JP3579952B2 - Signal conversion method - Google Patents

Description

[0001]
[Industrial applications]
The present invention relates to a signal conversion method for converting a pulse train signal in which a plurality of pulses are continuous in time series into a predetermined format.
[0002]
[Prior art]
One of the pulse train signals in which a plurality of pulses are continuous in time series is a test pattern for verifying a logic circuit. The test pattern referred to here is used as test data for verifying whether or not the designed logic circuit operates normally by a logic circuit simulator incorporating the logic of the logic circuit when designing the logic circuit. This is a general term for an input signal data group and an output signal data group expected as a normal response thereto.
[0003]
Conventionally, this test pattern has often been created manually by a graphic editor or the like.However, as the logic circuit becomes larger, the test pattern to be created becomes more and more complex, and the length of the test pattern also increases. Since the length of a test pattern from 100 bits to thousands of bits and the number of parallel input of test patterns to a logic circuit tend to increase, it takes a lot of time to create and edit test patterns. It is getting. In addition, such a complicated test pattern requires a large amount of work time for error checking after creation.
[0004]
Therefore, in order to reduce the test pattern creation load, create a library of the most frequently used basic types of signal change types, call those types from the library as needed, and create a test pattern by combining them Several proposals have been made regarding the method of doing so. However, in this conventional method, signals are separated at equal intervals and the divided signals are registered in the library as basic patterns. Therefore, when this is called and used, even if the signal does not change, the corresponding basic pattern Need to be called one by one and combined. As a result, the data amount of the created test pattern increases.
[0005]
Japanese Patent Application Laid-Open No. H5-159013 discloses a method of making a library of basic patterns that are also frequently used. However, since the combination and registration of patterns are performed manually, the workload is large, and the If the length is long, there is a problem that the possibility of an operation error increases and the data amount also increases.
In addition to this, in order to reduce the storage capacity of the input test pattern, a method of compressing the data amount by storing only the time data of the change point has been proposed. Since the information is not stored, it can be applied only when the signal has only two values, that is, whether the signal is on or off, and there is a disadvantage that information other than the information that the value of the signal is on or off cannot be stored.
[0006]
[Problems to be solved by the invention]
The present invention has been made in view of the above circumstances, and has as its object to provide a signal conversion method for compressing a data amount by converting a pulse train signal in which a plurality of pulses are continuous in time series into a predetermined format.
[0007]
[Means for Solving the Problems]
To achieve the above object, the signal conversion method of the present invention provides a plurality of basic patterns in which a plurality of types of changes in a pulse train signal in which a plurality of pulses are continuous in a time series are categorized into a plurality of types, and the pulse train signal is time-converted. By applying the signal in each section to any of the above basic patterns while sequentially dividing, the pulse train signal is obtained by attaching time data representing the time of change of the signal in the section to the applied basic pattern. , And is converted into a series of basic patterns and time data assigned to each basic pattern.
[0008]
Here, a first time between a last change point of a signal in a first time interval that is earlier in time among the adjacent time intervals and a contact point between the adjacent time intervals, and The pulse train signal of the pulse train signal is so set that the first change point of the signal in the second time interval that is later in time among these adjacent time intervals and the second time between the contact point and the second point are the same time. Preferably, a break point is defined.
[0009]
Further, each of the plurality of pulse train signals is converted into a sequence of the basic pattern and time data added to each basic pattern, and a sequence of a basic pattern corresponding to a first pulse train signal of the plurality of pulse train signals is converted. When at least a part of the sequence of the basic pattern corresponding to the second pulse train signal of the plurality of pulse train signals matches, at least a part of the sequence of the basic pattern corresponding to the first pulse train signal It is also a preferable embodiment to replace with a notation indicating that it matches the sequence of the basic pattern corresponding to the second pulse train signal.
