[Nek5000-users] visit visualization question

[nek5000-users at lists.mcs.anl.gov]

Thank you Josh. This is very useful information. I will try it out on my
problem.

So this means visit does not make use of the high order definition of the
solution but simply does linear interpolation ???

Regards
praveen

On Fri, Nov 8, 2013 at 11:59 PM, <nek5000-users at lists.mcs.anl.gov> wrote:

> Praveen,
>
> Another thing you could do is to output your solution onto a uniformly
> spaced grid within each element(as opposed to the standard GLL points).  I
> have found this to be particularly useful in cases where relatively large,
> high order elements are used, like I was seeing for my Euler solver.
>
> To do this, in usrchk, include the following: (I'm doing this out of
> memory, so you may want to double check this)
>
> subroutine userchk()
>
>   include 'SIZE'
>   include 'TOTAL'
>   include 'RESTART'    ! Not 100% sure about this one
>
>   integer mynxo
>
>   mynxo = 10
>   ifreguo = .true.    ! Tells prepost to interpolate to uniformly spaced
> grid in each element
>   nxo = mynxo      ! nxo is the number of points (in each direction) to
> use within the elements
>
> end
>
> Theoretically, nxo can be anything from 2 (which would map your solution
> to linear finite elements) to lxo (which is set in your SIZE file).
>  Although, as I'm writing this, there may be a hard-limit on nxo set
> somewhere in the source code, so maybe the developers could verify this for
> me?
>
> Using this will output your field files to  "_reg" files.  Keep in mind
> that these files will be for output purposes only--restarting from these
> would incur some potentially heavy interpolation errors.
>
> Hope this helps!
>
> Josh
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/nek5000-users/attachments/20131109/86cf5815/attachment.html>