isbaker007 wrote:
--- In femm@xxxx, David Meeker <dmeeker@xxxx> wrote:I don't know of anyone who has successfully ported femm to run native on Linux. The best bet might be to try and recompile the source using winelib, which could work better than running the unmodified executables in Wine. Someone might have gotten it to work, but has it has been noted (see http://groups.yahoo.com/group/femm/message/999) if someone has been successful in performing a port, they might keep quiet about it because of licensing issues related to Microsoft...
> Hello David,I saw your reply on the questions of a Linux version of software and
the effort to create a 3D version.I am interested in porting femm to run on native Linux (as opposed to
WINE). I'm also interested in developing a 3D version. Its clear from
your comments that there's a fair bit of effort required for each of
these. I was wondering if anyone has either of these projects off the
ground? If so, I'd like to get involved in them.
More on 3D stuff in a later message.
The femm solver might be amenable to being modified for some sort of cluser computing. The solver is really a command-line application that is wrapped up as a Windows dialog application, and it would be easy enough to unwrap it to get a command-line C++ program that would run on just about any platform. In fact, in an older version of the program, the solver was a command line program (I changed it because I couldn't get a call to a command line program to behave the same on all installations of windows, leading to a lot of problems with things working fine on my machine, but not on some other machine running a different version of Windows.) I don't really know much about cluster computing, however, so it would be hard for me to say how it would have to be modified to be amenable to running on a cluster.
I note your point about the compute power required to perform the 3D
calcs. It so happens that I also have an interest in cluster
computing and I am doing initial test runs on a modest cluster of my
own. I'm hoping that it will provide the necessary horse power to do
the job. If no one else is working in this area, I'd like to have a
go. If I can get it all going on the cluster, I'll be able to make it
available to the on-line community later in the year (this is when I
hope to connect the cluster to the internet via web site front-end.)So can you point me in the direction of someone working in the area
or alternatively, is there any support for further development in
these direction?Kind Regards, Ian Baker
Dave.
--
David Meeker
http://femm.berlios.de/dmeeker