First-Order Stokes Model With Glen’s Law Viscosity Available BCs: No-Slip Basal Sliding Stress-Free Open Ocean

1. Data-Centric Algs Graph Algorithms SVDs Map-Reduce Linear Programming Network Models Development of the Albany/FELIX Land Ice Dycore using Software Components A.G. Salinger, I. Kalashnikova, M. Perego, R.S. Tuminaro, M.S. Eldred and J.D. Jakeman, Sandia National Laboratories S. Price, M. Hoffman, Los Alamos National Laboratories UQ: Bayesian Calibration Dycore Interfaces and Meshes Convergence & Scalability Objectives Albany/FELIX Ice Sheet Dycore Component-Based Strategy The Albany/FELIX solver can be driven by a CISM or MPAS-LI interface: • Develop: robust and scalable unstructured-grid finite element ice sheet code: • Stand-alone steady-state model for initialization and calibration • Dynamic model when linked to MPAS-LI or CISM for advection • Future land ice component of DOE-ACME earth system model • Support:DOE climate missions, such as providing Sea Level Rise predictions • Leverage:software and expertise from SciDAC Institutes (FASTMath, QUEST, SUPER) and hardware from DOE Leadership Class Facilities • Funding:“PISCEES” SciDAC Application Partnership (DOE’s BER + ASCR divisions) • PIs: S. Price and E. Ng; collaboration with ORNL, LANL, LBNL, UT, FSU, SC, MIT, and NCAR CISM MPAS-LI LandIce_model simple_glide Albany/FELIX (C++) velocity solve CISM (Fortran) Thickness evolution, temperature solve, coupling to ESM MPAS/Land Ice (Fortran) Thickness evlolution, temperature solve, coupling to ESM C++/Fortran interface,mesh conversion C++/Fortran interface,mesh conversion CISM • Structured rectangles • Extruded to Hexs MPAS • Unstructured polygons • Dual mesh of triangles • Extruded to Tets We are beginning to do dynamic runs: Analysis Tools (black-box) Optimization Regional Refinement: Utilities Composite Physics UQ (sampling) Input File Parser MultiPhysics Coupling Parameter Studies Parameter List System Models Component-based approach enables rapid development of new production codes embedded with transformationalcapabilities V&V, Calibration Memory Management System UQ OUU, Reliability I/O Management Analysis Tools (embedded) Communicators Mesh Tools Mesh I/O PostProcessing Nonlinear Solver Inline Meshing Visualization Time Integration Partitioning Sandia’s components effort includes ~100 interoperable libraries Verification Continuation Load Balancing Model Reduction Sensitivity Analysis Adaptivity Stability Analysis Remeshing Mesh Database Software Quality Constrained Solves Grid Transfers Mesh Database Optimization Version Control Quality Improvement Geometry Database UQ Solver Regression Testing DOF map Solution Database Build System Linear Algebra T= 0 yr T=70 yr Local Fill Backups Data Structures Discretizations Verification Tests Iterative Solvers Discretization Library Physics Fill Mailing Lists Direct Solvers Field Manager Element Level Fill Unit Testing Eigen Solver Material Models Bug Tracking Derivative Tools Preconditioners Objective Function Performance Testing Sensitivities Matrix Partitioning Constraints Code Coverage Derivatives Scalability results over 4 mesh bisections: How many vertical layers do you need? Convergence study for GIS 1km mesh: Architecture- Dependent Kernels Error Estimates Porting Adjoints MMS Source Terms Web Pages UQ / PCE Propagation Multi-Core Release Process Accelerators 3D Mesh convergence study for GIS model gives theoretical 2nd-Order rate 4 cores 334K DOFs (8km GIS, 5 layers) 16384 cores 1.12B DOFs (0.5km GIS, 80 layers) 84 First-Order Stokes Model With Glen’s Law Viscosity Available BCs: No-Slip Basal Sliding Stress-Free Open Ocean Defining a UQ workflow for stochastic inversion of Basal sliding coefficients: 1. Model Reduction (KLE) 2. PCE Emulator 3. MCMC Calibration using Emulator Ongoing Work • Finite Element Discretization (Hex, Tet) • Parallel, Unstructured Grid with Partitioning • Automatic Differentiation for Jacobians • Globalized Newton’s Method Nonlinear Solves • Preconditioned Krylov Iterative Solvers • Performance-Portable Kernels (in progress) • Software tools: git / cmake / ctest / jenkins Solution Verification using manufactured solutions Robust Nonlinear Solves using Homotopy Continuation • Mature dynamic evolution capability under MPAS • Perform deterministic and stochastic initialization runs • Improve coupling to full earth system model • Finish conversion to performance-portable kernels • We acknowledge the contributions of our PISCEES collaborators, including B. Lipscomb, K. Evans, P. Worley, M. Norman, M. Gunzberger, and C. Jackson, and our many Trilinos/Dakota collaborators, including E. Phipps and L. Swiler g=10-10 g=10-1.0 g=10-2.5 g=10-6.0 g=10-10 SAND 2014-xxxxP