NonlinearSolve.jl introduces a suite of highly efficient open-source nonlinear equation solvers in Julia, addressing the scalability challenges faced in scientific and engineering disciplines. The tool stands out with its unified API, runtime-based algorithm selection, GPU-accelerated computation, and support for large-scale problems using sparse automatic differentiation and Jacobian-free Krylov methods. Outperforming established tools like Sundials and MINPACK, NonlinearSolve.jl offers unparalleled robustness and efficiency in solving benchmark and real-world problems, unlocking new possibilities in modeling and simulation in various fields. A significant advancement that researchers and practitioners can benefit from in their computational toolkit.
https://arxiv.org/abs/2403.16341