~~CLOSETOC~~ {{TOC:wide}} \\ ====== Solvers ====== In Scan&Solve 2014, users have the ability to choose among 3 different algorithms for solving the linear system of equations that represent the structural simulation solution. The algorithms are identified by the following abbreviations: - **DSS** - Direct Sparse Solver - **ISS** - Iterative Sparse Solver - **SnS** - Scan&Solve Solver For the numerically-inclined, the DSS and ISS solvers are from [[https://software.intel.com/en-us/intel-mkl|Intel's Math Kernel Libraries]]. ===== Selecting a Solver ===== Switching among solvers is done in the Linear Solver group of the Settings tab. Once a solver is chosen, the selection is stored for subsequent simulations on that part. {{:switch_solvers_1.png}} ---- ==== Option 1: Direct Sparse Solver ==== The **direct sparse solver** is the fastest solver, but it can also consume large amounts of memory. Because of this memory consumption, the DSS solver works best on simulations requiring resolutions less than 200000 - at least on typical computer hardware possessing 8-12GB RAM. ---- ==== Option 2: Iterative Sparse Solver ==== The **iterative sparse solver** is of intermediate speed and is relatively frugal with memory. It is a faster option than the default Scan&Solve solver but it doesn't support animation of the solution process. ---- ==== Option 3: Scan&Solve Solver ==== The **Scan&Solve solver** is a robust iterative solver that supports animation of the solution process while being very frugal with memory. During the solution phase, intermediate solutions are iteratively improved until they meet the precision setting on the Settings tab. Because the intermediate solutions do not always progress directly to the solution, the position of the progress meter in the Scan&Solve Specify tab can vary back and forth. The tolerance for terminating the iterations is adjusted by the Precision spinner. The default setting of 1e-09 is a good value for most simulations. By reducing the precision, solutions will be delivered faster but the accuracy of the solutions may suffer.