[0010]
Note that the pulse train signal in the signal conversion method of the present invention is not only a binary signal that transitions between two states of a certain value and another value, but also has, for example, an undefined state or a high impedance state. Such a signal may be used.
[0011]
[Action]
According to the signal conversion method of the present invention, a plurality of basic patterns are prepared by classifying in advance the types of changes of the pulse train signal, and the pulse train signal is converted into a series of basic patterns and time data assigned to each basic pattern. In consideration of the time axis, the type of change of the signal that exists infinitely can be expressed by a combination of a predetermined limited type of basic pattern and time data. Data volume is greatly reduced.
[0012]
Further, in the process of dividing the pulse train signal into a series of basic patterns, the last change point of the signal in the first time interval that is earlier in time between two adjacent time intervals, The first time between the contact points of the separation sections and the second time between the first transition point of the signal in the second separation section that is later in time and the contact point are the same time. When the break point of the pulse train signal is determined so as to satisfy the following condition, only one of the first time and the second time, which are substantially the same time data, is stored as time data for both purposes, and Can be saved, and the data amount is further compressed.
[0013]
In addition, each of the plurality of pulse train signals is converted into a basic pattern sequence and time data attached to each basic pattern, and at least a basic pattern sequence corresponding to a first pulse train signal of the plurality of pulse train signals is converted. When a part matches the sequence of the basic pattern corresponding to the second pulse train signal among the plurality of pulse train signals, the at least a part of the sequence of the basic pattern corresponding to the first pulse train signal is changed to If it is replaced with a notation indicating that it matches the sequence of the basic pattern corresponding to the second pulse train signal, the “at least a part” is replaced with the sequence of the basic pattern having a large data amount, It is expressed by the notation that they match, and the data amount is further compressed.
[0014]
【Example】
Hereinafter, examples of the present invention will be described.
FIG. 1 is a block diagram of a test pattern generation device including an embodiment in which the signal conversion method of the present invention is applied to a test pattern for a logic circuit simulator.
Here, the types of changes in the test pattern in which a plurality of pulses are successive in time series are classified in advance, categorized as a plurality of basic patterns, and the categorized basic patterns are stored in a basic pattern table shown in FIG. 10 is stored.
[0015]
As shown in FIG. 1, in addition to the basic pattern table 10, the test pattern generation apparatus creates test pattern input data by combining graphic data or character data by operating a keyboard or a mouse. The input data input means 11 fits the signal in each section interval to one of the basic patterns while separating the input test pattern in time series, and the signal in the section interval is applied to the fitted basic pattern. Data conversion means 12 for converting a test pattern into a series of basic patterns and time data added to each basic pattern by attaching time data representing the time of change, and registering the converted data. Conversion data registration file 20 and the converted data Conversion data registration means 13 for registration, conversion data reading means 14 for reading conversion data registered in the conversion data registration file 20, and test pattern generation means 15 for generating a test pattern from the read conversion data are provided. ing.
[0016]
Note that the data input means 11 can also accept an output signal from a logic circuit simulator (not shown) as a test pattern.
FIG. 2 is a diagram showing the contents of the basic pattern table 10 of FIG.
As described above, the basic pattern table 10 classifies and classifies the types of changes in the test pattern, and stores them in a library.
[0017]
As shown in FIG. 2, in this embodiment, fifteen types of basic patterns are prepared, and each basic pattern is denoted by reference numerals 1, 2, 3,... Can be identified by codes 1A, 1B, 1C, 1D, 2A, 2B, 2C, 3A,..., 4D in which codes A, B, C, D,. .
[0018]
Code 1A represents a basic pattern in which the signal changes from 0 to 1 and further changes from 1 to 0, code 1B represents a basic pattern in which the signal changes from 1 to 0 and further changes from 0 to 1, and code 1C Represents a basic pattern in which a signal changes from a high impedance state to 1 and further changes from 1 to a high impedance state, and code 1D represents a basic pattern in which a signal changes from a high impedance state to 0 and further changes from 0 to a high impedance state. Represents a pattern.
