Contribution to the Methods and Algorithms of Automatic Generation of Functional Tests for Processors

The dissertation thesis is aimed at automatic functional test generation methods for processors. Testing processors needs continually new test generation methods, algorithms and test application techniques for their verification, manufacture testing and reliable life-time run. The functional tests are mainly generated over an instruction set architecture and processor model description. They run in similar form and frequency as real programmes. Such tests are categorized as the software-based tests. A metric for quality evaluation of the software-based tests is obviously done by code coverage of a described processor model. A simulation tool is used with the mentioned metric calculation. A new test generation method is presented in the dissertation thesis. It uses VHDL model of processors and genetic algorithms with adaptation of various evolutionary strategies. In addition fitness function is defined for evaluation of generated tests by genetic algorithms and a new method for creation of a starting test (the first population) has been proposed. The described methods are suitable for automatic generation of software-based tests for processors. Functionality and effectiveness of the developed methods were evaluated in implemented software system AGenMIX over one type of RISC processors. Two representative experiments are shown in the paper.