Impact of Model Size on Test Generation
One of the measures to track the size of a model when it comes to test generation complexity is the number of test targets.
Number is shown after model loading in Test Targets view
The more test targets the longer model loading will take (model loading time increases linearly)
The more test targets a model has the longer it will take to generate tests.
It is important to understand that
There is no general relationship between the number of test targets to test generation time but there is a relationship based on modeled functionality
o Assume a model A with 500 targets which takes x seconds to generate tests and model B (modeling different functionality) also with 500 targets taking y seconds
o A and B are not likely to take the same time for test generation (x ≠ y)
o But, if A is modified to only have 400 targets then A will take less time to generate; similarly if targets are increased to 600 it will take longer. Note: Assumption here is for the same test data and advanced option settings!
Recommendation: When extending modelled functionality, split the model into multiple models once the model complexity exceeds the agreed definition (For e.g., test generation exceeds “reasonable test generation time” or a limit like 2,000 targets). |
Related content
Copyright © 2023 Conformiq and its subsidiaries. All rights reserved.