[Bug 715314] Review Request: metis - A set of serial programs for partitioning graphs, partitioning finite element meshes, and producing fill reducing orderings for sparse matrices

[bugzilla at redhat.com]

Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug.


https://bugzilla.redhat.com/show_bug.cgi?id=715314

Ankur Sinha <sanjay.ankur at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |CLOSED
         Resolution|                            |CANTFIX
        Last Closed|                            |2011-07-01 11:48:25

--- Comment #4 from Ankur Sinha <sanjay.ankur at gmail.com> 2011-07-01 11:48:25 EDT ---
(In reply to comment #3)
> It still doesn't look like the licensing is Free here, but I'll wait to see
> what clarification comes from upstream.

Here's the clarification(quoting):

"
Hi Ankur

Metis is used as the main strategy for mesh partitioning i.e. its relevant only
for parallel computations. We have a historical OK from Metis developers to
distribute it with Elmer. However, I understand that this does not make Metis
free software. There is an alternative package, Scotch [1], which to my
understanding provides similar functionality. It is used in the Debian build of
Elmer. We should try to invest some time to test these packages and use perhaps
it instead of Metis. As a quick solution I think that building ElmerGrid
without Metis is OK. Just change the #define PARTMETIS in femelmer.h to zero
and there is some alternative text printed in case Metis is called. 

Best regards,
Peter

[1] http://www.labri.fr/perso/pelegrin/scotch/
"

I'm going to go look at scotch. This one's a CANTFIX :/

Thanks all for the inputs :)

Regards,
Ankur

-- 
Configure bugmail: https://bugzilla.redhat.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.