertion point of the algorithm information of the application with a tag and represents the hardware architecture. When the source code is updated, the conversion device 100 extracts the algorithm information corresponding to the tag from the source code reflecting the update. The conversion device 100 inserts the extracted algorithm information into the corresponding portion of the offload description template to update the offload description. As a result, the conversion device 100 can improve the update efficiency of the corresponding offload description when the application is updated.

First, the source code and the offload description will be described with reference to FIGS. 2 and 3. FIG. 2 is a diagram showing an example of source code and offload description. The source code 10 and the offload description 11 shown in FIG. 2 perform the same processing. The source code 10 is described in C, and the offload description 11 is described in SystemC. Further, the offload description 11 is a program implemented in the target hardware or a compileable source code of circuit data.

Since the offload description 11 expresses the architecture on the hardware, the points shown in the descriptions 12 to 14 are different from the source code 10. Description 12 is a description relating to input / output, and is a pointer variable in the source code 10 and a FIFO (First In First Out) variable in the offload description 11. The description 13 is a description relating to the loop structure, and is three for statements in the source code 10 and one while statement in the offload description 11. The description 14 is a description relating to the variable name and the access to the variable, and is derived from the descriptions 12 and 13.

FIG. 3 is a diagram showing an example of algorithm information and an offload description template. The offload description 15 shown in FIG. 3 can be classified into the algorithm information 16 and the offload description template 17. The algorithm information 16 indicates a part of the offload description 15 where an operation is performed on the data. The offload description template 17 indicates in the offload description 15 a control statement including input / output and memory declaration statements, input / output and memory access, and input / output and memory access. The control statement includes, for example, a loop such as a for statement, a while statement, and a until statement, and a conditional branch such as an if statement, an else statement, and a switch statement.

Next, the configuration of the conversion device 100 will be described. As shown in FIG. 1, the conversion device 100 includes a communication unit 110, a display unit 111, an operation unit 112, a storage unit 120, and a control unit 130. In addition to the functional units shown in FIG. 1, the conversion device 100 may have various functional units of known computers, such as various input devices and audio output devices.

The communication unit 110 is realized by, for example, a NIC (Network Interface Card) or the like. The communication unit 110 is a communication interface that is connected to another information processing device by wire or wirelessly via a network (not shown) and controls information communication with the other information processing device. The communication unit 110 receives, for example, a source code update patch from another terminal. Further, the communication unit 110 transmits the updated offload description to another terminal.

The display unit 111 is a display device for displaying various information. The display unit 111 is realized by, for example, a liquid crystal display or the like as a display device. The display unit 111 displays various screens such as a display screen input from the control unit 130.

The operation unit 112 is an input device that receives various operations from the user of the conversion device 100. The operation unit 112 is realized by, for example, a keyboard, a mouse, or the like as an input device. The operation unit 112 outputs the operation input by the user to the control unit 130 as operation information. The operation unit 112 may be realized by a touch panel or the like as an input device, or the display device of the display unit 111 and the input device of the operation unit 112 may be integrated.

The storage unit 120 is realized by, for example, a semiconductor memory element such as a RAM (Random Access Memory) or a flash memory (Flash Memory), or a storage device such as a hard disk or an optical disk. The storage unit 120 includes a source code storage unit 121, a tag insertion patch storage unit 122, a template storage unit 123, an algorithm information storage unit 124, and an offload description storage unit 125. Further, the storage unit 120 stores information used for processing in the control unit 130.

The source code storage unit 121 stores the source code of the application. Further, the source code storage unit 121 stores the source code updated by the source code update patch.

The tag insertion patch storage unit 122 stores a tag insertion patch for inserting an extraction tag that specifies a portion in which a source code algorithm is described. The first tag insertion patch is generated based on the source code of the first edition and the source code with the tag inserted by manually inserting the extraction tag into the source code of the first edition. Further, the source code of the first edition means that the offload description is generated first, and the source code of the first edition may be any version of the application itself.

The template storage unit 123 stores the offload description template. The offload description template indicates the insertion point of the algorithm information of the application with a tag and represents the architecture of the hardware.

The algorithm information storage unit 124 stores the algorithm information obtained by extracting the algorithm portion of the source code.