[0019]
Code 2A represents a basic pattern in which a signal changes from 0 to an undefined state and further changes from an undefined state to 0, and code 2B represents a basic pattern in which a signal changes from 1 to an undefined state and further changes from an undefined state to 1. The code 2C represents a basic pattern in which a signal changes from a high impedance state to an undefined state and further changes from an undefined state to a high impedance state. The code 2D is undefined.
[0020]
Further, code 3A represents a basic pattern in which the signal changes from 0 to 1, code 3B represents a basic pattern in which the signal changes from 1 to 0, and code 3C changes the signal from the high impedance state to 1. The code 3D indicates a basic pattern in which a signal changes from a high impedance state to 0.
Code 4A represents a basic pattern in which a signal changes from 1 to a high-impedance state, code 4B represents a basic pattern in which a signal changes from 0 to a high-impedance state, and code 4C represents a signal in which the signal remains at 1. The code 4D represents a basic pattern that does not change while the signal remains at 0.
[0021]
How the basic pattern table 10 is used will be described later.
FIG. 3 is a diagram showing the contents of the conversion data registration file 20 of FIG.
The conversion data registration file 20 is for storing the conversion data registered by the conversion data registration unit 13 described later, and each basic pattern sequence 31, 32, 33 constituting the conversion data is stored in the conversion data registration file 20. . Can be identified by codes 01, 02, 03, 04,...
[0022]
How the conversion data registration file 20 is used will be described later.
Hereinafter, the operation of this embodiment will be described with reference to FIG.
First, a test pattern is created and input by the data input means 11 of FIG.
Next, the input test pattern is time-separated in order from the beginning by the data conversion means 12, and the signal in each delimitation section is applied to the basic pattern stored in the basic pattern table 10. In fitting to the basic pattern, a type in which the state of the signal changes, for example, the value of the signal changes from 1 to 0, without considering the position of the change point on the time axis of the signal in the delimiting section. It is determined whether or not the type of change between the signal in the segment interval and the basic pattern matches by paying attention only to the type of change such as changing to 1. Is applied to the corresponding basic pattern.
[0023]
On the other hand, the signal in the delimiter section is applied to the basic pattern, and time data representing the time of change in each delimiter section is stored with the applied basic pattern.
FIG. 4 is a detailed flowchart of the conversion process from the test pattern to the basic pattern sequence and time data in the data conversion unit 12 of FIG.
[0024]
FIG. 5 is a schematic diagram showing a process in which a test pattern is divided in time series, and a signal in each segment is applied to the basic pattern and converted into a series of basic patterns. FIG. FIG. 7 is a schematic diagram showing a process of extracting time data of a signal change in each delimiting section of FIG.
Hereinafter, a process of converting a test pattern into a basic pattern sequence and time data will be described with reference to FIGS.
[0025]
First, in step S41 of FIG. 4, a test pattern is read, and in step S42, a change point of a signal is extracted in a time series from the test pattern, and an operation of dividing the signal in a time series is performed.
The extraction of the change point is performed as follows. In the test pattern shown in FIG. 5A, extraction of a change point of a signal is started from a start point S, first, a first change point 51 is extracted, then, a second change point 52, and a third change point 53 Thus, three change points adjacent in time series are extracted. Along with the extraction of the change points, the type of signal change at each change point is extracted. That is, at the first change point 51, the signal changes from 0 to 1, at the second change point 52, the signal changes from 1 to 0, and at the third change point 53, Is extracted as a type in which the signal changes from 0 to 1.
[0026]
The change type at the first change point and the change type at the third change point among the three consecutive change points are compared, and the change type at the first change point and the change type at the third change point are compared. Are the same, a section including the first change point and the second change point is defined as one delimiting section. On the other hand, when the type of change at the first change point is different from the type of change at the third change point, a section including only the first change point is defined as one delimiting section.
