RANDOM TESTING OF THE DATA PROCESSING SECTION OF A MICROPROCESSOR

The principle of random testing is as follows : a random input sequence is applied s imultaneously to both a circuit under test and a reference circuit. The outputs are compared. In t he case of microprocessors, we apply a sequence of random instructions with random data. The research aim is to calculate the test length (i.e. the number of instructions) for given instruction probabilities. The paper deals with the faults in the data processing section, which is modeled by a qraph containing registers and operators. We consider functional level fault models in registers and operators. We use the notion of Minimal Detectinq Transition Sequences (MDTS) .