Toward NRPyElliptic-Multipatch: Conformally Curved Binary Black Hole Initial Data for Next-Gen Numerical Relativity

Conformally flat (CF) binary-black-hole initial data (BBH ID) impose a spin cap (χ<0.93) and can exhibit larger early-time unphysical transients ("junk" radiation) when evolved forward in time, limiting precision BBH gravitational-wave inference. We present progress toward a multipatch version of NRPyElliptic---packaged as a lightweight, dependency-free library in the spirit of the BHaHAHA apparent-horizon finder---that leverages BlackHoles@Home multicoordinate, multipatch grids to generate conformally curved (CC) BBH ID in ~1 GB of memory. This extends the original single-patch CF solver to CC with multipatch capabilities, enabling high-mass-ratio CF and CC configurations while remaining straightforward to integrate with existing numerical relativity (NR) codes. The Einstein constraints are solved via hyperbolic relaxation, accelerated by coarse-to-fine (multigrid-inspired) sequencing and locally accelerated wave speeds matched to local grid spacing, yielding fast, robust convergence from poor initial guesses. These design choices enable state-of-the-art BBH ID on systems ranging from desktop computers to individual MPI ranks in HPC simulations and provide a drop-in interface for current NR infrastructures. Ongoing work targets GPU acceleration of the multipatch backend, positioning NRPyElliptic-Multipatch as a step toward code-agnostic, community-scale ID infrastructure and increasingly realistic NR simulations.