This chapter presents experimental results obtained from true circuits. In most cases the results obtained from deterministic and random testing are compared.

We insist on the fact that these results are not obtained from simulation but from circuits. Simulation is an extremely powerful tool which is very efficient for example for calculating a fault coverage or estimating detection probabilities. However, it is not adapted to comparing random and deterministic tests. As a matter of fact, a strong point of random testing is its ability to detect faults even if they are not targeted (Property 5.2 in Section 5.3.2). On the other hand, any comparison between random and deterministic testing based on simulations is biased since a simulation can only take into account faults devised by somebody.

We first present results obtained by the author’s team, in cooperation with various companies, in Sections 11.1 to 11.3. Other results presented in the literature are then given in Section 11.4.