The final merger of two black holes will emit more energy than all the stars in the observable universe combined. This energy will come in the form of gravitational waves, which are a key prediction of Einstein's general relativity and a new tool for exploring the universe. Observing these mergers with gravitational wave detectors, such as the ground-based LIGO and the space-based LISA, requires knowledge of the radiation waveforms. Since these mergers take place in regions of extreme gravity, we need to solve Einstein's equations of general relativity on a computer.

For more than 30 years, scientists have tried to compute black hole mergers using the methods of numerical relativity. The resulting computer codes were long plagued by instabilities, causing them to crash well before the black holes in the binary could complete even a single orbit. Within the past few years, however, this situation has changed dramatically, with a series of remarkable breakthroughs. This talk will focus on new simulations that are revealing the dynamics and waveforms of binary black hole mergers, and their applications in gravitational wave detection, testing general relativity, and astrophysics.