[FLASH-USERS] Tracking one cell value in time

[Sai Siva]

Dear Klaus,

Thank you for your suggestions.

On Tue, Jun 18, 2019, 10:01 PM Klaus Weide <klaus at flash.uchicago.edu> wrote:

> On Tue, 18 Jun 2019, Sai Siva wrote:
>
> > I want to track one particular cell value at all the times during the
> > simulation.
> >
> > I could able to get it from *ViSit--> Query options. *However this gives
> a
> > desired cell value at *output times* only. As a result values at other
> > times are unknown.
> >
> > Is there any way in the FLASH code so that I can get one cell value at
> all
> > the times.
>
> There are some routines included with FLASH that should help you do this.
> Have a look at
>
>    Driver_computeCellLocations.F90
>
> and, especially if you just want to write out certain values after each
> time step,
>
>   IO_writeSingleCells.F90 .
>
> The existing implementations are examples, they have some hardcoded values
> (positions) in them and were originally written for 1 1D case. So you will
> have to make some changes.
>
> Note that Driver_computeCellLocations needs to be called collectively by
> all MPI ranks.
>
> This may look overly complicated. The advantage is that you get a
> mechanism for tracking the same physical location across multiple changes
> of the grid during refinement and movement of blocks between ranks.
>
> Klaus
>

On Tue, Jun 18, 2019, 10:01 PM Klaus Weide <klaus at flash.uchicago.edu> wrote:

> On Tue, 18 Jun 2019, Sai Siva wrote:
>
> > I want to track one particular cell value at all the times during the
> > simulation.
> >
> > I could able to get it from *ViSit--> Query options. *However this gives
> a
> > desired cell value at *output times* only. As a result values at other
> > times are unknown.
> >
> > Is there any way in the FLASH code so that I can get one cell value at
> all
> > the times.
>
> There are some routines included with FLASH that should help you do this.
> Have a look at
>
>    Driver_computeCellLocations.F90
>
> and, especially if you just want to write out certain values after each
> time step,
>
>   IO_writeSingleCells.F90 .
>
> The existing implementations are examples, they have some hardcoded values
> (positions) in them and were originally written for 1 1D case. So you will
> have to make some changes.
>
> Note that Driver_computeCellLocations needs to be called collectively by
> all MPI ranks.
>
> This may look overly complicated. The advantage is that you get a
> mechanism for tracking the same physical location across multiple changes
> of the grid during refinement and movement of blocks between ranks.
>
> Klaus
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://flash.rochester.edu/pipermail/flash-users/attachments/20190618/bf6c600e/attachment.htm>