JPH05215819A - Timing set allocation treating method - Google Patents

Abstract

PURPOSE:To allocate timing sets in reduced manhours while taking users into consideration by inputting specifications of various kinds of testers and information on the user designated timing sets, and allocating automatically the timing sets according to specifications of tests. CONSTITUTION:Data being a treating object are taken in from a test specification data file 10 from which respective signals are outputted in a periodic unit of respective tests and to which operation data values for tester resources are inputted, a tester specification data file 11 to which specifications of various kinds of testers are inputted and a timing set data file 12 being designated by users, and a timing set allocation treatment 13 is carried out so as to satisfy specifications of tests, and the result is outputted to a file 14. This output result becomes an input data for a test program.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a timing set distribution processing method in a process of automatically generating a test specification which is input data when a test program for testing an LSI is generated.

[0002]

2. Description of the Related Art Japanese Unexamined Patent Publication (Kokai) No. 60-80299 discloses a method for grouping similar parts into one group to reduce the number of setups for the inserter.

[0003]

However, in the above-mentioned conventional technique, the limitation at the time of creating a group is that the number of parts belonging to one group is limited, and the finally obtained number of groups is not limited. Moreover, the user's wishes could not be reflected.

The present invention provides a method for automatically performing timing set distribution processing based on specifications of various testers and user-specified timing set information as input data. The purpose is to

[0005]

According to the present invention, the timing set is automatically distributed according to the tester specifications which are the input data. Therefore, by changing the tester specifications, it is possible to distribute the timing set corresponding to a plurality of types of testers.

[0006]

EXAMPLES Examples of the present invention will be described below.

FIG. 1 is a system configuration showing an embodiment of timing set distribution processing according to the present invention. Test specification data file 1 in which the operation data value to the tester resource for outputting each signal in each test cycle unit is input
0, the tester specification data file 11 in which the specifications of various testers are input, and the timing set data file 12 specified by the user are taken in, and the timing set distribution processing is performed so as to satisfy the test specifications. 13, the result is output to the file 14. This output result becomes the input data for generating the test program.

[0008]

[Table 1]

Table 1 is an example of the test specification data file among the three input data described in FIG. In each test cycle unit, these consist of the type of tester resource used in the test cycle, the value set in the resource, the variable if defined, and an arithmetic expression for obtaining the value. The timing set represents a set of data set in the tester resource in units of this test cycle. In the present invention, the value of each test cycle for each tester resource is referenced to check whether the same timing set can be distributed to each test cycle, and to allocate the number of timing sets to be used to the minimum. is there.

[0010]

[Table 2]

Table 2 is an example of a tester specification data file of the input data described in FIG. Restrictions for each tester model, such as the type and number of timing sets, rules for allocating timing sets, etc., are defined. Before executing the distribution processing, the target tester is determined, the information is read, and the processing is performed according to the information. If this data file is prepared for each tester model, even if the model of the tester changes, it is not necessary to change the internal processing by only changing the file read at the time of processing to the file for that tester.

[0012]

[Table 3]

Table 3 shows user-specified timing set data among the input data. This table consists of data for each test cycle, data for setting a user-specified flag specified in the test cycle, and items for inputting the timing set name to be actually set. There are four methods for specifying the timing set to be input in this table. One is single allocation. This means that the designated test cycle independently allocates one timing set even if a common timing set with another test cycle can be allocated. The second option is to specify no allocation. This is used when it is not necessary to allocate a timing set for a test cycle. The third is a case where the timing set number is designated by the test cycle. The last is the designation of common allocation. Used when specifying the same timing set in multiple test cycles.

The similarity between test cycles is used as an evaluation function for determining which test cycle and which test cycle can allocate the same timing set.

A method of calculating the similarity between the test cycles will be described. This is done using the test specification data file in Table 1. Table 4 shows the values set in the tester resource data in each test cycle.

[0016]

[Table 4]

What is set in the tester resource is not only an actual value, but also an arithmetic expression when the value is obtained, and a variable that has been set. Only the set value was used.

Here, the fact that there is a similarity between the test cycles means that one of the following three conditions is satisfied.

