1: \begin{abstract}

2: Elliptic partial differential equations must be solved numerically for

3: many problems in numerical relativity, such as initial data for every simulation

4: of merging black holes and neutron stars.

5: Existing elliptic solvers can take multiple days to solve these problems at high

6: resolution and when matter is involved, because they are either hard to

7: parallelize or require a large amount of computational resources.

8: Here we present a new solver for linear and nonlinear elliptic problems that is

9: designed to scale with resolution and to parallelize on

10: computing clusters. To achieve this we employ a discontinuous Galerkin

11: discretization, an iterative multigrid-Schwarz preconditioned Newton-Krylov

12: algorithm, and a task-based parallelism paradigm. To accelerate convergence of the elliptic

13: solver we have developed novel subdomain-preconditioning techniques.

14: We find that our multigrid-Schwarz preconditioned elliptic solves achieve

15: iteration counts that are independent of resolution, and our task-based parallel

16: programs scale over 200 million degrees of freedom to at least a few thousand

17: cores. Our new code solves a classic initial data problem for binary black holes

18: faster than the spectral code \spec{} when distributed to only eight

19: cores, and in a fraction of the time on more cores.

20: It is publicly accessible in the next-generation \spectre{}

21: numerical relativity code.

22: Our results pave the way for highly parallel elliptic solves in numerical

23: relativity and beyond.

24: \end{abstract}

25: