Hi, thanks for your reply.
All the programs are installed as 32 bit.
I should add that I also installed the python-only version of the program using
Cantera-Python-2.2.0-x86-py2.7.msi
and the startup tests ran fine from the python command line.
I could in principle run with python, I have to admit that I feel more comfortable with octave and thought I'd ask about octave since it is a reasonably popular free alternative to matlab, and if nobody asked before this could be useful to others. I realized that the mex dll, having been compiled to talk with matlab, might not quite work with octave. I should add that this is the "standalone" version of octave, not the one embedded in cygwin. I also have cygwin installed, so in principle I could try any modifications you can suggest for running from cygwin with octave. But I am far from proficient, I can handle some c but c++ is still on my to-learn list.
Since there are so many potential threads to this discussion I might as well add two more :-)
First I am wondering how the mex interface is set up, since python appears to be required still with the matlab installation (as suggested by the use of a windows environment flag for the python path). If mex is used to speak with the c++ routines directly, what is the role of python? This is more out of curiosity.
I want to apply cantera to a set of chemical equilibrium problems of the type n*a + m*b <--> c, and am finding with my scripts using Newton-Raphson (such as octave's standard routine fsolve) don't converge. I think the problem is that some of the concentrations become negligible and the computation become unstable. I've tried a lot of freely available alternative matlab/octave routines and also minimizing the quadratic version of the problem but not much helps. I figured cantera might do the trick. Any comments on this specific application would be welcome.
Finally, I am also wondering how much numerical accuracy improves with the 64 bit versions.
Thanks again,
Christoph