Abstract

The ability to theoretically model chemical and biological processes is a key to understand nature and to predict experiments. Unfortunately, this type of computational modeling is very data and computation extensive. However, the worldwide computing grid can now provide the necessary resources. It is therefore a primary goal of current research to utilize these facilities. Here, we present a coupling of the GAMESS quantum chemical code to the Nimrod/G grid distribution tool. As an example, it is applied to the parameter scan of an effective group difference pseudopotential (GDP). This represents the initial step in the parameterization of a capping atom for hybrid quantum mechanics-molecular mechanics (QM/MM) calculations of complex molecular systems. The results give hints to the underlying physical forces of functional group distinctions and provide starting points for later parameter optimizations. The technology demonstrated here significantly extends the manageability of accurate, but costly quantum chemical calculations and is thus valuable for a wide range of applications which involve thousands of independent runs.