June 2020 to-do list

[jehicken]

• Create a factory for solvers, and have that factory provide the adept::Stack.
• Eliminate the serial version of mach (i.e. get rid of the relevant cpp define guards)
• Merge dev into euler fully
• Consider using res->GetEnergy() for the change in entropy calculation.
• We need a better way of updating the dt in the nonlinear mass integrator (does not generalize right now)
• turn sandbox examples into regression tests before switching to parallel mfem. (See @tuckerbabcock comment below)
• Remove steady option, and instead use a new ODE solver to implement Newton and pseudo-transient continuation.
• The bilinear and nonlinear forms are initialized (i.e. pointers are not nullptr) regardless of whether or not they are used. Can we avoid this?

[jehicken]

For the factory, here is a good reference:
https://isocpp.org/wiki/faq/strange-inheritance#calling-virtuals-from-ctor-idiom

[jehicken]

More notes on the factory:

std::unique_ptr<AbstractSolver> AbstractSolver::create()
{
  std::unique_ptr<AbstractSolver> solver( /*...use a factory to create a Derived object via new...*/ );
  solver->initDerived();
  return solver;
}

[tuckerbabcock]

More TODO:

• turn sandbox examples into regression tests before switching to parallel mfem (I think we can look to MFEM's examples for a template as they can be run as regression tests)

[geyan3566]

TODO:

• Get rid of petsc ?

• Make gd solver parallel and compatible w/o pumi mesh. (This depends on whether the bug is fixed or not in ParPumiMesh).

• In our odl branch, develop the function that assembles interface integrator for ParNonlinearForm::GetGradient().

[tuckerbabcock]

@geyan3566 regarding point 3, don’t make the fix on the odl branch, make it on some other branch and merge that into odl

[jehicken]

@tuckerbabcock : This is your reminder to checkout out the factory branch and get createSolver working with Joule.