In this section, we bring our activities together as the integrated approach outlined in Figure 1. We must make choices from the many different applications and target machines in order to provide a project focus. We will work with DARPA and the HPCC community in choosing exemplars that are synergistic with other activities. For the near future, we have as a natural baseline choice the machines chosen for the DOE ASCI program, and for a ten year horizon the target architectures from the PetaFlop process that were reported at the Frontiers'96 conference.
We expect that it will be straightforward to collaborate as a team, since we have great experience in doing this with other projects. As well as the standard (and increasingly sophisticated) electronic mechanisms, we expect that the team will need to hold meetings on a semi-annual basis. These will be especially important at the start of the project when more frequent get-togethers may be necessary - some of these will be at DARPA PI meetings.