The offload description storage unit 125 stores the offload description generated based on the offload description template and the algorithm information.

The control unit 130 is realized by, for example, using a CPU, an MPU (Micro Processing Unit), or the like to execute a program stored in an internal storage device using the RAM as a work area. Further, the control unit 130 may be realized by, for example, an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA. The control unit 130 has an acquisition unit 131, an extraction unit 132, and an update unit 133, and realizes or executes the functions and operations of information processing described below. The internal configuration of the control unit 130 is not limited to the configuration shown in FIG. 1, and may be any other configuration as long as it is configured to perform information processing described later.

The acquisition unit 131 acquires the first edition source code, the tag-inserted source code, and the offload description template from another terminal, for example, via the communication unit 110. The acquisition unit 131 stores the acquired first edition source code and the offload description template in the source code storage unit 121 and the template storage unit 123, respectively. The tag-inserted source code acquired here is the source code of the first edition with the extraction tag manually inserted. In addition, the offload description template acquired here describes a tag indicating the insertion point instead of the algorithm information for the offload description manually generated from the source code of the first edition.

The acquisition unit 131 generates a tag insertion patch based on the source code of the first edition and the source code in which the tag has been inserted. The acquisition unit 131 stores the generated tag insertion patch in the tag insertion patch storage unit 122.

The extraction unit 132 acquires a source code update patch from another terminal, for example, via the communication unit 110. The extraction unit 132 reads the source code to be patched from the source code storage unit 121, and applies the source code update patch to the read source code. That is, the extraction unit 132 updates the source code based on the source code update patch. The extraction unit 132 stores the updated source code in the source code storage unit 121.

The extraction unit 132 refers to the tag insertion patch storage unit 122, and applies the tag insertion patch to the updated source code. Since the tag insertion patch is difference information, the tag insertion location can be specified with high probability even if the source code is updated. The extraction unit 132 determines whether or not the tag insertion patch has been applied. When the extraction unit 132 determines that the tag insertion patch cannot be applied, it outputs a message to the effect that the offload description template needs to be updated and displays it on the display unit 111.

On the other hand, when it is determined that the tag insertion patch can be applied, the extraction unit 132 generates the tag-inserted source code. Further, the extraction unit 132 outputs a message to the effect that the offload description template does not need to be updated to the display unit 111 and displays it. The extraction unit 132 may generate the tag-inserted source code by using the abstract syntax tree of the algorithm information of the source code instead of the tag insertion patch. In the patch, it may not be possible to use it unless the source code line, indentation, argument name, etc. match. On the other hand, in the abstract syntax tree, it is possible to extract the algorithm information and generate the tagged source code even if the description is slightly changed.

The extraction unit 132 extracts the algorithm information from the generated source code with the tag inserted. That is, the extraction unit 132 extracts the algorithm information by extracting the part surrounded by the tag in the tag-inserted source code. The extraction unit 132 stores the extracted algorithm information in the algorithm information storage unit 124. When the extraction unit 132 stores the algorithm information in the algorithm information storage unit 124, the extraction unit 132 outputs an update instruction to the update unit 133.

In other words, when the source code is updated, the extraction unit 132 extracts the algorithm information corresponding to the tag from the source code reflecting the update. Further, when the source code is updated, the extraction unit 132 inserts a tag for extracting algorithm information into the source code reflecting the update, and extracts the algorithm information corresponding to the tag. Further, the extraction unit 132 inserts the extraction tag by using a patch for inserting the extraction tag of the algorithm information. Further, the extraction unit 132 inserts an extraction tag using an abstract syntax tree of algorithm information. Further, the extraction unit 132 notifies that the offload description template needs to be updated when there is a part in the source code reflecting the update where the tag for extracting the algorithm information cannot be inserted.

When the update instruction is input from the extraction unit 132, the update unit 133 reads the offload description template and the algorithm information from the template storage unit 123 and the algorithm information storage unit 124, respectively. The update unit 133 inserts the algorithm information into the tag portion of the read offload description template to generate the offload description. The update unit 133 stores the generated offload description in the offload description storage unit 125.

