[ncl-talk] memory limit for gsn_csm_contour_map of large triangulate mesh?

David Brown dbrown at ucar.edu
Tue Oct 6 12:13:48 MDT 2015


Hi Ivica,
What kind of fill are you using? With raster fill we have successfully
rendered grids with 65 million nodes. There is no hard limit for
raster fill. Also you can speed up the rendering using OpenMP, by
setting the environment variable OMP_NUM_THREADS to a some number
greater than 1.
With area fill or cell fill you would be limited to somewhere in the
30 - 40 million node range and area fill in particular would be
painfully slow, but it should still work. The number of nodes you have
should plot pretty easily. If you would like us to take a look, upload
 the data and your script using the instructions at
http://www.ncl.ucar.edu/report_bug.shtml.
 -dave

On Tue, Oct 6, 2015 at 9:53 AM, Ivica Janeković <ivica.jan at gmail.com> wrote:
> Dear support,
> I am trying to contour field from my finite element model and seems I
> hit the limit with number of elements/nodes.
> The grid has 820k nodes and 1,4m elements; when using
> gsn_csm_contour_map it plots just randomly picked triangles. The same
> script is working fine for smaller grids (so script I believe is fine).
> I have the latest ncl version (6.3.0) on debian 64 with a
> lots of memory (32G).
>
> Is there limit on the number of nodes/elems to plot within ncl, and if
> there is how to expand?
>
>
> Thanks in advance,
> Ivica
> _______________________________________________
> ncl-talk mailing list
> ncl-talk at ucar.edu
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk


More information about the ncl-talk mailing list