[gradsusr] high resolution file plotting constant value

Thu Mar 14 14:12:31 EDT 2013

Hi Carissa,

    From my experience, you are right.  It's a small stack problem.  The
jpeg2000 compression
uses the Jasper library which requires a large stack when processing large
grids.  You can
change the stack size by ulimit -s XXXXX.  (Only works once per terminal
session.)

   As for the NOMADS implementation, you have two choices

    (1) increase the stack size on the NOMADS machine. Not sure how this
affects the memory footprint.
           (probably OS dependent)
    (2) change the sst from jpeg2000 packing into complex packing
          wgrib2 IN.grb -set_grib_type c2 -grib_out OUT.grb

   The advantage of (2) is that complex packing is much faster than
jpeg2000 to decode.

     Wesley Ebisuzaki

On Thu, Mar 14, 2013 at 12:25 PM, Jennifer Adams <jma at cola.iges.org> wrote:

> Hi, Carissa --
> I can draw your temperature field on my mac laptop, but get the same error
> you do on my linux server ("Constant field. Value= 271.35"). There is a
> non-informative message "jpc_decode return" that comes from the grib2
> library. The 'jpc_decode' routine is part of the jasper library, which is
> required by the grib2 library, but I can't find anything about it in the
> jasper documentation. In the library's code, there are comments that this
> routine is the JPEG-2000 decoder. So the problem likely has something to do
> with the JPEG compression. I have no idea why it would work on AIX and OSX
> but not linux, but I don't think ulimit or the GrADS version has anything
> to do with it (I was testing with GrADS 2.0.2, but the grib2 I/O code
> hasn't changed in a long time). Can you try using wgrib2 or some other tool
> to change the compression and see if that makes a difference? I don't have
> any other bright ideas … perhaps Wesley can be more helpful.
> --Jennifer
>
>
>
> On Mar 14, 2013, at 11:17 AM, Carissa Klemmer - NOAA Federal wrote:
>
> Users,
>
> I work for NCEP on the team that supports the operational NOMADS site. We
> are in the process of adding a new product to GrADS-OpenDAP but are running
> into issues. I have included the syntax for the control file used below to
> make the index file. Once the control and index file are created I attempt
> to view the data plotted through grads.
> Here is the problem. I can not get the data to plot anything but a
> constant value on our sandbox nomads server or on my desktop. I can however
> see the plot successfully on our operational supercomputer. I have
> suspicions of the ulimit stack size, but have not been able to confirm this
> due to the test machines system constraints. So can anyone else get this
> data to successfully plot? If so please share your wisdom. Why does the
> exact same control file and grads version work on one system but not
> another?
>
> The supercomputer is running grads version 2.0.a7 on AIX with ulimit -s of
> 4000000.
> NOMADS sandbox is running 2.0.a3.p3 on LINUX with ulimit -s of 16832
> I have downloaded both 2.0.2 and 2.0.a7 to my desktop (LINUX) with a
> ulimit -s 8000.
>
> 1. The grib2 file: rtgssthr_grb_0.083.grib2
> which can be found for download under:
> http://nomads.ncep.noaa.gov/pub/data/nccf/com/gfs/prod/sst.20130313/
>
> 2. The control file (sst.ctl) was created from g2ctl as follows:
> dset rtgssthr_grb_0.083.grib2
> index sst.idx
> undef 9.999E+20
> title /com/gfs/prod/sst.20130312/rtgssthr_grb_0.083.grib2
> * produced by g2ctl v0.0.8.2
> * command line options: /com/gfs/prod/sst.20130312/rtgssthr_grb_0.083.grib2
> * griddef=1:0:(4320 x 2160):grid_template=0:winds(N/S): lat-lon grid:(4320
> x 2160) units 1e-06 input WE:NS output WE:SN res 48 lat
> 89.9                  58000 to -89.958000 by 0.083000 lon 0.042000 to
> 359.958000 by 0.083000 #points=9331200:winds(N/S)
>
> dtype grib2
> ydef 2160 linear -89.958000 0.083
> xdef 4320 linear 0.042000 0.083000
> tdef 1 linear 00Z12mar2013 1mo
> zdef 1 linear 1 1
> vars 1
> TMPsfc   0,1,0   0,0,0 ** surface Temperature [K]
> ENDVARS
>
> 3. to create the index file I ran:
> >gribmap -0 -v -i sst.ctl
>
> 4.To view the output
> >grads
> >d tmpsfc
>
> As even more proof that I somewhat know what I am doing, I can
> successfully create a control and index file using the same steps as above
> and view pretty graphics for the half degree SST, rtgssthr_grb_0.5.grib2.
>
> Thanks for any insight.
>
> --
> Carissa Klemmer
> NCEP Central Operations
> Production Management Branch Dataflow Team
> 301-683-3835 _______________________________________________
> gradsusr mailing list
> gradsusr at gradsusr.org
> http://gradsusr.org/mailman/listinfo/gradsusr
>
>
> --
> Jennifer M. Adams
> IGES/COLA
> 4041 Powder Mill Road, Suite 302
> Calverton, MD 20705
> jma at cola.iges.org
>
>
>
>
> _______________________________________________
> gradsusr mailing list
> gradsusr at gradsusr.org
> http://gradsusr.org/mailman/listinfo/gradsusr
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://gradsusr.org/pipermail/gradsusr/attachments/20130314/78cf5e24/attachment-0003.html 