Challenges in Numerical Relativity: Intermediate mass-ratio binaries

As gravitational wave detectors continually improve, numerical relativists must also improve their methods to generate wavesforms of sufficient accuracy for gravitational wave analysis and tests of general relativity.  The large computational cost in generating waveforms comes from both the high level of adaptability needed for intermediate mass-ratio black hole binaries, as well as the long run-times that are needed to produce waveforms that can be matched to post-Newtonian waveforms.  Dendro-GR is a new computer code for solving the Einstein equations, which was designed to anticipate some of these challenges.  Dendro-GR solves the Einstein equations on an unstructured grid using wavelet-based refinement.  It has achieved high scalability and computational performance with binary black hole runs up to mass ratio $q=32$,  and results of these simulations will be presented.  We will also discuss new directions to further improve Dendro-GR's computational performance, such as new methods for time adaptivity and improved finite-difference stencils.  Some initial results of these tests will be presented.