Re: [hermes3d] Refinement towards origin

[moritz braun]

Dear Pavel

I have investigated at the eigenvector  obtained for the
simple system of the harmonic oscillator in three dimensions
with the code as in the tests subdirectory of
/hermes/hermes3d/examples/schroedinger-eigen-refine-towards-origin/
as submitted some time ago.

For uniform refinement, there is convergence to the analytical solution
1/pi^{3/4} exp(-x^2/2) at the points  on the x-axis for which the wave function values are dumped via the out_fn_vtk function to a file 

However for uniform refinement the wave function comes out assymetric  on the x-axis, i.e. the refinement breaks
the proper symmetry of the system.
This is quite startling!!

Any ideas what could be wrong?

regards

Moritz


--
Prof M Braun         Tel.:27-12-4298006/8027
Physics Department  Fax.: 27-12-4293643
University of South Africa (UNISA)     
moritz...@gmail.com
P.O. Box 392    
0003
UNISA
 South Africa

[Pavel Solin]

Hi Moritz,

On Sun, Mar 20, 2011 at 9:16 AM, moritz braun <moritz...@gmail.com> wrote:
> Dear Pavel
>
> I have investigated at the eigenvector  obtained for the
> simple system of the harmonic oscillator in three dimensions
> with the code as in the tests subdirectory of
> /hermes/hermes3d/examples/schroedinger-eigen-refine-towards-origin/
> as submitted some time ago.
>
> For uniform refinement, there is convergence to the analytical solution
> 1/pi^{3/4} exp(-x^2/2) at the points  on the x-axis for which the wave
> function values are dumped via the out_fn_vtk function to a file
>
> However for uniform refinement the wave function comes out assymetric  on
> the x-axis, i.e. the refinement breaks
> the proper symmetry of the system.
> This is quite startling!!
>
> Any ideas what could be wrong?

We are now finalizing a transition to new object-oriented weak forms,
so I do not have time for anything at the moment. I just checked that
the example compiles in the old master but not in the weakform branch.

Can you send some screenshots to illustrate the problem?

Pavel

> regards
>
> Moritz
>
>
> --
> Prof M Braun         Tel.:27-12-4298006/8027
> Physics Department  Fax.: 27-12-4293643
> University of South Africa (UNISA)
> moritz...@gmail.com
> P.O. Box 392
> 0003
> UNISA
>  South Africa
>
>

> --
> hermes3d mailing list
> herm...@googlegroups.com
> http://groups.google.com/group/hermes3d?hl=en

--
Pavel Solin
University of Nevada, Reno
Home page: http://hpfem.org/~pavel
FEMTEC 2011: http://hpfem.org/events/femtec-2011/
Hermes: http://hpfem.org/
FEMhub: http://femhub.org/

[moritz braun]

On Sun, Mar 20, 2011 at 9:36 PM, Pavel Solin <so...@unr.edu> wrote:

Hi Moritz,

On Sun, Mar 20, 2011 at 9:16 AM, moritz braun <moritz...@gmail.com> wrote:
> Dear Pavel
>
> I have investigated at the eigenvector  obtained for the
> simple system of the harmonic oscillator in three dimensions
> with the code as in the tests subdirectory of
> /hermes/hermes3d/examples/schroedinger-eigen-refine-towards-origin/
> as submitted some time ago.
>
> For uniform refinement, there is convergence to the analytical solution
> 1/pi^{3/4} exp(-x^2/2) at the points  on the x-axis for which the wave
> function values are dumped via the out_fn_vtk function to a file
>
> However for uniform refinement the wave function comes out assymetric  on
> the x-axis, i.e. the refinement breaks
> the proper symmetry of the system.
> This is quite startling!!
>
> Any ideas what could be wrong?

We are now finalizing a transition to new object-oriented weak forms,
so I do not have time for anything at the moment. I just checked that
the example compiles in the old master but not in the weakform branch.

Can you send some screenshots to illustrate the problem?

