Parameters and algorithms have been improved for geometry optimizations in delocalized coordinates. This means that fewer steps are needed, especially for weakly bound or floppy systems and strict convergence demands. The optimization in delocalized coordinates can also be used in Transition State searches.
