JP2004355130A - Design method and apparatus of integrated circuit device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To achieve code optimizing for both of design data and a software. <P>SOLUTION: A design apparatus of an integrated circuit device optimizes both of design data 100 and a verification vector 101 by being provided with a code optimizing means 104 for deleting or commenting out a non-execution code description and by storing code coverage information generated by a code coverage measuring means 102 into a code coverage database 103 to use the database 103. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a technology for performing code optimization of code descriptions of design data and verification vectors using code coverage.
[0002]
[Prior art]
2. Description of the Related Art Conventionally, in designing an integrated circuit device, hardware design is performed using a hardware description language such as Verilog / VHDL. Then, in order to confirm whether the design data functions correctly, a verification vector is created and verified using a software language such as a hardware description language or C / C ++ or a verification-only language. I have. At that time, a technique called code coverage is used to quantitatively estimate how much the design data can be verified by the verification vector.
[0003]
In code coverage technology, when all code descriptions of design data to be verified are verified with a verification vector, executed descriptions (executable code descriptions) and non-executable descriptions (non-executable code descriptions) are measured. Then, the coverage ratio of how much of the code is the execution code is calculated and used to improve the quality of design data.
[0004]
The prior art will be described with reference to FIG. The hardware design data U1 and the verification vector U2 described in RTL (Register Transfer Level: register transfer level) are input to the code coverage measuring unit U3, and the code coverage is measured. The redundant description in the design data U1 is measured, and the redundant description of the verification vector U2 is deleted by the code optimizing means U4 from the code coverage information. That is, the verification vector U2 is optimized using the code coverage information, and the number of verification steps is reduced (for example, see Patent Document 1).
[0005]
[Patent Document 1]
JP 2001-14365 A (page 3-4, FIG. 1-2)
[0006]
[Problems to be solved by the invention]
However, the above-described conventional configuration has a problem that code optimization of a verification vector is possible, but code optimization of design data is not possible. In addition, there is a problem that the design data cannot be applied to software because it is intended only for hardware.
[0007]
The present invention has been made in view of such circumstances, and an object of the present invention is to enable code optimization of design data and software.
[0008]
[Means for Solving the Problems]
In order to solve the above problems, the present invention takes the following measures.
[0009]
As a first solution, a method for designing an integrated circuit device according to the present invention is a method for measuring an executable code description and a non-executable code description using a verification vector for each code description of design data describing a function of an integrated circuit device. Obtaining the coverage information, storing the code coverage information obtained by the measurement, and performing code optimization by deleting the non-executable code description from the design data based on the code coverage information. It is characterized by the following.
[0010]
The invention of the above integrated circuit device design method can be developed as an integrated circuit device design device as follows. In other words, the integrated circuit device design apparatus according to the present invention uses design data describing the function of the integrated circuit device and a plurality of verification vectors for verifying the design data, so that each code description of the design data is Code coverage measuring means for measuring an executable code description and a non-executable code description by a verification vector, code coverage storage means for storing code coverage information generated by the code coverage measuring means, and the code coverage information based on the code coverage information. Code optimization means for deleting the non-executable code description from the design data.
[0011]
According to this configuration, in the verification of the design data using the verification vector, the non-executable code description, which is the code description that has never been executed, in the design data is redundantly described using the code coverage information. And delete it. As a result, the code optimization of the design data can be realized.
[0012]
In the above, in the code optimizing step, it is preferable that a mode of deleting the non-executable code description is changed according to an implementation. That is, there are hardware and software as implementation forms. In the case of hardware, non-executable code description is deleted in a mode suitable for hardware, and in the case of software, non-executable code is deleted in a mode suitable for software. It is useful to delete the execution code description. At a more specific level, the following are preferred:
[0013]
That is, when the implementation form is hardware, a step of inlining each function call, a step of individually measuring code coverage for each function code description inlined, and a function from the code coverage information obtained by the measurement. Deleting the non-executable code description for each code description. Inline expansion is one of code optimization methods, and expands a function body at a function call position. Thereby, optimization in consideration of hardware implementation is realized.
