Journal ArticlesuperB/NRPy: scalable, task-based numerical relativity for 3G gravitational wave scienceJadoo, Nishita (ORCID:000900031699277X); Jacques, Terrence Pierre (ORCID:0000000289930567); Etienne, Zachariah B (ORCID:0000000268389185)Modern gravitational-wave science demands increasingly accurate and computationally intensive numerical relativity (NR) simulations. The Python-based, open-source<monospace>NRPy</monospace>framework generates optimized C/C++ code for NR, including the complete NR code<monospace>BlackHoles@Home</monospace>(<monospace>BH@H</monospace>), which leverages curvilinear coordinates well-suited to many astrophysical scenarios. Historically,<monospace>BH@H</monospace>was limited to single-node<monospace>OpenMP</monospace>CPU parallelism. To address this, we introduce<monospace>superB</monospace>, an open-source extension to<monospace>NRPy</monospace>that enables automatic generation of scalable, task-based, distributed-memory<monospace>Charm++</monospace>code from existing<monospace>BH@H</monospace>modules. The generated code partitions the structured grids used by<monospace>NRPy</monospace>/<monospace>BH@H</monospace>, managing communication between them. Its correctness is validated through bit-identical results with the standard<monospace>OpenMP</monospace>version on a single node and via a head-on binary black hole simulation in cylindrical-like coordinates, accurately reproducing quasi-normal modes (up to<inline-formula><tex-math><CDATA/></tex-math><math overflow='scroll'><mrow><mi>ℓ</mi><mo>=</mo><mn>8</mn></mrow></math></inline-formula>). The<monospace>superB/NRPy</monospace>-generated code demonstrates excellent strong scaling, achieving an ≈45× speedup on 64 nodes (7168 cores) compared to the original single-node<monospace>OpenMP</monospace>code for a large 3D vacuum test. This scalable infrastructure benefits demanding simulations and lays the groundwork for future multi-patch grid support, targeting long inspirals, extreme parameter studies, and rapid follow-ups. This infrastructure readily integrates with other<monospace>NRPy</monospace>/<monospace>BH@H</monospace>-based projects, enabling performant scaling for the general relativistic hydrodynamics code<monospace>GRoovy</monospace>, and facilitating future coupling with GPU acceleration via the<monospace>NRPy-CUDA</monospace>project.</p>Classical and Quantum Gravity2025-07-2310661788Classical and Quantum Gravity42151550060264-9381https://doi.org/10.1088/1361-6382/adee712508377; 2004311; 2411068; 2108072; 2409654National Science Foundation