There are two hydro projects right now: reviving an old perturbation code, and writing a fully relativistic 3D hydro code (the "R" code).The perturbation code is a project that was originally done about fifteen years ago...many PhD theses have come out of the this code, though mine probably won't be one of those. Essentially, it takes a spherically symmetric hydrodynamic model of something interesting (say, a stellar core collapse), and adds a nonspherical perturbation in order to compute the gravitational wave signal of the collapsing object. We do things this way because a spherically symmetric system cannot produce any gravitational waves, but a nonspherical hydrodynamical simulation is much more difficult to program.
The R code will be a general-purpose program to compute 3D fluid flows in regions of large gravitational fields. Eventually, this will be attached to the G code that this group is developing, to calculate the dynamics of systems with both fluid flows and large gravitational fields--say, for example, the collision of two neutron stars to form a black hole.
Perturbation Code The Perturbation code has been revived, and using the hydrodynamics from the olld hydro code that was written to work with it, it does very well, exactly reproducing the results of those runs. With Doug Swesty's more advanced hydro, though, it doesn't do very well at all. It generates large amounts of noise in the signal, and it isn't clear where the noise is coming from. Work continues (at a slow pace) to debug this problem, since once this is figured out, this project will be essentially complete.R Code The R code has been written, and passes a 1D relativistic and nonrelativistic shock tube problem easily. The only error in the relativistic case appears to be from the simplistic artificial viscosity scheme we're using; use of a full tensor AV should clear up the problem, though we might not bother since this is well outside the regieme where we expect to be running the code in the near future. It remains to finish testing the code on a series of progressively more difficult test problems, then we will attach it to the G code.