Collaborative Research: Elements: Portable Resistive GRMHD Cyberinfrastructure for Multi-Messenger Astrophysics

Mergers of neutron stars and pairs of supermassive black holes are among the most energetic events in the universe. They can produce gravitational waves, bright flashes of light, and fast jets; neutron-star mergers can also create heavy elements found in planets and living things. Scientists use large computer simulations to understand the hot, magnetized gas around these objects, but many current simulations treat that gas as a perfect electrical conductor. This simplification makes magnetic-field reconnection, a key source of heat, light, and outflows, occur as an uncontrolled byproduct of numerical error rather than through a physically modeled process. This project develops open-source software that lets researchers model magnetic energy release with physically meaningful inputs and reproducible tests. The project advances the national interest by improving computational tools for scientific discovery, strengthening high-performance computing, and training students in computational science. It also broadens participation in science through activities that include research experiences for Deaf and Hard of Hearing students, outreach to rural K-12 students in Idaho, and public visualizations of cosmic collisions through accessible performances and outreach events. This project develops portable cyberinfrastructure for resistive general-relativistic magnetohydrodynamics (GRMHD), the computational framework used to simulate magnetized, electrically resistive fluids in strong gravity. A team of researchers extends the open-source GRHayL library with fourth-order, low-diffusion numerical methods and a resistive GRMHD module with two runtime-selectable conductivity prescriptions. The software uses a vector-potential formulation to preserve the divergence-free magnetic-field constraint, implicit-explicit time integration to handle stiff electric-field relaxation, and diagnostics that compare numerical resistivity with the prescribed physical conductivity. A human-supervised workflow uses large language models to assist code development, testing, documentation, review, and porting while leaving scientific and algorithmic decisions to human researchers. The new capabilities are integrated into the AsterX software system for production simulations and into BlackHoles@Home/GRoovy for single-node development, testing, optimization, and teaching. The project verifies the shared kernels through standard tests, cross-code comparisons, and central processing unit (CPU) and graphics processing unit (GPU) parity tests. Public deliverables include tagged software releases with digital object identifier snapshots, Apptainer containers, verification inputs and reference outputs, documentation, tutorials, and curated benchmark and diagnostic data products. This award by the Office of Advanced Cyberinfrastructure is jointly supported by the Physics at the Information Frontier Program in the Physics Section of the Directorate for Mathematical and Physical Sciences. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.