Solver algorithms ================= Reference for the analysis types and backend algorithms femorph-solver ships. The :doc:`user-guide ` section walks through how to drive each solver; this page lists the underlying algorithm and its sources. .. toctree:: :maxdepth: 1 :caption: Analysis types static modal cyclic transient .. toctree:: :maxdepth: 1 :caption: Backends linear_backends eigen_backends Quick reference table --------------------- .. list-table:: :header-rows: 1 :widths: 20 26 30 24 * - Analysis - Entry point - Algorithm - Backend chain * - :doc:`static` - :meth:`Model.solve ` - Sparse direct factor + back-substitute - :doc:`linear_backends` * - :doc:`modal` - :meth:`Model.modal_solve ` - Shift-invert Lanczos / block-Davidson / LOBPCG - :doc:`eigen_backends` × :doc:`linear_backends` * - :doc:`cyclic` - :meth:`Model.cyclic_modal_solve ` - Per-harmonic real 2n-augmented Lanczos (Grimes-Lewis-Simon 1994) - :doc:`eigen_backends` × :doc:`linear_backends` * - :doc:`transient` - :meth:`Model.transient_solve` (roadmap) / modal-superposition (TA-14) - Newmark-β / generalised-α (direct); modal-superposition (decoupled SDOFs) - :doc:`linear_backends` How the dispatch chains plug together ------------------------------------- * :doc:`linear_backends` — one chain (Pardiso → CHOLMOD → MUMPS → UMFPACK → SuperLU) used by every analysis that factors a sparse matrix. * :doc:`eigen_backends` — modal / cyclic-modal pick ARPACK / PRIMME / LOBPCG. The shift-invert path inside the Lanczos / block-Davidson loop calls back into the linear chain, so the same Pardiso / CHOLMOD factor accelerates every shift-invert matrix-vector product. See the per-chapter cite blocks for full algorithmic references. .. seealso:: :doc:`/user-guide/solving/index` — the user-facing how-to for each entry point.