[GRASS-user] unstable results with r.param.scale ?

[Anna Petrášová]

Don't have time right now, but running it with valgrind might show the
problem.

Anna

On Wed, Dec 9, 2020 at 8:27 AM Vincent Bain <bain at toraval.fr> wrote:

> Hello Vincent,
>
> in case it could be related to my recent issue with r.param.scale:
>
> https://lists.osgeo.org/pipermail/grass-user/2020-October/081788.html
>
> In my case there was most probably something wrong with my OpenMP
> configuration.
>
> HTH,
> Vincent.
>
> Le mercredi 09 décembre 2020 à 13:59 +0100, Vincent Godard a écrit :
> > Dear Grass users,
> >
> > I've been using r.param.scale to compute various type of curvatures
> > over
> > a 1 m LiDAR DEM (FCELL) (from opentopography.org) :
> >
> > r.param.scale --o -c input=dem output=curvature size=15 method=profc
> > r.info -r curvature
> >
> > I usually get reasonable results like (checked with independent
> > calculations) :
> >
> > min=-0.203124601340989
> > max=0.0685868278408049
> >
> > But when running this command repeatedly, I randomly (usually every
> > 3rd
> > or 4th time) get very different results like :
> >
> > min=-127112309.998249
> > max=131128979.492382
> >
> > I am running GRASS 7.8.4 on Ubuntu 20.04. The problem is also
> > present
> > when running r.param.scale from QGIS or rgrass7 (in a R session).
> >
> > I have replicated the problem with another high resolution DEM of
> > different source (still FCELL), and also with a 30 m SRTM DEM
> > (CELL),
> > but in this case the occurrence of these odd results is less
> > frequent
> > and they are not contrasting as much with the expected outcome
> > ("only"
> > by a 1 or 2 order of magnitude). The problem is present with or
> > without
> > the -c option, and seems only to affect the computation of
> > curvatures
> > (profc,planc,longc,crosc,minic,maxic).
> >
> > Has anyone encountered something like that, or have any idea about
> > what
> > might be going on?
> >
> > Thanks in advance,
> >
> > Vincent
> >
>
> _______________________________________________
> grass-user mailing list
> grass-user at lists.osgeo.org
> https://lists.osgeo.org/mailman/listinfo/grass-user
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/grass-user/attachments/20201209/55620ffa/attachment-0001.html>