David Neilsen

Main Research Areas

• Binary black holes and neutron stars
• Numerical General Relativity
• Computational Physics

I study the dynamics of merging black-hole and neutron-star binaries. At the time of merger, these binaries generate intense gravitational waves that can now be detected. I use numerical methods to solve the Einstein equations of general relativity to model these binaries, and predict the pattern of gravitational radiation. We use supercomputers to run large-scale simulations of the mergers.

I enjoy developing new computational methods for general relativity. I have been working with collaborators to develop the new Dendro-GR code for relativistic astrophysics. This code uses wavelets to construct an efficient unstructured, adaptive computational grid. The Dendro computational framework gives us excellent scaling and performance on massively parallel supercomputers.

A Brief CV

A World Line in Numerical Relativity

I first became intrigued with general relativity as an undergraduate student at BYU, when I realized that my physics classes hadn't covered gravity beyond Newton's Law of Gravitation in the first year. I tried to remedy the situation by taking a graduate general relativity class taught by Kent Harrison in my third year. While I had bitten off more than I could chew, I became fascinated by Einstein's insight that gravitational phenomena could be explained as the curvature of spacetime.

I continued studying general relativity in graduate school, first as a masters student with Kent Harrison, and then as a PhD student with Matthew Choptuik at the University of Texas at Austin. At Texas, I began learning numerical relativity, the art and science of using computers to solve the Einstein equations. It was the best of times, because of the excitement as many groups were working hard to solve the Einstein equations for generic black hole mergers—and it was the worst of times, because no one knew how to do it then. I did something a little different and studied gravitational critical collapse in perfect fluids for my dissertation project.

After graduation, I began work on the black hole merger project as a postdoctoral researcher in Austin with Richard Matzner. We worked on the Agave code at the time, in the days when black hole simulations were done on grids with 65³ points on Cray T3Es. After a few years in Austin, I moved across the state line to Baton Rouge to work with Luis Lehner at LSU. We used numerical analysis to find better techniques for solving the equations. We began developing the HAD code during this time, resulting in a collaboration that has lasted many years.

I began teaching at BYU in 2004, where I have been ever since. I still enjoy studying black holes and neutron stars with numerical relativity, though my interests sometimes fluctuate between the two. While working on neutron star mergers, I realized that our conventional numerical algorithms were not performing well on the new computer hardware. I became interested in developing algorithms for the next generation of supercomputers. I became intrigued with using wavelets to generate sparse computational meshes. This led to new opportunities, and eventually a very productive collaboration with Hari Sundar in the School of Computing at the University of Utah. We are developing the new Dendro-GR code to meet the challenges in numerical relativity over the next decade.