High performance grid computing: Getting HPC and HTC all together

The recently developed Grid Framework, GriF, is a Workflow Management System designed to evaluate Service and User Quality when helping the Grid users (especially those of the COMPCHEM Virtual Organization) in optimizing the return of the parameter sweeping studies. In this perspective, GriF has now been extended to allow access to High Performance supercomputers from the Grid. The implementation of such High Performance Grid has requested a tight collaboration of various experts of the Italian Grid Initiative to the end of bridging the gLite middleware with that of some supercomputer platforms. The advantages achievable in this way are illustrated by discussing the use case of atom-diatom quantum reactive scattering calculations.