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.