Dear Pavel

I have plotted the eigen functions obtained at the points along the x-axis dumped in the vtk output files
both for the case
of two homogenous refinements of the initial grid as given box.mesh3d , command line parameter 2
and for  three refinements towards the origin,  command line parameter 1, the plots are attached.
I have also looked at the second case along the y and z axis and it has the same assymetry.
The green function is the analytical expression for the ground state along the x axis.
I hope these plots make my previous e-mail a bit clearer.

regards

Moritz 

[moritz braun]

Dear Pavel

I need a bit of help on how to
move to the new branch in git, since
as I understand it
one can only have one fork at a given time
Since I do not want do overwrite my stuff that i kept syncronizing with master
I am not sure how to go about it!
I guess I need
a) rename my local repo
and
b) start a new branch in my private fork and
work on it in a different workspace on my harddrive.
Thus it will not affect the branch I worked on
or do you have a better suggestion?

Thanks for your assistance!

regards

Moritz 

On Mon, Mar 21, 2011 at 7:16 PM, Pavel Solin <so...@unr.edu> wrote:

On Mon, Mar 21, 2011 at 9:54 AM, moritz braun <moritz...@gmail.com> wrote:
>
>
> On Mon, Mar 21, 2011 at 6:52 PM, Pavel Solin <so...@unr.edu> wrote:
>>
>> Moritz,
>>  I have one more suggestion. Can you prepare a 2D example
>> that does exactly the same thing, to see whether this behavior
>> would be there as well? In 2D it would be possible for me to get
>> involved. If it worked, I would take the time to investigate the
>> 3D problem.
>>
> Dear Pavel
>
> a two dimensional example would may be a good way
> forward since I could then compare the "hand made" refinement towards a
> point
> with the built-in one
>
> I will come back to you in a few days hopefully on this matter.

OK. A 2D example is easier to begin with, and I will be
able to learn it in much shorter time. It would be great if you
could do it in the new weakform branch here

g...@github.com:karban/hermes.git

This is the new master. Forms are done differently (in 2D)
You will be able to learn this in tutorial examples or
benchmarks.

Pavel

> regards
>
> Moritz
>>
>> Pavel

>>
>> On Mon, Mar 21, 2011 at 9:50 AM, Pavel Solin <so...@unr.edu> wrote:
>> > Hi Moritz,
>> >

>> > On Mon, Mar 21, 2011 at 6:57 AM, moritz braun <moritz...@gmail.com>

>> > Thanks for the images. Due to other commitments I can't promise
>> > though that I can take enough time to dive into your problem and
>> > solve it.
>> >
>> > BTW, Ondrej run away from my group. Sorry I bothered you with
>> > a request to help him previously. This is not needed anymore.
>> >
>> > Pavel

[Erick]

Hi Prof. Braun,

I have done something similar to this. If its losing your stuff
you're worried about, why don't you just "cd" to your home directory
and while there create a new directory (say "new_repo" or whatever)
and re-clone the hermes repository there. After re-cloning the hermes
repository do "git co -b [branch name goes here]", and you will have a
new fresh working environment to work in and you wont have to worry
about losing your "other" data; after this you could also "pull" from
whatever branch you'd like to have the latest work of that branch.
Also create a new repository at your remote github "https://github.com/
[your address is here]" so you will have a clean remote repo to push
your new stuff/work to, and you could also use to fork to a new
branch. I hope this helps a little bit.

Erick

On Mar 25, 10:08 am, moritz braun <moritz.br...@gmail.com> wrote:
> Dear Pavel
>
> I need a bit of help on how to
> move to the new branch in git, since
> as I understand it
> one can only have one fork at a given time
> Since I do not want do overwrite my stuff that i kept syncronizing with
> master
> I am not sure how to go about it!
> I guess I need
> a) rename my local repo
> and
> b) start a new branch in my private fork and
> work on it in a different workspace on my harddrive.
> Thus it will not affect the branch I worked on
> or do you have a better suggestion?
>
> Thanks for your assistance!
>
> regards
>
> Moritz
>
>
>
> On Mon, Mar 21, 2011 at 7:16 PM, Pavel Solin <so...@unr.edu> wrote:

> > On Mon, Mar 21, 2011 at 9:54 AM, moritz braun <moritz.br...@gmail.com>

> > >> > On Mon, Mar 21, 2011 at 6:57 AM, moritz braun <moritz.br...@gmail.com

>
> > >> > wrote:
>
> > >> >> On Sun, Mar 20, 2011 at 9:36 PM, Pavel Solin <so...@unr.edu> wrote:
>
> > >> >>> Hi Moritz,
>
> > >> >>> On Sun, Mar 20, 2011 at 9:16 AM, moritz braun <

