AbstractsChemistry

One of the goals of modern quantum chemistry is to simulate actual chemical experiments. In order to study species closer to real life systems and bulk environments there is a need for methodological developments. There are two ways to approach large systems with a given level of accuracy: conceptual changes to quantum chemistry methods or algorithmic developments for current methods. Many scientists believe that the conceptual changes truly increase the size of the systems one can study. With more or less advanced approximations to the method it is possible to increase the efficiency of calculations orders of magnitude. The implementation and algorithms fall down in the priority list, as advanced algorithmic developments are time consuming and usually lead to lower efficiency increases than conceptual changes. In this work it is shown that algorithmic developments cannot be neglected, and that even simple changes help in utilizing the power of modern computers and can also increase the efficiency by orders of magnitude. In this work new algorithmic developments are presented and used for solving various timely chemical problems. One of the goals of modern quantum chemistry is to simulate actual chemical experiments. In order to study species closer to real life systems and bulk environments there is a need for methodological developments. In this work a series of methodological developments was carried out and directed into efficient use of modern super-computers. Developments on a borderline between theoretical chemistry and computer science were used for solving various timely chemical problems in both, fundamental and more applied atmospheric parts of the field of chemistry. Problems that previously were hard to even approach were solved using a combination of smart algorithmic approaches and modern Finnish super-computing infrastructure.