Inti Pelupessy
Inti Pelupessy
this is related to #23
The alternative to this is being able to restart using the bmi getters and setters (which offers more flexibility). The rationale of seperate function for writing and reading the state...
Currently this can be emulated with e.g. a 1D grid where you ignore the "gridiness." More elegant would be to allow for non-grid data stores.
I think for boundaries: these can be implemented as seperate grids, ie one or more boundaries are refered to with different grid indices. These can be (depending on the implementation...
Is it documented somewhere how to handle/ interface with MPI parallel codes?
the problem with the pickle seems that a pickled numpy array is quite big compared to raw binary (the unit overhead is small)
on the other hand, you can use pickle with binary protocol (1 or 2): ```S = pickle.dumps(y,1)``` this is quite compact: 8726 bytes for 1000 double array+units.kms
is the solution ok (pickle with binary protocol)? so we can close..
yea..I think it has stuff in there that is used in other places though..