Software Engineering

Computational Science

• Benchmark Tool Development - Cigma (CIG Model Analyzer)
  • Collect all existing CIG benchmark cases, along with cigma comparison scripts, into a centralized location in svn repository (expected completion, January 2010)
  • Introduce Cigma as a TeraGrid scientific gateway project, providing ease of use (no installation necessary) and ability to just upload data and run error analysis remotely. (Expected completion, January 2010)
  • Incorporate regular benchmark tests into the CIG buildbots (Expected completion, January 2010)
  • Improve performance for simpler grid types, other than unstructured (Expected completion, January 2010)

Computational Seismology

• Currently we are developing SPECFEM3D_SESAME: a flexible 3D acoustic, (an)elastic, and poroelastic spectral-element solver with adjoint capabilities. Like SPECFEM2D, the new 3D solver will be able to take regular and unstructured meshes produced by Cubit and partitioned by SCOTCH, ZOLTAN, or METIS. (Expected completion, TBA)

More information:
Software Information
Repository

Short-Term Tectonics

• Short-Term Tectonics Priorities

More information:
PyLith Development Plans
PyLith/Lithomop Repository

Long-Term Tectonics

• Merge Gale back with the VPAC code. This should give Gale restart capability, and improved PETSc compatibility (Expected completion, January 2010)
• Deformed lower boundaries (Expected completion, January 2010)
• Interface with custom surface process models in parallel (Expected completion, January 2010)
• User-specified non-uniform mesh spacing (Expected completion, January 2010)

More information:
Gale Development Proposal and Plans
Software Information
Repository

Mantle Convection

• Implementing Anelastic Liquid Approximation (Expected completion, February 2010)
• Test ConMan on different systems and run benchmarks, compile with different tools (Expected completion, February 2010)
• Dave May's new approach with preconditioners using PETSc to CitcomS (Expected completion, Winter 2010)

More information:
CitComS
CitComCU
CIG-MC Working Group

Magma Dynamics

The stgMadds project has already developed benchmark (a) (magmatic solitary waves) for 2 and 3D solitary waves and has begun benchmark (c) for 2D ridge models with forced adiabatic melting.

Currently, VPAC has developed codes for
Milestone 1: Isoviscous Stokes in 2 and 3D for Ridge geometries with Q2-Q1 elements
Milestone 2: Isoviscous Stokes flow, zero permeability, non-zero porosity
Milestone 3: Post-processing for melt velocities
Milestone 4: Magmatic Solitary wave codes

and is working on coupling Milestone 1 and 4 into Milestone 5. (Expected completion, TBA)

More information:
Computational Magma Dynamics Workshop 2006 report