Part 5 answers: How do I gain confidence the design is right, not just lucky?
One passing test proves little; confidence comes from a ladder where each rung catches bugs the lower rungs miss. This part climbs from tests that can fail through self-checking, random stimulus, assertions and coverage to a first look at formal methods.
Every rung is motivated by a real mini_controller bug that the rungs below it let through.
What | Detail |
|---|---|
Time | About 2.5 hours. |
You will build | A self-checking random testbench that catches planted bugs, with assertions and a coverage report. |
You are done when | Your testbench fails on every bug mutant and the coverage report shows no unexplained holes. |
Write a self-checking test with a reference model that fails on every real-bug mutant.
Add assertions and coverpoints that catch a bug directed simulation missed.
Explain what formal asks that simulation cannot answer.
# | Lesson | What it covers | Level |
|---|---|---|---|
5.1 | Nine planted bugs show why a pass alone proves nothing. | W | |
5.2 | Build the checker stage by stage against the golden model. | W | |
5.3 | Why one seed proves little, and how to replay a failure. | W | |
5.4 | One rule written as a property, and when it fires. | W | |
5.5 | Read the hole list and close it. | W | |
5.6 | Orientation only: what formal tools prove and when to ask for one. | W |
Lessons are in learning order: each one sets up the problem the next one solves. The glossary and the command cheat sheet are open to you throughout; if something fails, go to the troubleshooting table.
This course stays on the front-end start-up path and links out instead of repeating. These lessons build on or continue what this part teaches.