Directed vs Random & the Coverage-Driven Loop
VLSI → Coverage-Driven Verification
Two testing strategies, closed into one loop
Directed tests check cases you already thought of. Constrained-random tests explore the ones you didn't. Coverage-driven verification (CDV) closes the loop between them: randomize stimulus, measure what you actually hit with coverage, then write more constraints or directed tests to fill the holes — repeating until coverage is closed.
Why this is how modern DV teams actually sign off
"How do you know you've tested enough?" is a core interview question — "I close coverage" is the answer that signals real verification experience, because it means you have an objective, measurable stopping condition instead of just "we ran it a lot and it seemed fine."
Try it — step through one loop iteration
Run constrained-random tests against the current RTL — not one test, a whole regression of randomized seeds.
Completion saves automatically after you read through to the end.
Next lesson →