Modeling the radiation generated by accreting matter is an important step towards realistic simulations of black hole accretion disks, especially at high accretion rates. To this end, we have recently added radiation transport to the existing general relativistic magnetohydrodynamic code, Cosmos++. However, before attempting to model radiative accretion disks, we have tested the new code using a series of shock tube and Bondi (spherical inflow) problems. The four radiative shock tube tests, first presented by Farris et al. (2008), have known analytic solutions, allowing us to calculate errors and convergence rates for our code. The Bondi problem only has an analytic solution when radiative processes are ignored, but is pertinent because it is closer to the physics we ultimately want to study. In our simulations, we include Thomson scattering and thermal bremsstrahlung in the opacity, focusing exclusively on the super-Eddington regime. Unlike accretion onto bodies with solid surfaces, super-Eddington accretion onto black holes does not produce super-Eddington luminosity. In our examples, despite accreting at up to 300 times the Eddington rate, our measured luminosity is always several orders of magnitude below Eddington.

Using optimal matched filtering, we search 25 hours of data from the LIGO 40-meter prototype laser interferometric gravitational-wave detector for gravitational-wave chirps emitted by coalescing binary systems within our Galaxy. This is the first test of this filtering technique on real interferometric data. An upper limit on the rate R of neutron star binary inspirals in our Galaxy is obtained: with 90% confidence, R< 0.5/hour. Similar experiments with LIGO interferometers will provide constraints on the population of tight binary neutron star systems in the Universe.