Here, a specific example of the conversion from the source code to the offload description will be described with reference to FIGS. 4 to 9. FIG. 4 is a diagram showing an example of inserting an extraction tag. In the source code 10 shown in FIG. 4, line numbers L6, L9, and L12 are assumed to be algorithm parts. When the source code 10 is the source code of the first edition, the tag for extracting the algorithm information is manually inserted, and the tag-inserted source code 18 is generated. When the source code 10 is the source code of the second edition or later, the extraction unit 132 generates the tag-inserted source code 18 by applying the tag insertion patch.

The tag for extracting algorithm information is enclosed in a function pair for each part of the algorithm. That is, the extraction tag is expressed by, for example, calling the function pair of "// _OFFLOAD_BEGIN ()" and "// _ OFFLOAD_END ()". Further, in order to identify the part of the algorithm part, IDs such as "offload_1", "offload_2", and "offload_3" are given to the argument of the function pair.

FIG. 5 is a diagram showing an example of a tag insertion patch. The tag insertion patch 19 shown in FIG. 5 can be generated by obtaining the difference between the source code of the first edition and the source code in which the tag has been inserted by a diff program. Once the tag insertion patch 19 which is the difference information is generated, the extraction unit 132 corresponds to the second and subsequent editions by applying the tag insertion patch 19 to the source code of the second and subsequent editions using a patch program. You can generate source code with tags inserted.

FIG. 6 is a diagram showing an example of algorithm information. The algorithm information 20 shown in FIG. 6 corresponds the portion surrounded by the function pair of the extraction tag from the tag-inserted source code with the ID of the extraction tag. The extraction unit 132 is, for example, from the tag-inserted source code 18 shown in FIG. 4, “c [i] = a” surrounded by the function pairs of “// _OFFLOAD_BEGIN (offload_1)” and “// _OFFLOAD_END (offload_1)”. [i] + b [i]; "is extracted as an algorithm. For example, as shown in the algorithm information 20, the extraction unit 132 extracts the algorithm "c [i] = a [" following the tag "// _OFFLOAD_BODY (offload_1)" that takes an ID for identifying the part of the algorithm part as an argument. Copy "i] + b [i];". The extraction unit 132 similarly extracts the algorithm for the IDs “offload_2” and “offload_3” and copies them in association with the tag to generate the algorithm information 20.

FIG. 7 is a diagram showing an example of inserting glue logic. The offload description 11 shown in FIG. 7 is a raw offload description immediately after the rewriting work is manually performed from the source code 10. On the other hand, since the algorithm information 20 is extracted from the source code 10, there is no one-to-one correspondence with the offload description 11 regarding the description of the algorithm part. Therefore, a glue logic for bridging is inserted so that the description of the algorithm information 20 can be directly inserted into the offload description by manual work. The offload description 21 is consistent with the algorithm information 20 by inserting glue logic into the offload description 11.

Glue logic insertion involves, for example, three steps: (1) defining an array with the number of elements "1", (2) defining a loop variable, and (3) replacing it with an input / output (FIFO) access function. In the example of FIG. 7, as (1), a [1], b [1], c [1], d [1], and x [1] are defined in the offload description 11. As (2), i = 0, j = 0, and k = 0 are defined in the offload description 11. As (3), the access method of the FIFO of the offload description 11 is replaced with afifo.read (), bfifo.read (), xfifo.write (). Glue logic itself is likely to be deleted by the optimization function of the high-level synthesis tool, and it is considered that it is unlikely to cause performance deterioration.

FIG. 8 is a diagram showing an example of an offload description template. The offload description template 22 shown in FIG. 8 is obtained by deleting the algorithm description from the offload description 21 by leaving the tag “// _OFFLOAD_BODY ()”. That is, the description other than the tag "// _OFFLOAD_BODY ()" of the offload description 21 is a description representing the hardware architecture.

FIG. 9 is a diagram showing an example of generation of an offload description. As shown in FIG. 9, the update unit 133 inserts the algorithm description of the corresponding ID of the algorithm information 20 into each tag “// _OFFLOAD_BODY ()” of the offload description template 22, and the offload description 23. To generate. In the explanation of FIG. 9, since the algorithm information 20 extracted from the source code 10 is used, the offload description 23 has the same description as the offload description 21, but the algorithm information based on the updated source code is used. If used, it will be an offload description to which the update has been applied.

