GLS · Topic
X-Propagation in Gate-Level Simulation
Where X comes from in gate-level simulation, how it spreads, and the pessimism and optimism that make RTL and gate-level disagree.
X in gate-level simulation is not a bug in itself — it is the simulator being honest that a value is unknown. The difficulty is that RTL simulation is often optimistic where gate level is pessimistic: an RTL `if` with an X condition quietly takes the else branch, while the equivalent gates produce X on the output. So a design that looked clean at RTL can light up at gate level, and the difference is a modelling artefact rather than a new defect.