|
Boost Users : |
Subject: Re: [Boost-users] [BGL] visualizing large graphs
From: Cedric Laczny (cedric.laczny_at_[hidden])
Date: 2010-10-01 04:30:00
Hi,
after having had some trouble with the installation, I could finally test
Tulip. My graph is fairly small at the moment, with about 10k nodes and 20k
edges. I'll have to test if it would work in cytoscape, but I doubt it. In
that sense, I really like Tulip because it behaves nicely and I can imagine
that it will do this also for larger graphs. Although, I don't know how large
they really can get.
What kind of disappointed me are the layout algorithms in Tulip. Even though
there might be some plugin-layouts which I haven't tested, this is really
nicely working in cytoscape. Installing e.g. the HistogramView-plugin failed
due to not understandable reasons.
In total, I would say that Tulip is not really helpful in my case, sadly.
Therefore, I hope that some of you may have other solutions or suggestions to
the problem.
Best,
Cedric
On Thursday, 30. September 2010 22:52:39 Marsh Ray wrote:
> On 09/29/2010 04:36 AM, Adam Spargo wrote:
> > Hi Cedric,
> > I have been trying to use a program called Tulip to visualise my graphs,
> > it claims to work up to millions of nodes, but as yet I still get
> > crashes for relatively small graphs.
>
> That was my experience with Tulip as well.
>
> > My main problem has been that all
> > my machines with enough RAM are remote and the x-connection is too slow.
> > I have put some more RAM in my desktop but now the graphics card seems
> > to struggle. Waiting for a new graphics card ... Tulip does support
> > batch mode, so I could run on a big memory machine and output hardcopy,
> > but I haven't had a chance to really learn it interactively yet.
>
> I've been playing a little with OpenGL and large graphs. Without making
> a huge amount of code, a simple massless electrostatic model can make
> cool looking layouts and animations.
>
> Having that third dimensions makes the layout problem much much easier.
> It would be interesting to see what you get if you simply squash the
> result back down to two.
>
> Anyone on this list could probably do something in C++ that gets the job
> done without crashing faster than it took me to finally give up after
> trying all those student-project-like viewers written in Java.
>
> > I'll let you know if I have any success.
>
> Yes, please keep us posted.
>
> - Marsh
> _______________________________________________
> Boost-users mailing list
> Boost-users_at_[hidden]
> http://lists.boost.org/mailman/listinfo.cgi/boost-users
Boost-users list run by williamkempf at hotmail.com, kalb at libertysoft.com, bjorn.karlsson at readsoft.com, gregod at cs.rpi.edu, wekempf at cox.net