Test vector generation method

Abstract

The invention relates to a method for generating test vector, which comprises: building an orthogonal table database; according to the test conditions, finding input level values and relative input parameters; selecting the orthogonal tables with same level values, and using highest input level value as the initial present input level value; finding the correct orthogonal table in the orthogonal table group according to present input level value; judging if the met orthogonal table subgroup is found in the orthogonal table group, if it is, judging if present input level value is the lowest level value, if it is, using the orthogonal table with minimum test times as the result to be feedback, or else, continuously searching in the results according to the second lower input level value, or completing the process; then generating test vector according to the result orthogonal table to process test. The invention can reduce the number of vector, the test time and cost, with confirmed effect.

Description

technical field [0001] The invention relates to the field of testing, in particular to a method for generating test vectors in performing ASIC verification or software testing. Background technique [0002] During application-specific basic circuit (ASIC) verification or software testing, various parameters need to be considered. There are many combinations of different values ​​of various parameters, and each combination is called a test vector. Suppose there are 3 parameters A, B, C, and each parameter has 3 values ​​A1, A2, A3; B1, B2, B3; C1, C2, C3, then their combinations are as many as 3 3 = 27 types, that is, there are 27 test vectors. If these 27 test vectors are tested during ASIC verification or software testing, it will consume a lot of resources and time, and the cost of testing is very high. [0003] In the actual test process, it is often encountered that some parameters have a wide range of values. For example, the packet length of an Ethernet packet may a...

AI Technical Summary

Problems solved by technology

[0008] It can be seen from the above technical solutions that the existing technology generates test vectors by random means, which cannot effectively generate test vectors for the test target. Although the randomness can be constrained, in practice, the random coverage of the test vectors often cannot expected result

[0009] Moreover, the existing methods do not have certain rules and theoretical basis when defining the test target. For the combination of various parameters, due to the huge number of combinations, it is impossible to test each combination. Should appear, can not give a reasonable goal, and does not give a theoretical basis for the generation method

Moreover, the existing technology can only constrain a single parameter, and the constraints of each parameter are irrelevant. In fact, in most cases, a system is formed by the combination of multiple parameters. The existing technology can only pass enough random parameters. To achieve the coverage target, with the increase of test vectors, the test time and cost will be higher

Images