ODE Solver for transient and nonlinear processes – core functionality
Created by: chleh
This is a splitoff from #1040, i.e., it contains the parts that have been reviewed yesterday by @wenqing and @endJunction. The comments from the review are in d6131c9, all new modifications come after them. I hope I did not forget to address any of them.
One comment I did not address is that about the ordering of BLAS function arguments. Currently that follows the ordering of the respective PETSc functions. I propose to keep it like that for the time being. Maybe we can change it later on.
As with #1040 the following is the state:
- LisMatrix/LisVector is not supported
- The PETSc build compiles, but many methods are only stubs. @wenqing, maybe we can work on that later on.
- The unit test only makes sure that things compile. Currently it only produces result files that are not compared to the analytical solutions. A strict test has to be implemented later on.