[Nek5000-users] History points

Thu May 6 09:01:53 CDT 2010

I will add this routine including some corrections to the latest repo. 
It should be considered as a first step trying to replace the old history point handling.

Stefan

On May 6, 2010, at 3:49 PM, nek5000-users at lists.mcs.anl.gov wrote:

> Hi Guys,
> 
> yes the history point module needs a major overhaul. I don't like the way it works!
> In fact we could do a much better job using our interpolation tool.  
> 
> I just coded something up (without testing the code snipplet) very quicky but I think I should work out of the box. 
> 
> How does it work:
> 
> - create a 'hpts.in' file
> line 1: number of points
> line 2:  X,Y,Z coordinates of a point 1
> ....
> line n+1: X,Y,Z coordinates of point n 
> 
> - set lpart >= n in the SIZE file
> 
> - call hpts() in userchk() 
> 
> The results will be written to 'hpts.out' . 
> 
> Enjoy
> -Stefan
> 
> 
> ----------------------------------------------------------------------------------------------------------------------
> 
>      subroutine hpts
> c
> c     dump velocity, temperature and ps-scalars for a given list
> c     of points (read from file hpts.in) into a file (hpts.out)
> c
>      INCLUDE 'SIZE'
>      INCLUDE 'TOTAL'
> 
>      parameter(nmax=lpart,nfldmax=ldim+ldimt)
>      parameter(mi=4,mr=1+2*ldim+nfldmax)
>      real    rTL(mr,nmax)
>      integer iTL(mi,nmax)
>      common /itlcb/ iTL
>      common /rtlcb/ rTL
> 
>      common /outtmp / wrk(lx1,ly1,lz1,lelt,nfldmax)
> 
>      character*80 fname
>      integer icalld,npoints,icount
>      save icalld,npoints,icount
>      data icalld  /0/
>      data npoints /0/
>      data icount  /0/
> 
>      nxyz  = nx1*ny1*nz1
>      ntot  = nxyz*nelt
> 
>      if(icalld.eq.0) then
>        icalld = 1
>        npoints = 0
> 
>        write(6,*) 'reading hpts.in'
>        open(50,file='hpts.in',status='old')
>        read(50,*) npoints
>        write(6,*) 'found ', npoints, ' points'
>        do i = 1,npoints
>           read(50,*) (rTL(1+j,i),j=1,ndim)
>        enddo
>        close(50)
> 
>        bb_tol = 1e-8  ! bounding-box test tolerance
>        call intpts_setup(bb_tol)
> 
>        open(50,file='hpts.out',status='new')
>        write(50,'(A)') '# time  vx  vy  [vz]  T  PS1   PS2 ...'
>      endif
> 
>      nflds  = nfield + ndim-1 ! number of fields you want to interpolate
> 
>      ! pack working array
>      call copy(wrk(1,1,1,1,1),vx,ntot)
>      call copy(wrk(1,1,1,1,2),vy,ntot)
>      if(if3d) call copy(wrk(1,1,1,1,2),vz,ntot)
>      do i = 1,nfield-1
>         call copy(wrk(1,1,1,1,ndim+i),T(1,1,1,1,i),ntot)
>      enddo
> 
>      ! interpolate
>      call intpts(wrk,nflds,iTL,mi,rTL,mr,npoints,nmax)
> 
>      ! write interpolation results to file
>      if(nid.eq.0) then
>        do ip = 1,npoints
>           write(50,'(1p20E15.7)') time,
>     &      (rTL(1+2*ndim+ifld,ip), ifld=1,nflds)
>        enddo
>      endif
> 
>      return
>      end
> 
> 
> 
> On May 6, 2010, at 3:24 PM, nek5000-users at lists.mcs.anl.gov wrote:
> 
>> P.S.
>> 
>> Yes, Frank, the exact coordinates of the history points are written only when the code rans.
>> 
>> But if for history points you choose the collocation points on the boundary of a spectral element -- those that have I,J,K equal to either 1 or lx1 -- than you can find out their coordinates from .rea ASCII file after the line
>> 
>> *** MESH DATA ***
>> 
>> where coordinates of the element edges are written.
>> 
>> Aleks
>> 
>> 
>> 
>> 
>> 
>> ----- Original Message -----
>> From: nek5000-users at lists.mcs.anl.gov
>> To: nek5000-users at lists.mcs.anl.gov
>> Sent: Wednesday, May 5, 2010 4:45:11 PM GMT -06:00 US/Canada Central
>> Subject: Re: [Nek5000-users] History points
>> 
>> Hi Aleks,
>> 
>> My question is what is the logfile?  Perhaps you mean the output that
>> NEK writes to the screen?  If so, then one would know the location of
>> the point at which data is being dumped only after entering the i,j,k
>> and element # of the point in the *.rea file....  
>> 
>> Cheers,
>> Frank
>> 
>> On Wed, 2010-05-05 at 23:37 +0200, nek5000-users at lists.mcs.anl.gov
>> wrote:
>>> Thanks Aleks.
>>> Concerning grep, it is a linux command that searches for the string
>>> History in the logfile. 
>>> _______________________________________________
>>> Nek5000-users mailing list
>>> Nek5000-users at lists.mcs.anl.gov
>>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users
>> -- 
>> Frank Herbert Muldoon, Ph.D. Mechanical Engineering
>> Technische Universität Wien (Technical University of Vienna)
>> Inst. f. Strömungsmechanik und Wärmeübertragung (Institute of Fluid
>> Mechanics and Heat Transfer)
>> Resselgasse 3
>> 1040 Wien
>> Tel: +4315880132232
>> Fax: +4315880132299 
>> Cell:+436765203470
>> fmuldoo (skype)
>> http://tetra.fluid.tuwien.ac.at/fmuldoo/public_html/webpage/frank-muldoon.html
>> 
>> _______________________________________________
>> Nek5000-users mailing list
>> Nek5000-users at lists.mcs.anl.gov
>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users
>> _______________________________________________
>> Nek5000-users mailing list
>> Nek5000-users at lists.mcs.anl.gov
>> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users
> 
> _______________________________________________
> Nek5000-users mailing list
> Nek5000-users at lists.mcs.anl.gov
> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users