2007-02-26 16:06:30 UTC
To everyone,
I've been considering doing a rewrite of DUNS, and would like to solicit people's opinions about what should or should not be done. Currently, I think it would be best to rewrite it as Fortran90 code. Parallelism would be implemented through OpenMP, instead of the domain decomposition done currently, which in my opinion never really worked well anyways. I'd also like to take out the dependency on ChemKin for the chemical properties. We only use a small portion of the chemkin library, and the few functions that we use can be redone much more efficiently for CFD.
A different direction to take would be to do the rewrite completely in C. I have quite a bit of experience with C, so it would turnout fine, but others might prefer remaining Fortran centric.
The main algorithm: diagonalized-implicit, upwind, structured multi-block, with all the currently available thermodynamic models would remain the same.
Doug
