Overview
Markov Model Stimulus Generation refers to the use of Markov-model-based techniques to guide stimulus generation during hardware verification. In the cited evidence, this approach appears as feedback-adjusted Markov Models used by the tool StressTest to verify microprocessors.
The broader verification problem addressed by this class of techniques is automated test application: given available directed or parameterized constrained-random sequences and functional coverage goals, the goal is to apply stimuli in a way that improves verification coverage across multiple trials.
Feedback-adjusted Markov-model flow
The evidence describes StressTest as a tool based on feedback-adjusted Markov Models. Its stimulus-generation process is closed-loop:
- An engineer provides a template describing points of interest in the design under verification, also called probe nodes.
- During simulation, the tool observes activity at the probe nodes.
- Closed-loop feedback techniques adjust or direct the test-generation engine.
- The generator is steered toward stimuli that produce higher switching activity at the probe points.
This makes the approach an on-the-fly optimization method for stimulus constraints rather than a purely static or unguided random-generation method.
Role in verification automation
The evidence places this approach in the test-application phase of design verification automation. Automation is described as important for handling modern design complexity and the growing number of test scenarios. In particular, automated test application aims to achieve coverage closure with reduced regression time.
Reported benefit
The cited source reports that StressTest achieved better coverage in fewer cycles than random generation techniques when using feedback-adjusted Markov Models and probe-node activity feedback.