Planning

The planning is such that in the first two years of the project most work lies in subjects (1) and (2), with some first work carried out in subject (3). In the second half of the project this will be the other way around.

Year 1

Phase 1 of N-body lab hardware: GRAPE attached to single workstation; detailed design of the cluster of GRAPES. Realization of two kernels: direct method and 1st order Barnes-Hut. Realization of combined stellar evolution with stellar dynamics code.

Year 2

Phase 2 of N-body lab hardware: addition of cluster of general purpose computers, the DAS node. Parallelization over cluster, focusing on the tree codes; general purpose versus dedicated computing. Upgrading of stellar evolution and stellar dynamics code. Production runs on the evolution of star clusters with a few ten-thousand stars, and a variety of initial fractions of binaries.

Year 3

Phase 3 of N-body lab: addition of more general GRAPE (if available) to handle other than 1/r potentials; addition of GRAPE boards to several nodes of the cluster. Higher order Barnes-Hut methods, and/or other efficient N-body kernels in the N-body lab. More extended production runs on systems of tens of thousands of stars, with a variety of fractions of binaries (large open clusters or small globulars).

Year 4

Final production runs on "real" globular clusters or Galactic nucleus (a million stars), which start out with a variety of initial fractions of binary systems; study of the formation and evolution of single and binary compact stars (neutron stars, black holes), over a period up to the age of the Universe ($\sim 10^{10}$ years).