[0027]
In FIG. 5A, since the type of change at the first change point 51 and the type of change at the third change point 53 are both types of change from 0 to 1, the first change point 51 And a second transition point 53 are separated by a break point E defined as follows. That is, the break point E between the adjacent break section R1 and the break section R2 in FIG. 5A is a signal of the signal in the first break section R1 that is earlier in time among the two break sections R1 and R2. A first time from the last change point 52 to a contact point E where the break section R1 and the break section R2 are in contact with each other, and a signal in the second break section R2 which is later in time among the two break sections R1 and R2. Are determined so that the second time from the first change point 53 to the contact point E is the same as each other. In other words, the time between the last change point 52 of the signal in the first segment R1 that is earlier in time and the first change point 53 of the signal in the second segment R2 that is later in time is 2 The time of equal division is defined as a break point E.
[0028]
The method of determining the breakpoint when the type of change at the first change point is different from the type of change at the third change point is the same as described above, and the signal of the signal within the first break section that is earlier in time is also determined. A time point at which the time between the last change point and the first change point of the signal in the second time interval that is later in time is bisected is set as the break point.
Subsequent to the extraction of the first delimiter section, the second delimiter section is extracted. In order to cut out the second break section, the break point E is set as the start point of the second break section, and the change type at the first change point 53 and the third break point are set in the same manner as in the cut of the first break section R1. Is compared with the type of change at the change point 55. Thereafter, the second and third delimiting sections R2 and R3 are sequentially extracted in the same manner as when the first delimiting section R1 is extracted.
[0029]
Next, in step S43, each of the delimited sections R1, R2, R3,... Cut out as described above is applied to one of the basic patterns in the basic pattern table 10 in FIG. In step S44, the sequence is converted into a sequence of the fitted basic patterns. That is, from the test pattern shown in FIG. 5A in which six change points 51, 52, 53, 54, 55, and 56 continue in chronological order starting from the starting point S, three delimiting sections R1, R2 and R3 are cut out. Here, as shown in FIG. 5B, each of the delimiting sections R1, R2, R3,... Is set to the basic pattern of the code 1A stored in the basic pattern table 10. It is fitted and converted into a series of basic patterns consisting of three basic patterns 1A, as shown in FIG. 5 (c).
[0030]
On the other hand, in step S45, time data is generated from the time of each of the above change points.
FIG. 6 is a schematic diagram illustrating a process of generating time data from the time of each change point. FIG. 6 shows the test pattern of FIG. 5A from the start point S to the fourth change point 54. Time data is extracted from each of the sections R1, R2, R3,... Divided in step S42. Here, the first delimiting section R1 corresponds to the case where the type of change at the first change point and the type of change at the third change point are the same, and two change points are included in the delimited section R1. Since it exists, a time T1 from the start point S to the first change point 51, a time T2 from the first change point 51 to the second change point 52, and a time T3 from the second change point 52 to the break point E Is extracted. These three times T1, T2, and T3 become time data corresponding to the first element 1A of the series of the basic pattern in FIG. 5C. On the other hand, if the type of change at the first change point is different from the type of change at the third change point, there is only one change point in the delimited section, so the time from the start point S to the first change point is Data and two pieces of time data from the first change point to the break point are extracted.
[0031]
Next, the time data of the second segment R2 is extracted using the segment E between the first segment R1 and the second segment R2 as the starting point of the second segment R2, and thereafter, similarly. The time data of the third and fourth delimiting sections are extracted.
By the way, as described above, the break point E is a time obtained by dividing the time between two adjacent change points 52 and 53 into two equal parts. For example, the time data T3 of the first break section R1 in FIG. , And the same as the time data T4 of the second section R2. Therefore, in the present embodiment, in step S46, the time data T3 of the first time interval R1 and the time data T4 of the second time interval R2, which are earlier in time, of the adjacent time intervals are shared. The data amount is compressed by storing only the time data.
