The optimal solver for a given pr oblem depends not only on the equations being solv ed\, but the boundary conditions\, discretization\ , parameters\, problem regime\, and machine archit ecture. This interdependence means that \\textit{a priori} selection of a solver is a fraught activi ty and should be avoided at all costs. While there are many packages which allow flexible selection and (some) combination of linear solvers\, this un derstanding has not yet penetrated the world of no nlinear solvers. We will briefly discuss technique s for combining nonlinear solvers\, theoretical un derpinnings\, and show concrete examples from magm a dynamics. \; \;

The sa me considerations which are present for solver sel ection should also be taken into account when choo sing a discretization. However\, scientific softwa re seems even less likely to allow the user freedo m here than in the nonlinear solver regime. We wil l discuss tradeoffs involved in choosing a discret ization of the magma dynamics problem\, and demons trate how a flexible mechanism might work using ex amples from the PETSc libraries from Argonne Natio nal Laboratory. LOCATION:Seminar Room 1\, Newton Institute CONTACT:info@newton.ac.uk END:VEVENT END:VCALENDAR