[0014]
Further, in the case where the implementation form is software, by executing a step of inlining a function call when execution speed priority is selected and a step of making a function call when code size priority is selected, is there. For the function call step, the step of measuring code coverage in the function call for each function call and the non-executable code description common in the function call of the same function are deleted based on the code coverage information obtained by the measurement. And performing the steps.
[0015]
In the case of giving priority to the code size, the function entities are combined into one without performing inline expansion, and the functions are referred to in the form of a function call. In the code optimization, since one function code is shared by a plurality of function calls, code coverage information of the function code is recorded for each of the plurality of function calls, and only a common non-executable portion is deleted from the function code. Since the optimization is to delete the non-executable code description after sharing one function code by a plurality of function calls, the code size has priority.
[0016]
In the case where execution speed is prioritized, inline expansion is performed in the same way as in the case of hardware, instead of function calls, and all non-executable code descriptions in individual function codes after inline expansion are deleted. This optimizes the part to be deleted for each function. Since the optimization is to delete the non-executable code description for each function, the execution speed is prioritized.
[0017]
As described above, more efficient code optimization can be performed by selecting a deletion method suitable for each mounting method for optimization including not only hardware but also software.
[0018]
In the above, in the code optimizing step, the non-executable code description may be commented out for the future instead of deleting the non-executable code description.
[0019]
As a second solution, the integrated circuit device designing method according to the present invention measures an executable code description and a non-executable code description using a plurality of verification vectors for each code description of design data describing the function of the integrated circuit device. Obtaining code coverage information, storing the code coverage information obtained by the measurement, and verifying the executable code description and the non-executable code description in the stored code coverage information. Integrating the vectors into one. In this case, it is preferable that a plurality of verification vectors having the same code coverage information be integrated into one based on the code coverage information measured for each of the plurality of verification vectors.
[0020]
The invention of the above integrated circuit device design method can be developed as an integrated circuit device design device as follows. In other words, the integrated circuit device design apparatus according to the present invention uses design data describing the function of the integrated circuit device and a plurality of verification vectors for verifying the design data, so that each code description of the design data is Code coverage measuring means for measuring the execution code description and the non-executable code description by the verification vector, code coverage storage means for storing code coverage information generated by the code coverage measurement means, and storage by the code coverage storage means. Integrated means for integrating a plurality of verification vectors having the same executable code description and non-executable code description in the obtained code coverage information into one.
[0021]
Using the code coverage information acquired and stored by the code coverage measurement, the execution code descriptions are the same among the plurality of verification vectors, and the non-execution code descriptions are the same. , The plurality of substantially identical verification vectors are integrated into one. Thus, code optimization for the verification vector can be realized.
[0022]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, a method of designing an integrated circuit device according to an embodiment of the present invention will be described with reference to FIGS. It should be noted that the present invention is not limited to this embodiment at all, and can be implemented in various modes without departing from the gist.
[0023]
FIG. 1 is a block diagram showing an example of an apparatus and method for designing an integrated circuit device according to an embodiment of the present invention. In FIG. 1, reference numeral 100 denotes design data describing a function of a system, hardware, or software to be covered; 101, a verification vector for verifying the function of the design data 100; And a verification vector 102, code coverage measuring means for measuring whether each code line of the design data 100 is executed or not by the verification vector 101, and 103 is code coverage information output from the code coverage measuring means 102. Is a code coverage database that stores code. The code optimization unit 104 performs code optimization of the design data 100 and the verification vector 101 using the code coverage database information.
[0024]
An example of the operation of the method for designing the integrated circuit device configured as described above will be described with reference to FIGS.
[0025]
FIG. 2 is a flowchart for measuring the code coverage by inputting the design data 100 and the verification vector 101 in the code coverage measuring means 102 and optimizing the code of the design data 100 and the verification vector 101 by the code optimizing means 104.
[0026]
In step 200, the code coverage measuring means 102 inputs the design data 100 and the verification vector 101, and measures the code coverage using the design data 100 and the verification vector 101.
[0027]
Next, in step 201, the code coverage information, which is the code coverage measurement result, is registered in the code coverage database 103.