[0032]
In this way, by converting the test pattern into a series of basic patterns and time data attached to each basic pattern, the types of signal changes which are infinite when considering the time axis can be changed to these 15 types of basic patterns. It is possible to express the data by converting it into converted data composed of a pattern and time data, and it is possible to greatly reduce the data amount of a test pattern having a huge data amount. Test data generation, storage, input / output, and the like can be efficiently executed by the converted data having the data amount compressed as described above.
[0033]
As shown in FIG. 2, the basic pattern table 10 stores a basic pattern having a high degree of data compression, that is, a basic pattern having a large number of change points, to a basic pattern having a low degree of compression, that is, a basic pattern having a small number of change points. In order to apply the signal in each segment to the basic pattern, the matching is started from the basic pattern stored at the top of the basic pattern table 10 and the signal in the segment is stored. If they match, the matching basic pattern is applied to the signal in the above-mentioned section, and if they do not match, the matching is sequentially performed so as to match with the next basic pattern. This makes it possible to convert the test pattern into a series of basic patterns with a high degree of compression. For example, in FIG. 2, code 1A represents a basic pattern in which the test pattern changes from 0 to 1 and further changes from 1 to 0, while a basic pattern in which the test pattern changes from 0 to 1 is represented by code 3A. The defined basic pattern in which the test pattern changes from 1 to 0 is defined in code 3B. Therefore, the signal in the section where the test pattern changes from 0 to 1 and further changes from 1 to 0 is stored in the front without being applied to the basic pattern of the codes 3A and 3B stored in the rear. Is applied to the basic pattern of the code 1A. If the codes 1A, 3A, and 3B are stored in the order of 3A, 3B, and 1A instead of the above order, the signal in the delimiter section where the test pattern changes from 0 to 1 and further changes from 1 to 0 is the code. This is applied to the two basic patterns of 3A and code 3B, and the data compression degree is reduced as compared with the case where the data is stored in the order shown in FIG.
[0034]
As for the time data, the two basic patterns of the code 3A and the code 3B require six time data, whereas the basic pattern of the code A1 requires only four time data. By storing basic patterns in order, data can be compressed with a high degree of compression.
Next, conversion data including a plurality of test patterns converted into a series of basic patterns and time data is registered in the conversion data registration file 20 for each test pattern by the conversion data registration unit 13 of FIG.
[0035]
When registering the conversion data in the conversion data registration file 20, when there are a plurality of test patterns, the conversion data corresponds to a certain one of the plurality of test patterns (this is referred to as a first test pattern). If a part or all of the series of basic patterns to be performed matches a series of basic patterns corresponding to another test pattern (referred to as a second test pattern) among the plurality of test patterns, the first test is performed. The matching part of the series of the basic pattern corresponding to the pattern is registered after being replaced with a notation indicating that it matches the series of the basic pattern corresponding to the second test pattern. Hereinafter, the procedure for registering the conversion data will be described including a specific example of this point.
[0036]
FIG. 7 is a flowchart showing a detailed procedure of conversion data registration by the conversion data registration means 13 of FIG.
FIG. 8 is a schematic diagram showing a process of additionally registering a new basic pattern sequence in the conversion data registration file 20 of FIG.
The newly registered basic pattern sequence 40 is assumed to be composed of four elements of codes 1A, 1A, 1A, and 1B, as shown in FIG. 8B. In registering the basic pattern sequence 40, the three basic pattern sequences 31, 32, and 33 already registered in the conversion data registration file 20 are referred to (step S61 in FIG. 7).
[0037]
The registered basic pattern sequence 40 is first compared with the basic pattern sequence 31 (step S62 in FIG. 7). This comparison starts with a comparison between the first elements of both streams, and then a comparison is made between the second element and the third element (step S63 in FIG. 7).
