First-principles calculations for materials science: their power and limitations

Summary form only given. The use of parallel computers and that of sophisticated computational algorithms has made first-principle calculations feasible also for realistic models of systems of technological interest. Several successes have been obtained recently. In spite of this, there are still a number of problems to be solved, before this type of approach can assume a leading role in the investigation of materials. This talk will report on recent applications of parameter-free molecular dynamics to a variety of systems, with emphasis on the method, on the comparison of the results with experiment, on its useful outputs and also on its current limitations.