One is when the same value is set in the same tester resource, the second is when a value is set only in one test cycle, and the third is when neither value is set. .. If any of the three test resources is satisfied by all the test resources, the similarity between the two test cycles can be calculated. If all three conditions are not satisfied, that is, if two test cycles have different values set for the same tester resource, they cannot allocate one timing set. In this case, -1 is set based on the fact that there is no similarity between the two test cycles.

An example in which the above three conditions are satisfied for all test resources, and the number of tester resources for which no value is set in either tester resource in the two test cycles is used as the similarity. Is shown in Table 5.

[0021]

[Table 5]

That is, the similarity between the test periods TO_1 and TO_5 in Table 4 is 2 of TEST_RES_5 and TEST_RES_7.

In this method of calculating the degree of similarity, there are cases where both the value and the arithmetic expression are viewed, and even if the values are the same, if the arithmetic expression is different, the set value is different. This allows the user to select which similarity to use before executing this process.

The flow of the timing set distribution process using the above similarity calculation method will be described with reference to FIG.

First, 3 which becomes the input data of this processing at 21
Read two files. Next, at 22 the consistency of the user-specified specification file among the input files is checked. An unallocated flag is set 23 in each test cycle that is input data. The designated timing set is assigned to the designated period from the user designated information 24. The following processes from 26 to 28 are repeated until the timing set is distributed to all the test periods at 25.25 or the number of timing sets becomes insufficient. At 26, the similarity is calculated between the test cycles in which the unallocated flag is set, and one of the test cycles having the maximum combination of the similarities is selected, which is called a nuclear cycle.
At 27, the following processes from 29 to 210 are performed until there are no more test cycles that can be grouped with the nuclear cycle. Two
At 9, the similarity between the nuclear cycle and the unallocated test cycle is calculated.
The test cycle having the maximum similarity calculated in 210 is selected, and the test cycle is set as the same group. Further, new nuclear cycle data is created from the data of the test cycle and the data of the nuclear cycle, and the distributed flag is attached to the selected test cycle. When the processing of 27 is finished, the test cycle that can be defined by the same timing set is determined.

Using the data in Table 4, the process of allocating the actual timing set is performed. First, the similarity calculation result table shown in Table 5 is created from the data of each test cycle of Table 4 by the similarity calculation method described above. Next, the test cycle of the combination having the maximum similarity is obtained from the similarity calculation result table. In this case, the test cycles are TO_2 and TO_4. One of these TO_2 is registered in GROUP1 as a nuclear cycle. It is shown in Table 6.

[0027]

[Table 6]

[0028]

[Table 7]

Further, a distribution completion flag is set in TO_2. GROUP1
And the test cycle in which the unallocated flag is set, that is, TO_1, TO_
Calculate the similarity with 3, TO_4, TO_5. The results are shown in Table 7.

From the result, the test cycle TO_4 having the highest similarity to GROUP1 is selected. Put the data of the selected test cycle in GROUP1 and set the allocated flag in TO_4. This process is repeated, and when the similarity of the unallocated test cycle becomes all −1, the selection of the test cycle to which one timing set can be assigned ends. That is, the same timing set can be assigned to the test cycle included in GROUP1. This process is performed until the timing sets are distributed to all the test cycles or until the condition that the number of timing sets defined in the tester specification data file is not exceeded is satisfied. In this way, the number of timing sets can be minimized by searching for test cycles that can distribute the same timing set and locking them together.

[0031]

According to the present invention, the number of man-hours can be reduced because the timing set that satisfies the tester specifications can be automatically distributed in consideration of the user designation information.

[Brief description of drawings]

FIG. 1 is a system configuration diagram of the present invention.

FIG. 2 is a diagram showing a processing flow of a distribution processing apparatus of the present invention.

Claims (1)

[Claims] 1. In a device for generating a test program for testing the function of an LSI, in a process for automatically generating test timing data which is input data when a test program is generated, specifications of various testers and each test cycle unit A signal operation data and a means for fetching the specifications specified by the user as input data; a means for calculating the similarity to the resource of the tester during each test cycle;
A timing set distribution processing method, comprising means for allocating a minimum timing set by calculating a period that can be defined by the same timing set from the calculated similarity.