In the example shown here, a part of the basic pattern sequence 40 shown in FIG. 8B, that is, the first to third elements and the basic pattern sequence already registered in the conversion data registration file 20. 31 matches the first element to the third element of the basic pattern sequence 40, and is replaced with the code 01 representing the position of the basic pattern sequence 31 in the conversion data registration file 20 (FIG. 7). Step S64). Next, the replaced code (01) and the fourth element (1B) of the series 40 of the basic pattern that has not been replaced are combined to form the series shown in FIG. A series 34 of basic patterns as shown in the bottom line of FIG. 8A is newly registered in the conversion data registration file 20. A code 04 representing the sequence is assigned to the sequence 34 of the basic pattern (step S65 in FIG. 7).
[0038]
As described above, initially, the basic pattern sequence 40 composed of four elements is converted into the basic pattern sequence 34 composed of two elements, so that the data amount of the test pattern is further compressed.
If some or all of the elements of the registered basic pattern sequence 40 do not match the registered basic pattern sequence 31, the registered basic pattern sequence 40 is: When a part or all of the elements of the registered basic pattern sequence 40 are compared with the registered second basic pattern sequence 31 and do not match the registered second basic pattern sequence 32, Is compared with the registered basic pattern series 33 with the registered basic pattern series 33. If some or all of the elements of the registered basic pattern sequence 40 do not match any of the registered basic pattern sequences, the registered basic pattern sequence 40 remains in its code at step S65. In step S65, a code indicating the position in the conversion data registration file 20 is added to the series of the basic pattern and registered in the conversion data registration file 20.
[0039]
In this way, when a part or all of the elements of the sequence of the registered basic pattern matches all the elements of the sequence of the registered basic pattern, the matching part of the sequence of the registered basic pattern is matched. By replacing the code with the code representing the sequence of the registered basic pattern, the data amount of the sequence of the registered basic pattern can be significantly reduced. When a new basic pattern is registered, even if the series of the basic pattern is further compressed as described above, no special processing is performed on the time data attached to the basic pattern. The data is registered in the conversion data registration file 20 together with the series of the basic pattern.
[0040]
Next, a description will be given of a method of restoring converted data consisting of a series of basic patterns and time data created by the signal conversion method of the present invention to an original test pattern. In this restoration processing, processing reverse to the data conversion processing in the data conversion means 12 and the registration processing in the converted data registration means 13 is performed.
First, the converted data reading means 14 shown in FIG. 1 reads the first converted data (basic pattern sequence 31 and corresponding basic data 31) from among the plurality of converted data registered in the converted data registration file 20 shown in FIG. Time data).
[0041]
Next, the test pattern generation unit 15 searches for elements that match the codes assigned to the basic pattern sequence in the converted data registration file 20 from among the read basic pattern sequence elements. For example, first, each of the three elements 1A, 1A, 1A of the sequence 31 of the first basic pattern in FIG. 8A is converted into a code 01, 02, 03 added to the sequence of each basic pattern in the conversion data registration file 20. , 04 in order. As a result, none of the three elements 1A, 1A, 1A of the basic pattern sequence 31 match any of the codes 01, 02, 03, 04 assigned to each basic pattern sequence in the conversion data registration file 20. Then, the collation processing of the basic pattern sequence 31 with the conversion data registration file 20 ends.
[0042]
Next, the basic pattern sequence 31 is restored to the original test pattern format as shown in FIG.
This restoration processing is performed with reference to the basic pattern table 10. First, each of the three elements 1A, 1A, 1A of the series 31 of the basic pattern shown in FIG. 5 (c) is associated with the three basic patterns 1A shown in FIG. 5 (b). The test pattern shown in FIG. 5A is restored from the time data (T1, T2, T3,... In FIG. 6) stored in the conversion data registration file 20 together with each element of the series of the basic pattern. I do.
