[ncl-talk] gsn_contour_shade is SLOW!

Anne anne.seidenglanz at unive.it
Thu Feb 14 10:35:25 MST 2019


Hi,

I would like to pick up an issue on scripts running slow that contain the
gsn_contour_shade function (see conversation below from 2011).
My question is: has there been found any solution to this problem?? I am
trying to plot a panel plot with 6 plots which each of them has the
significance stippling  overlaid, but really takes forever..:/

I used

      opt at gsnShadeHigh = 4

and

         plot(0) = gsn_csm_contour_map(wks, var_diff_ensmean(0,:,:), res)
         plot0   = gsn_csm_contour(wks, prob_nov, res2)
         plot0   = gsn_contour_shade(plot0, 10e-7, 1.96, opt)
        overlay(plot(0),plot0)


Any ideas?:)
Thanks
Anne



Hi Steve,

I would like to know why it takes so long too! I see you are doing raster
contours, which would have been the first thing I'd ask you to try (the
default "area fill" filled contours can be slow).

Can you poing me to the script and data on mirage? Feel free to send me a
personal email on this.

Thanks,

--Mary

On Apr 11, 2011, at 11:37 AM, Stephen Yeager wrote:

*> *
*> Hello: *
*> *
*> I've attached an ncl script which can be run on mirage. It generates *
*> a 16-panel plot. Once the computation is done (a minute, perhaps), it
takes about 10-20 seconds to generate the final plot. *
*> *
*> However, when I add shading onto the contour plots using either *
*> ShadeLtContour() or gsn_contour_shade() *
*> *
*> it takes more than 30 minutes to finish! I like the resulting page of
plots, but would like to know why it takes so long to complete, and how I
might speed it up. *
*> *
*> Thanks, *
*> Steve *
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.ucar.edu/pipermail/ncl-talk/attachments/20190214/37a60fb4/attachment.html>


More information about the ncl-talk mailing list