[0028]
Next, in step 202, it is determined whether or not to perform the code optimization on the design data 100. If the code optimization is not performed, the process is terminated. Proceed to step 203.
[0029]
Next, in step 203, the code of the design data 100 is optimized using the information of the code coverage database 103.
[0030]
In the next step 204, it is determined whether or not code optimization is to be performed on the verification vector 101. If code optimization is not performed, the process is terminated. To step 205.
[0031]
Next, in step S205, the code optimizing unit 104 performs code optimization of the verification vector 101 using information of the code coverage database 103.
[0032]
FIG. 3 is a flowchart showing in detail the process of code optimization of the design data 100 in step 203 of FIG. The design data 100 is input, and the processing is started from the first line of the design data 100.
[0033]
In step 300, it is determined whether or not the code description is a function part. If it is determined that the code description is not a function part, the process proceeds to step 301. If it is determined that the code description is a function part, the process proceeds to step 304.
[0034]
In step 301, it is determined whether or not the line has been executed. If the line has not been executed, the process proceeds to step 302. If the line has been executed, the process proceeds to step 303.
[0035]
If the code description is not a function part and a non-executable code description in the above flow, the process proceeds to step 302, where the code is deleted or commented out for the future and code optimization is performed. Proceeding to 303, the code advances one line and returns to step 300.
[0036]
On the other hand, if the code description is not a function part and the line is executed in the above flow, the code optimization step 302 is jumped to the step 303, the code is advanced by one line, and the process returns to the step 300.
[0037]
If it is determined in step 300 that the code description portion is a function portion and the process proceeds to step 304, the following processing is executed.
[0038]
In step 304, it is determined whether the function is implemented by hardware or software. If the function is implemented by hardware, the process proceeds to step 305. If the function is implemented by software, the process proceeds to step 308.
[0039]
In the case of hardware implementation, in step 305, inline expansion for expanding the function body at the function call position is performed.
[0040]
Next, in step 306, code optimization is performed by deleting or commenting out the non-executable code description for each function by performing inline expansion, and then proceed to step 307.
[0041]
In step 307, it is determined whether or not the last line of the code. If not, the process proceeds to step 303, advances the code by one line, and returns to step 300.
[0042]
If it is determined in step 304 that the implementation is implemented by software, in step 308, it is determined whether the execution speed or the code size is prioritized as a code optimization method. If the execution speed is prioritized, the process proceeds to step 305 as in the case of hardware, and if the code size is prioritized, the process proceeds to step 309.
[0043]
In the above flow, when the code description part is implemented by software with a function part, and when the execution speed is prioritized as a code optimization method, as in the case of the hardware, the process proceeds to step 305 and the inline processing is performed. In step 306, code optimization is executed by deleting or commenting out the non-executable code description for each function expanded inline. Thereafter, the last line of the code is determined in step 307, and if it is not the last line, the process advances by one line of code in step 303 and returns to step 300.
[0044]
On the other hand, if the code description part is implemented by software with the function part in the above flow and the code size is prioritized as a code optimization method, the process proceeds to step 309 to reduce the function entity to one. And convert it to a function call that uses it in common.
[0045]
Next, in step 310, code optimization is performed by deleting or commenting out the non-executable code description in common from all the function calls in the commonly used functions. Thereafter, the last line of the code is determined in step 307, and if it is not the last line, the process advances by one line of code in step 303 and returns to step 300.
[0046]
FIG. 4 is a block diagram showing a function call of design data. 400 is design data, and 401 is a code description that describes the substance of the function call called by the design data 400.
[0047]
FIG. 5 is a block diagram showing inline expansion of design data. Reference numeral 500 denotes design data, and reference numeral 501 denotes an entity declaration section of a function which is inline-expanded in the design data 500. The function entity part of the entity declaration part 501 is expanded as it is in the function call part of the design data 500 by inline expansion.
