Thread Subject: mesh generation

Hello,
does someone know why meshing on different machines and at different clock time results in different mesh?
I'm using initmesh, refinemesh ...
Thanks to all for hints

"Romeo " <romeo_aristogatto@yahoo.com> wrote in message <grj0ge$eke$1@fred.mathworks.com>...
> Hello,
> does someone know why meshing on different machines and at different clock time results in different mesh?
> I'm using initmesh, refinemesh ...
> Thanks to all for hints

There must be some random component in the
meshing algorithm.

John

"another nice mesh you've gotten me into..."

"John D'Errico" <woodchips@rochester.rr.com> wrote in message <grj2ce$2vr$1@fred.mathworks.com>...
> "Romeo " <romeo_aristogatto@yahoo.com> wrote in message <grj0ge$eke$1@fred.mathworks.com>...
> > Hello,
> > does someone know why meshing on different machines and at different clock time results in different mesh?
> > I'm using initmesh, refinemesh ...
> > Thanks to all for hints
>
> There must be some random component in the
> meshing algorithm.
>
> John
>
> "another nice mesh you've gotten me into..."

Hi John,
that's what I thought too but if i run the same code on the same machine at different clock times I get the same result, and all random generators in matlab automatically change the seeds from clock tables ....
But if I change machine, i.e. my laptop or parallel cluster machines, I get different mesh with the same input arguments ...
Whatever I didn't find in m-files any random commmands, because dealunay algorithms is fully deterministic, but may be I missed something ...
Any further idea will be appreciate, thanks

"Romeo " <romeo_aristogatto@yahoo.com> wrote in message <grj5a6$ouq$1@fred.mathworks.com>...
> "John D'Errico" <woodchips@rochester.rr.com> wrote in message <grj2ce$2vr$1@fred.mathworks.com>...
> > "Romeo " <romeo_aristogatto@yahoo.com> wrote in message <grj0ge$eke$1@fred.mathworks.com>...
> > > Hello,
> > > does someone know why meshing on different machines and at different clock time results in different mesh?
> > > I'm using initmesh, refinemesh ...
> > > Thanks to all for hints
> >
> > There must be some random component in the
> > meshing algorithm.
> >
> > John
> >
> > "another nice mesh you've gotten me into..."
>
> Hi John,
> that's what I thought too but if i run the same code on the same machine at different clock times I get the same result, and all random generators in matlab automatically change the seeds from clock tables ....

If you restart matlab, it always starts with the same
random seed though, unless you reset the seed in
your startup.m file.

> But if I change machine, i.e. my laptop or parallel cluster machines, I get different mesh with the same input arguments ...
> Whatever I didn't find in m-files any random commmands, because dealunay algorithms is fully deterministic, but may be I missed something ...

Actually, sometimes the delaunay tools can do a
random joggle, and I don't know if the joggle is
done using Matlab's random generator or not.

Next, I've seen MANY times when someone has
moved some data from one system to another.
They always say they have changed nothing, yet
they get different results on different machines.
In EVERY case, we have eventually shown how
they actually used different data.

For example, are you pasting in 4 digits in these
numbers into a different matlab session? This is
a common error to make.

John

Thanks a lot John, you gave me the typical hints that show the right way when there's only confusion =)