FIG. 10 is a diagram showing an example of the output of the patch program when the offload description template needs to be updated. The message 24 shown in FIG. 10 is an example of a message to be output when the extraction unit 132 determines that the tag insertion patch cannot be applied. The message 24 is a message when the patch program fails, and is a warning message including "FAIRLED". When the extraction unit 132 determines that the tag insertion patch cannot be applied, the message 24 may be displayed, or a message such as "The offload description template needs to be updated" is displayed based on the message 24. You may try to do it. When the extraction unit 132 determines that the tag insertion patch can be applied, it displays a message such as "It is not necessary to update the offload description template."

Subsequently, the case where the abstract syntax tree of the algorithm information of the source code is used instead of the tag insertion patch will be described with reference to FIGS. 11 to 13. FIG. 11 is a diagram showing an example of the difference in the abstract syntax tree. The source code 25 shown in FIG. 11 is described in C, and the abstract syntax tree 26 corresponds to the source code 25. Further, the source code 27 is a source code for updating the source code 25. The abstract syntax tree 28 corresponds to the source code 27. Comparing the source code 25 and the source code 27, the operator "+" is updated to "*". On the other hand, when the abstract syntax tree 26 and the abstract syntax tree 28 are compared, it can be seen that the description 29a relating to the operator is updated to the description 29b. That is, the descriptions 29a and 29b are the differences in the abstract syntax tree.

FIG. 12 is a diagram showing an example of a case where algorithm information is extracted using an abstract syntax tree. In the example of FIG. 12, the extraction unit 132 converts the source code 30 described in C into an abstract syntax tree 31 and searches for the description 32 corresponding to the algorithm. The description 32 corresponds to the assignment statement to the variable x in the function main of the source code 30. The extraction unit 132 extracts the algorithm information 33 based on the description 32. In the algorithm information 33, the tag is omitted. The extraction unit 132 generates the tag-inserted source code based on the extracted algorithm information 33. In this way, when the abstract syntax tree is used, the algorithm part can be specified not in the line unit like the patch but in the grammar unit of the programming language. Further, the extraction unit 132 can also perform programming language conversion by using an abstract syntax tree.

FIG. 13 is a diagram showing an example of programming language conversion. FIG. 13 is an example of converting the FORTRAN source code 34 into the algorithm information 37 described in C. The extraction unit 132 converts the source code 34 into an abstract syntax tree 35 and searches for the description 36 corresponding to the algorithm. The description 36 corresponds to the assignment statement to the variable x in the function main of the source code 34. The extraction unit 132 extracts the algorithm information 37 based on the description 36. The extraction unit 132 can perform language conversion from FORTRAN to C by inserting the algorithm information 37 into the template of the source code described in C. The extraction unit 132 may generate an offload description based on the algorithm information 37.

Next, the operation of the conversion device 100 of the embodiment will be described. First, the initial extraction process will be described with reference to FIG. FIG. 14 is a flowchart showing an example of the initial extraction process of the embodiment.

The acquisition unit 131 acquires, for example, the source code of the first edition, the source code with the tag inserted, and the offload description template from another terminal (step S1). The acquisition unit 131 stores the acquired first edition source code and the offload description template in the source code storage unit 121 and the template storage unit 123, respectively.

The acquisition unit 131 generates a tag insertion patch based on the source code of the first edition and the source code in which the tag has been inserted (step S2). The acquisition unit 131 stores the generated tag insertion patch in the tag insertion patch storage unit 122.

The extraction unit 132 extracts the algorithm information from the tag-inserted source code (step S3). The extraction unit 132 stores the extracted algorithm information in the algorithm information storage unit 124. As a result, the conversion device 100 can store the source code of the first edition, the tag insertion patch, the offload description template, and the algorithm information, and prepare for the source code conversion processing of the second and subsequent editions. That is, the conversion device 100 can separately manage the algorithm information of the application and the offload description template representing the architecture on the hardware implementation.

Subsequently, the conversion process when the source code is updated will be described with reference to FIG. FIG. 15 is a flowchart showing an example of the conversion process of the embodiment.

The extraction unit 132 acquires, for example, a source code update patch from another terminal (step S11). The extraction unit 132 updates the source code based on the source code update patch (step S12). The extraction unit 132 stores the updated source code in the source code storage unit 121.