[0043]
When the matching process and the restoring process for the first basic pattern sequence 31 are completed in this way, the conversion data reading unit 14 shown in FIG. 1 uses the conversion data reading unit 14 shown in FIG. The third conversion data is read, and the test pattern generation unit 15 starts the collation processing and the restoration processing.
First, the two elements 1A and 3A of the second basic pattern sequence 32 are sequentially collated with the codes 01, 02, 03 and 04 assigned to the respective basic pattern sequences in the conversion data registration file 20. As a result, none of the elements 1A and 3A of the basic pattern sequence 32 match any of the codes 01, 02, 03 and 04 assigned to the respective basic pattern sequences in the conversion data registration file 20. The collation processing of the series 32 with the conversion data registration file 20 ends.
[0044]
Next, the basic pattern sequence 32 is restored to the original test pattern format as shown in FIG. The restoration procedure is the same as in the case of the series 31 of the basic pattern.
In this way, the processing is sequentially advanced for the third and subsequent basic pattern sequences 33, 34,..., But in the collation processing of the fourth basic pattern sequence 34, the elements of the basic pattern sequence 34 One of the codes matches one of the codes assigned to the series of each basic pattern in the conversion data registration file 20. That is, the two elements 01 and 1B of the fourth basic pattern sequence 34 are sequentially collated with the codes 01, 02, 03 and 04 assigned to the respective basic pattern sequences in the conversion data registration file 20. And the first element 01 of the basic pattern sequence 34 matches the code 01 assigned to the first basic pattern sequence of the conversion data registration file 20. Therefore, the element 01 at the head of the series 34 of the basic pattern is replaced with the element (1A, 1A, 1A) of the series 31 of the basic pattern corresponding to the code 01 added to the series. As a result, the fourth basic pattern sequence 34 is converted from 01, 1B to 1A, 1A, 1A, 1B, and the original basic pattern sequence 40 shown in FIG. 8B is obtained.
[0045]
Next, as in the case of the first basic pattern sequence 31, each of the four elements 1A, 1A, 1A, 1B of the basic pattern sequence 40 is stored in the basic pattern table 10 and the conversion data registration file 20. 5A, and a test pattern having the same format as that of FIG. 5A is restored.
The test pattern restored in this way is used as input data of the logic circuit simulator.
[0046]
In general, a test pattern having a format as shown in FIG. 5A is input to the logic circuit simulator, but the logic circuit simulator has the same function as that of the test pattern generation means 15. By having it, the conversion data in the conversion data registration file 20 can be directly input to the logic circuit simulator.
[0047]
By doing so, the converted data of the test pattern, whose data amount has been compressed by the signal conversion method of the present invention and registered in the converted data registration file 20, can be transferred to another storage medium while the data amount is compressed. It can be stored, copied, and stored, making it easier to handle test pattern data.
In addition, a new test pattern can be generated by manually executing the same processing as the conversion data reading means 14 and the test pattern generating means 15. That is, conversion data consisting of a series of basic patterns and time data is read from the conversion data registration file 20, and manually combined, copied, deleted, and modified to freely generate a new test pattern. Can be. The test pattern generated in this way can be registered in the conversion data registration file 20 via the data conversion unit 12 and the conversion data registration unit 13 as shown in FIG.
[0048]
【The invention's effect】
As described above, according to the signal conversion method of the present invention, the amount of data is compressed by converting the pulse train signal into a sequence of basic patterns and time data attached to each basic pattern, and as a result, the data generation , It is possible to reduce the work time for error checking and the like, and to reduce the capacity of the storage device and the storage medium.
[0049]
Further, when the pulse train signal is time-separated and the signals in each of the break intervals are applied to the basic pattern to make the pulse train signal a sequence of the basic pattern, the time interval among the above-mentioned adjacent break intervals that are earlier in time is used. A first time between the last transition point of the signal within one break section and a contact point between these adjacent break sections, and a second time-lag among the above break sections adjacent to each other. In the case where the break point of the pulse train signal is determined so that the first change point of the signal in the break section and the second time between the contact point are the same as each other, substantially the same break point is set. Either the first time or the second time, which is time data, can be stored as dual-purpose time data, and the other need not be stored, so that the data amount is further compressed, and result Data generation, working time, such as error checking can be further shortened, and it is possible to further reduce the capacity of the storage device and a storage medium.