[0048]
FIG. 6 is a block diagram showing a code optimization method when design data is expanded inline. Reference numeral 600 denotes design data before optimization, reference numeral 601 denotes a function body called from the design data 600, and reference numeral 602 denotes inline expansion of a function call of the design data 600 by selecting only a part actually executed by the function body 601. Is the design data that is optimized by performing This makes it possible to perform inline expansion optimized for each function call. Since optimization is performed after inline expansion for each function call, execution speed has priority.
[0049]
In FIG. 6, "case mode3:" and "code3;break;" of the non-executable code description that is never executed are not selected and are deleted as a result, thereby performing the code optimization.
[0050]
FIG. 7 is a block diagram showing a code optimization method when design data is implemented by a function call. Reference numeral 700 denotes design data before optimization, reference numeral 701 denotes a function body called from the design data 700, and reference numeral 702 denotes optimization by deleting a non-executable code description commonly once in a function call of the design data 700. This is the design data that was designed. This allows further optimization while keeping the code size small. That is, since optimization is performed to delete a non-executable code description after sharing one function code by a plurality of function calls, the code size is prioritized.
[0051]
In FIG. 7, "case mode3:" and "code3;break;" of the non-executable code description are deleted, and the code is optimized.
[0052]
FIG. 8 is a block diagram showing the code coverage database of the main function part of the design data in the information stored in the code coverage database 103 by the code coverage measuring means 102. Reference numeral 800 denotes a verification vector 1 including a parameter flag = 0, reference numeral 801 denotes a verification vector 2 including a parameter flag = 1, reference numeral 802 denotes a main function part of the design data, and reference numeral 803 denotes code coverage of the design data main function part 802. It is a code coverage database that stores measurement results.
[0053]
The code coverage database 803 stores the code coverage information of each of the verification vector 1 (800) and the verification vector 2 (801), and the function call portion indicates a pointer to the separately stored code coverage database.
[0054]
FIG. 9 is a block diagram showing the code coverage database of the function call part of the design data in the information stored in the code coverage database 103 by the code coverage measuring means 102. Reference numeral 900 denotes a verification vector 1 including a parameter flag = 0, reference numeral 901 denotes a verification vector 2 including a parameter flag = 1, reference numeral 902 denotes a main function part of design data, and reference numeral 903 denotes a function body called from the design data. 904 is a code coverage database that stores the code coverage measurement result of the call function body 903.
[0055]
The code coverage database 904 stores the code coverage information of each of the verification vector 1 (900) and the verification vector 2 (901), and further stores a code coverage measurement result for each function call called in the main function.
[0056]
In FIG. 9, for “case mode 3:” and “code 3; break;” of the non-executable code description, both the function call 1 and the function call 2 in the verification vector 1 are “0”, and the Both function call 1 and function call 2 are "0".
[0057]
FIG. 10 is a block diagram illustrating a method of deleting redundant verification vectors using a code coverage database storing code coverage measurement results of a plurality of verification vectors. A verification vector 1000 includes parameters flag1 = 0 and flag2 = 0, a verification vector 1001 includes parameters flag1 = 1 and flag2 = 0, and a verification vector 1002 includes parameters flag1 = 1 and flag2 = 1. Reference numeral 1003 denotes a main function part of the design data, and 1004 denotes a code coverage measurement result for each of the verification vector 1 (1000), the verification vector 2 (1001), and the verification vector 3 (1002) of the design data main function part 1003. Is a code coverage database in which is stored.
[0058]
When the coverage information of the verification vector 2 (1001) of the code coverage database 1004 is compared with the coverage information of the verification vector 3 (1002), the coverage information is completely the same. Therefore, it can be seen that one of them is a redundant verification vector, so that one of them can be deleted, thereby optimizing the verification vector.
[0059]
【The invention's effect】
According to the present invention, the code size and the execution speed can be improved by removing redundant codes for design data and verification vectors by using code coverage information.
[0060]
Also, by selecting a code optimization method between the case where the function part of the design data is implemented in hardware and the case where it is implemented in software, it is possible to handle both hardware and software design data. This has the effect that it can be performed.
[0061]
Furthermore, when implemented by software, by selecting a code optimization method based on evaluation indexes such as execution speed and code size, there is an effect that more excellent code optimization can be performed.