The extraction unit 132 refers to the tag insertion patch storage unit 122, and applies the tag insertion patch to the updated source code (step S13). The extraction unit 132 determines whether or not the tag insertion patch has been applied (step S14). When the extraction unit 132 determines that the tag insertion patch can be applied (step S14: affirmative), the extraction unit 132 generates the tag-inserted source code (step S15).

The extraction unit 132 outputs a message to the effect that the offload description template does not need to be updated to the display unit 111 and displays it (step S16). The extraction unit 132 extracts the algorithm information from the generated source code with the tag inserted (step S17). The extraction unit 132 stores the extracted algorithm information in the algorithm information storage unit 124, and outputs an update instruction to the update unit 133.

When the update instruction is input from the extraction unit 132, the update unit 133 reads the offload description template and the algorithm information from the template storage unit 123 and the algorithm information storage unit 124, respectively. The update unit 133 inserts the algorithm information into the tag portion of the read offload description template to generate the offload description (step S18). The update unit 133 stores the generated offload description in the offload description storage unit 125, and ends the conversion process.

On the other hand, when the extraction unit 132 determines in step S14 that the tag insertion patch cannot be applied (step S14: negation), the extraction unit 132 outputs and displays a message to the effect that the offload description template needs to be updated to the display unit 111. (Step S19), and the conversion process is terminated. As a result, the conversion device 100 can improve the update efficiency of the corresponding offload description when the application is updated. That is, the conversion device 100 can be flexibly updated in response to changes in the application or hardware. Further, the conversion device 100 can easily identify the target location of the updated source code when it is determined that the tag insertion patch cannot be applied. Therefore, the conversion device 100 can easily identify the changed part of the offload description template.

In the above embodiment, the case where the source code of the application is described in C or C ++ and the offload description is described in SystemC has been described, but the present invention is not limited to this. For example, as the source code of the application, source codes of various languages such as FORTRAN, COBOL, Java (registered trademark), Perl, Python, and Ruby can be used. Further, for example, as the offload description, an offload description of various languages such as SystemVerilog can be used.

As described above, the conversion device 100 indicates the insertion point of the algorithm information of the application corresponding to the offload description operated by the hardware based on the application having the source code with a tag and the offload representing the architecture of the hardware. Get the description template. Further, when the source code is updated, the conversion device 100 extracts the algorithm information corresponding to the tag from the source code reflecting the update. Further, the conversion device 100 inserts the extracted algorithm information into the corresponding portion of the offload description template to update the offload description. As a result, the conversion device 100 can improve the update efficiency of the corresponding offload description when the application is updated.

Further, when the source code is updated, the conversion device 100 inserts a tag for extracting algorithm information into the source code reflecting the update, and extracts the algorithm information corresponding to the tag. As a result, the conversion device 100 can acquire the algorithm information corresponding to the updated source code.

Further, the conversion device 100 inserts the extraction tag by using a patch for inserting the extraction tag of the algorithm information. As a result, the conversion device 100 can specify the description regarding the algorithm.

Further, the conversion device 100 inserts an extraction tag using an abstract syntax tree of algorithm information. As a result, the conversion device 100 can specify the description regarding the algorithm.

Further, the conversion device 100 notifies that the offload description template needs to be updated when there is a part in the source code reflecting the update where the tag for extracting the algorithm information cannot be inserted. As a result, the converter 100 can notify the user that there has been an update of a portion related to the hardware architecture.

Further, each component of each of the illustrated parts does not necessarily have to be physically configured as shown in the figure. That is, the specific form of distribution / integration of each part is not limited to the one shown in the figure, and all or part of them may be functionally or physically distributed / integrated in any unit according to various loads and usage conditions. Can be configured. For example, the extraction unit 132 and the update unit 133 may be integrated. Further, the illustrated processes are not limited to the above order, and may be performed simultaneously or in a different order as long as the processing contents do not contradict each other.

Further, the various processing functions performed by each device may be executed on the CPU (or a microcomputer such as an MPU or a MCU (Micro Controller Unit)) in whole or in any part thereof. In addition, various processing functions may be executed in whole or in any part on a program analyzed and executed by a CPU (or a microcomputer such as an MPU or MCU) or on hardware by wired logic. Needless to say, it's good.

