mt-metis orderings for The University of Florida Sparse Matrix Collection

Hello,
I'm trying to use mt-metis 0.4.3 for ordering matrices from The University of Florida Sparse Matrix Collection. I call mtmetis_nd() with the default settings. Running mt-metis using 16 cores, I get orderings with 10-30x non-zeros more than 1-core results.
For example, for matrix parabolic_fem (http://www.cise.ufl.edu/research/sparse/matrices/Wissgott/parabolic_fem.html) mt-metis gives 26,530,942 non-zeros using 1 core and 564,678,524 non-zeros using 16 cores.
What should I do to improve the quality of orderings?

Thanks in advance

RE: My apologies for this. It

My apologies for this. It looks like I had introduced a bug into the API call at somepoint. This should be fixed in version 0.4.4 which was just posted today. Please let us know if there are any other issues.

Thanks,

Dominique