> > moritz.br...@gmail.com>

> > >> >>> > moritz.br...@gmail.com

> > >> >>> > P.O. Box 392
> > >> >>> > 0003
> > >> >>> > UNISA
> > >> >>> >  South Africa
>
> > >> >>> > --
> > >> >>> > hermes3d mailing list
> > >> >>> > herm...@googlegroups.com
> > >> >>> >http://groups.google.com/group/hermes3d?hl=en
>
> > >> >>> --
> > >> >>> Pavel Solin
> > >> >>> University of Nevada, Reno
> > >> >>> Home page:http://hpfem.org/~pavel
> > >> >>> FEMTEC 2011:http://hpfem.org/events/femtec-2011/
> > >> >>> Hermes:http://hpfem.org/
> > >> >>> FEMhub:http://femhub.org/
>
> > >> >>> --
> > >> >>> hermes3d mailing list
> > >> >>> herm...@googlegroups.com
> > >> >>>http://groups.google.com/group/hermes3d?hl=en
>
> > >> >> --
> > >> >> Prof M Braun         Tel.:27-12-4298006/8027
> > >> >> Physics Department  Fax.: 27-12-4293643
> > >> >> University of South Africa (UNISA)

> > >> >> moritz.br...@gmail.com

> > > moritz.br...@gmail.com

> > > P.O. Box 392
> > > 0003
> > > UNISA
> > >  South Africa
>
> > --
> > Pavel Solin
> > University of Nevada, Reno
> > Home page:http://hpfem.org/~pavel
> > FEMTEC 2011:http://hpfem.org/events/femtec-2011/
> > Hermes:http://hpfem.org/
> > FEMhub:http://femhub.org/
>
> --
> Prof M Braun         Tel.:27-12-4298006/8027
> Physics Department  Fax.: 27-12-4293643
> University of South Africa (UNISA)

> moritz.br...@gmail.com

[moritz braun]

Hi Erick

Thanks for your hints

what I did in the end
is simply to create a new branch in place
and pull down the "new" master by Pavel Korban

regards

Moritz

moritz...@gmail.com

[Erick]

Nice, that works too.

Erick

[Pavel Solin]

Hi Moritz, how do you like the new forms?
Pavel

[moritz braun]

Dear Pavel

Have a good rest

regards

Moritz

On Sat, Mar 26, 2011 at 10:01 AM, Pavel Solin <so...@unr.edu> wrote:

I tried it and you are right. This must be the result of
a very recent change, I am copying Lukas to see whether
he has a clue (I have 1 am in the morning and I don't at
the moment).

Pavel

On Sat, Mar 26, 2011 at 12:51 AM, moritz braun <moritz...@gmail.com> wrote:
> Dear Pavel
>