By the way, various processes described in each of the above embodiments can be realized by executing a program prepared in advance on a computer. Therefore, in the following, an example of a computer that executes a program having the same functions as those of the above embodiments will be described. FIG. 16 is a diagram showing an example of a computer that executes a source code conversion program.

As shown in FIG. 16, the computer 200 has a CPU 201 that executes various arithmetic processes, an input device 202 that accepts data input, and a monitor 203. Further, the computer 200 includes a medium reading device 204 for reading a program or the like from a storage medium, an interface device 205 for connecting to various devices, and a communication device 206 for connecting to another information processing device or the like by wire or wirelessly. Has. Further, the computer 200 has a RAM 207 for temporarily storing various information and a hard disk device 208. Further, each of the devices 201 to 208 is connected to the bus 209.

The hard disk device 208 stores a source code conversion program having the same functions as the processing units of the acquisition unit 131, the extraction unit 132, and the update unit 133 shown in FIG. Further, the hard disk device 208 is for realizing a source code storage unit 121, a tag insertion patch storage unit 122, a template storage unit 123, an algorithm information storage unit 124, an offload description storage unit 125, and a source code conversion program. Various data are stored. The input device 202 receives, for example, input of various information such as operation information from a user of the computer 200. The monitor 203 displays various screens such as a display screen for the user of the computer 200, for example. For example, a printing device or the like is connected to the interface device 205. The communication device 206 has, for example, the same function as the communication unit 110 shown in FIG. 1 and is connected to a network (not shown) to exchange various information with other information processing devices.

The CPU 201 performs various processes by reading out each program stored in the hard disk device 208, expanding the program in the RAM 207, and executing the program. Further, these programs can make the computer 200 function as the acquisition unit 131, the extraction unit 132, and the update unit 133 shown in FIG.

The above source code conversion program does not necessarily have to be stored in the hard disk device 208. For example, the computer 200 may read and execute a program stored in a storage medium that can be read by the computer 200. The storage medium that can be read by the computer 200 is, for example, a portable recording medium such as a CD-ROM, a DVD (Digital Versatile Disc), or a USB (Universal Serial Bus) memory, a semiconductor memory such as a flash memory, a hard disk drive, or the like. .. Further, the source code conversion program may be stored in a device connected to a public line, the Internet, a LAN, or the like, and the computer 200 may read the source code conversion program from these and execute the program.

100 Conversion device 110 Communication unit 111 Display unit 112 Operation unit 120 Storage unit 121 Source code storage unit 122 Tag insertion patch storage unit 123 Template storage unit 124 Algorithm information storage unit 125 Offload description storage unit 130 Control unit 131 Acquisition unit 132 Extraction unit 133 Update Department

Claims (6)

Based on the application having the source code, the offload description template corresponding to the offload description to be operated by the hardware is obtained by indicating the insertion point of the algorithm information of the application with a tag and representing the architecture of the hardware.
When the source code is updated, the algorithm information corresponding to the tag is extracted from the source code reflecting the update.
The extracted algorithm information is inserted into the corresponding part of the offload description template to update the offload description.
A source code conversion program characterized by having a computer execute processing. In the extraction process, when the source code is updated, a tag for extracting the algorithm information is inserted into the source code reflecting the update, and the algorithm information corresponding to the tag is extracted.
The source code conversion program according to claim 1. In the extraction process, the extraction tag is inserted by using a patch for inserting the extraction tag of the algorithm information.
The source code conversion program according to claim 2. In the extraction process, the extraction tag is inserted using the abstract syntax tree of the algorithm information.
The source code conversion program according to claim 2. The extraction process notifies that the offload description template needs to be updated when there is a part in the source code reflecting the update where the tag for extracting the algorithm information cannot be inserted.
The source code conversion program according to any one of claims 2 to 4, wherein the source code conversion program is characterized by the above. Based on the application having the source code, the offload description template corresponding to the offload description to be operated by the hardware is obtained by indicating the insertion point of the algorithm information of the application with a tag and representing the architecture of the hardware.
When the source code is updated, the algorithm information corresponding to the tag is extracted from the source code reflecting the update.
The extracted algorithm information is inserted into the corresponding part of the offload description template to update the offload description.
A source code conversion method characterized by a computer performing processing.

Images