Subject: Re: [ublas] MUMPS and latest traits
From: Thomas Klimpel (Thomas.Klimpel_at_[hidden])
Date: 2009-02-19 06:24:14
Jesse Perla wrote:
> While I am using the new LAPACK bindings, I was also hoping to get
> MUMPS up and running. Is it likely that WinMumps + the old mumps
> bindings + boost 1.38 are going to be compatible with the current
> matrix/vector traits? Or should I probably stay away for now
Is there such a thing as WinMumps? I thought I would have to use the cygwin version of gfortran to compile mumps and then link against the resulting library with VC9 (more or less in the same way as described in the ATLAS documentation).
Karl Meerbergen wrote:
> I have no experience with bindings + mumps on windows. Someone else?
It's on my to do list, as a part of what I call "backlog". Give me time till next Monday, and I will report whether the regression tests passes.
To be honest, this was quite low on my to do list, but I have not yet hit any obstacles. (As I answered Mehran Ziadloo: "I don't even have a g95 on my windows. I only have a g77 (gcc 3.2) and a gfortran-4 (gcc 4.3.2). And I'm currently not planning on experimenting to link their output with VC++ projects.")
I currently omit the link step in the regression tests on VC9 for libraries that must be compiled with cygwin.
My initial plan was to first get everything work under cygwin before trying to cross link between VC9 and cygwin. The ATLAS documentation gives the impression that it should work, but Clint Whaley (the maintainer of ATLAS) seems to have entered some bugs against gcc for alignment issues that break the interoperability with VC for ATLAS (sorry for creating FUD). So I'm quite curious whether the regression tests will pass for the cygwin/VC9 combination.