> I managed to compile the nist-03 benchmark, but when trying to run it, it
> fails  with the following output
>
> I ---- Adaptivity step 1:
> I Solving on reference mesh.
> Traceback (most recent call last):
>   File unknown, address: 0x404859
>   File unknown, address: 0x7fe533fe2d8e
>   File unknown, address: 0x405e8d
>   File unknown, address: 0x7fe537382aae
>   File unknown, address: 0x7fe53739bd0f
>   File unknown, address: 0x7fe53739b848
>   File unknown, address: 0x7fe53739b1cc
>   File unknown, address: 0x7fe537395f1c
>   File unknown, address: 0x7fe5373952cf
>   File unknown, address: 0x7fe537388051
>   File unknown, address: 0x7fe5373695f3
>   File unknown, address: 0x7fe536e96d49
>   File unknown, address: 0x7fe536ea1132
>   File unknown, address: 0x7fe536ea0cd8
> E WeakForm::MatrixFormVol::ord must be overrided.
>
>
>
>
> On Sat, Mar 26, 2011 at 9:23 AM, Pavel Solin <so...@unr.edu> wrote:
>>
>> On Sat, Mar 26, 2011 at 12:14 AM, moritz braun <moritz...@gmail.com>
>> wrote:
>> > Dear Pavel
>> >
>> > Thanks for the instructions
>> >
>> > However, If I  work with a clone of this branch
>> > then I can't  submit my example for the refinement problem back to you!
>>
>> You can send a pull request to Pavel Karban's repo.
>>
>> Pavel
>>
>> >
>> > regards
>> >
>> > Moritz
>> >
>> > On Sat, Mar 26, 2011 at 9:12 AM, Pavel Solin <so...@unr.edu> wrote:
>> >>
>> >> After you clone hermes from Pavel Karban's weakform branch,
>> >> type cmake . but do not type make. Go for example to
>> >> hermes2d/benchmarks-nist, and type make there. This will build
>> >> just these benchmarks. They have been already updated.
>> >> Analogously, you can build just one of them by descending into
>> >> its directory and typing make there.
>> >>
>> >> On Sat, Mar 26, 2011 at 12:08 AM, moritz braun <moritz...@gmail.com>
>> >> wrote:
>> >> > Hi Pavel
>> >> >
>> >> > In order to check whether hese problems   are because of the way that
>> >> > I
>> >> > have
>> >> > pulled the new branch into my fork
>> >> > I will now make a clean clone of Pavel Karban;s  branch and
>> >> > look at it first.
>> >> >
>> >> > regards
>> >> >
>> >> > Moritz
>> >> >
>> >> > On Sat, Mar 26, 2011 at 8:57 AM, moritz braun
>> >> > <moritz...@gmail.com>
>> >> > wrote:
>> >> >>
>> >> >> Dear Pavel
>> >> >>
>> >> >> I am a bit confused now
>> >> >>
>> >> >> I ran
>> >> >> cmake .
>> >> >> and
>> >> >> make in the hermes top directory
>> >> >> as I normally
>> >> >> do
>> >> >> I don't know how to compile only the tutorials seperately
>> >> >>
>> >> >> regards
>> >> >>
>> >> >> Moritz
>> >> >>
>> >> >>
>> >> >> On Sat, Mar 26, 2011 at 8:51 AM, Pavel Solin <so...@unr.edu> wrote:
>> >> >>>
>> >> >>> On Fri, Mar 25, 2011 at 11:47 PM, moritz braun
>> >> >>> <moritz...@gmail.com>
>> >> >>> wrote:
>> >> >>> > Hi Pavel
>> >> >>> >
>> >> >>> > I am a bit confused as how to go about this now
>> >> >>> > I had to uncomment most of the tutorial directories in the
>> >> >>> > CMakeLists.txt
>> >> >>> > and it got to late for me and I went to bed
>> >> >>> > It is rather tedious to fiddle like this
>> >> >>>
>> >> >>> Why would you comment or uncomment anything
>> >> >>> in any CMakeLists.txt file. Hermes compiles, just
>> >> >>> some examples do not. So do not attempt to
>> >> >>> compile all examples and instead just work in
>> >> >>> the directory with a specific example of your interest.
>> >> >>>
>> >> >>> Pavel
>> >> >>>
>> >> >>> >
>> >> >>> > regards
>> >> >>> >
>> >> >>> > Moritz
>> >> >>> >
>> >> >>> > On Sat, Mar 26, 2011 at 8:42 AM, Pavel Solin <so...@unr.edu>
>> >> >>> > wrote:
>> >> >>> >>
>> >> >>> >> On Fri, Mar 25, 2011 at 11:19 PM, moritz braun
>> >> >>> >> <moritz...@gmail.com>
>> >> >>> >> wrote:
>> >> >>> >> > Hi Pavel
>> >> >>> >> >
>> >> >>> >> > Unfortunately most of the tutorials don't compile
>> >> >>> >> > I am not sure what is wrong!
>> >> >>> >>
>> >> >>> >> Not all tutorials have been converted yet. Try compiling
>> >> >>> >> just benchmarks-nist, they are all done.
>> >> >>> >> Pavel
>> >> >>> >>
>> >> >>> >> >
>> >> >>> >> > regards
>> >> >>> >> >
>> >> >>> >> > Moritz
>> >> >>> >> >
>> >> >>> >> > On Sat, Mar 26, 2011 at 6:18 AM, Pavel Solin <so...@unr.edu>