[0050]
Further, each of the plurality of pulse train signals is converted into a basic pattern sequence and time data added to each basic pattern, and one of the basic pattern sequences corresponding to the first pulse train signal of the plurality of pulse train signals is converted. When the part or all of the sequence matches the sequence of the basic pattern corresponding to the second pulse train signal among the plurality of pulse train signals, a part or all of the sequence of the basic pattern corresponding to the first pulse train signal is If it is replaced with a notation indicating that it matches the series of the basic pattern corresponding to the second pulse train signal, the data amount can be further compressed, and as a result, the work time for data generation, error check, etc. can be further reduced. And the capacity of the storage device and the storage medium can be further reduced.
[0051]
Further, by automating the signal conversion method of the present invention using a computer, the work load can be greatly reduced. In addition, since the signal conversion operation is accurately executed by the automation, an error check is omitted. You can also.
[Brief description of the drawings]
FIG. 1 is a block diagram of a test pattern generation apparatus including an embodiment in which a signal conversion method of the present invention is applied to a test pattern for a logic circuit simulator.
FIG. 2 is a diagram showing contents of a basic pattern table in one embodiment of the present invention.
FIG. 3 is a diagram showing a part of the contents of a conversion data registration file according to an embodiment of the present invention.
FIG. 4 is a flowchart showing a procedure for converting a test pattern into a basic pattern sequence and time data in one embodiment of the present invention.
FIG. 5 is a schematic diagram showing a process in which a change type of a signal in each delimiting section of a test pattern is applied to a basic pattern in one embodiment of the present invention.
FIG. 6 is a schematic diagram showing a process of converting time information of a signal change in each delimiting section of a test pattern into time data in one embodiment of the present invention.
FIG. 7 is a flowchart showing a detailed procedure of conversion data registration in one embodiment of the present invention.
FIG. 8 is a schematic diagram showing a process of registering a basic pattern sequence in a conversion data registration file according to an embodiment of the present invention.
[Explanation of symbols]
10 Basic pattern table
11 Data input means
12 Data conversion means
13 Conversion data registration means
14 Conversion data reading means
15 Test pattern generation means
20 Conversion data registration file
31, 32, 33, 34, 40 Series of basic patterns
51, 52, 53, 54, 55, 56 Change point

Claims (3)

Prepare a plurality of basic patterns in which the types of changes in the pulse train signal in which a plurality of pulses are chronologically continuous are categorized into a plurality,
Applying the signal in each segment to one of the basic patterns while dividing the pulse train signal in time series, and attaching time data representing the time of change of the signal in the segment to the fitted basic pattern. And converting the pulse train signal into a sequence of the basic patterns and time data added to each basic pattern. A first time between a last transition point of a signal in a first time interval that is earlier in time among the adjacent time intervals and a contact point between the adjacent time intervals; Of the pulse train signal so that the first change point of the signal in the second time interval that is later in time of the time interval and the second time between the signal and the contact point are the same as each other. The signal conversion method according to claim 1, wherein Each of a plurality of pulse train signals is converted into a sequence of the basic pattern and time data added to each basic pattern, and at least a sequence of the basic pattern corresponding to a first pulse train signal of the plurality of pulse train signals. When a part matches the sequence of the basic pattern corresponding to the second pulse train signal of the plurality of pulse train signals, the at least a part of the sequence of the basic pattern corresponding to the first pulse train signal is 3. The signal conversion method according to claim 1, wherein the description is replaced with a notation indicating that the sequence matches the sequence of the basic pattern corresponding to the second pulse train signal.

Images