Minutes 14 August 2006

Present: Shane, Karl, Travis, Peter.

Apologies: Poul.

An "audio recording":ftp://ftp.bioeng.auckland.ac.nz/cmiss/opencmiss/meeting-2006-08-14.ogg is available (updated 2006-08-15).

Desired libMesh features

• libMesh developers have wanted a distributed mesh and a test suite for some time and have discussed how these might be implemented. (Sundance appears to have the ability for a distributed mesh.)
• Although many interpolation functions are available (including Hermite), only Lagrange interpolation is available for geometries. LibMesh developers are happy with contributions to extend dependent variable functionality to geometric variables.

Managing our code (raised by Peter)

• Maintaining our changes to libMesh
  • patches if only small changes
  • import libMesh into our own repository if building large components.
• Maintaining our code that sits on top of libMesh
  • can put in examples for now while only a small number of examples.
  • common functions can go into a library later

Plan:

• Transient diffusion: assemble matrix only when necessary - hopefully by next week.
• Make this into an example in cmiss examples - infrastructure for this will take longer than setting up the libMesh code so start simple.
• Add cellml for monodomain reaction-diffusion.
• Think about geometries.

Threads

• PETSc is not threadsafe and does not use threads and its developers don't see threads as the appropriate model for PETSc. This makes it difficult to gain any advantage from threads within any package using PETSc.

Meeting finished 11:36 am.