[Brief description of the drawings]
FIG. 1 is a block diagram illustrating an example of an apparatus and method for designing an integrated circuit device according to an embodiment of the present invention. FIG. 2 is a flowchart illustrating an example of a code optimizing process according to an embodiment of the present invention. FIG. 4 is a flowchart illustrating an example of a design data optimizing method according to an embodiment of the present invention. FIG. 4 is an operation explanatory diagram illustrating an example of a design data function calling method according to an embodiment of the present invention. FIG. 6 is an operation explanatory diagram showing an example of a method of inline expansion of design data according to the embodiment. FIG. 6 is an operation explanatory diagram showing an example of a code optimization method at the time of inline expansion according to an embodiment of the present invention. FIG. 8 is an operation explanatory diagram showing an example of a code optimization method at the time of a function call in the embodiment. FIG. 8 shows an example of a code coverage database of a main function in the embodiment of the present invention. FIG. 9 is an operation explanatory diagram showing an example of a code optimization method for a function call part in the embodiment of the present invention. FIG. 10 is an example of code optimization of a redundant verification vector in the embodiment of the present invention. FIG. 11 is a block diagram showing an example of a conventional code optimization method.
REFERENCE SIGNS LIST 100 design data 101 verification vector 102 code coverage measuring means 103 code coverage database 104 code optimizing means

Claims (11)

Obtaining code coverage information by measuring an executable code description and a non-executable code description by a verification vector for each code description of the design data describing the function of the integrated circuit device;
Storing code coverage information obtained in the measurement,
Removing the non-executable code description from the design data based on the code coverage information and performing code optimization. The method of designing an integrated circuit device according to claim 1, wherein in the code optimizing step, an aspect of deleting the non-executable code description is changed according to an implementation form. The code optimizing step, when the mounting form is hardware,
A step of inlining each function call;
A step of individually measuring code coverage for each function code description expanded inline,
And deleting a non-executable code description for each function code description from the code coverage information obtained by the measurement. 3. The method of designing an integrated circuit device according to claim 2, wherein The code optimizing step, when the mounting form is software,
A step of inlining function calls when execution speed priority is selected;
The step of making a function call when the code size priority is selected is executed. The function call process includes:
Measuring code coverage within the function call for each function call;
And deleting a common non-executable code description in a function call of the same function based on the code coverage information obtained by the measurement. 6. The integrated circuit according to claim 1, wherein the code optimization step includes commenting out the non-executable code description instead of deleting the non-executable code description. How to design the device. Obtaining code coverage information by measuring an executable code description and a non-executable code description with a plurality of verification vectors for each code description of the design data describing the function of the integrated circuit device;
Storing code coverage information obtained in the measurement,
Integrating a plurality of verification vectors having the same executable code description and non-executable code description in the stored code coverage information into one, and integrating them into one. Design method. The method of claim 7, wherein the step of integrating the verification vectors integrates a plurality of verification vectors having the same code coverage information into one based on the code coverage information measured for each of the plurality of verification vectors. The design method of the integrated circuit device according to the above. Using the design data describing the function of the integrated circuit device and a plurality of verification vectors for verifying the design data, each code description of the design data is executed code description and non-execution code description by the plurality of verification vectors. Code coverage measurement means for measuring
Code coverage storage means for storing code coverage information generated by the code coverage measurement means,
A code optimizing means for deleting the non-executable code description from the design data based on the code coverage information. 10. The integrated circuit device designing apparatus according to claim 9, wherein the code optimizing unit performs comment-out on the non-executable code description instead of deleting the non-executable code description. Using the design data describing the function of the integrated circuit device and a plurality of verification vectors for verifying the design data, each code description of the design data is executed code description and non-execution code description by the plurality of verification vectors. Code coverage measurement means for measuring
Code coverage storage means for storing code coverage information generated by the code coverage measurement means,
Integrated means for integrating a plurality of verification vectors having the same executable code description and non-executable code description in the code coverage information stored by the code coverage storage means into one; Circuit device design equipment.

Images