[ncl-talk] ESMF regridding error? "Variable (opt) is undefined in line 1613 of ESMF-regridding.ncl
Mary Haley
haley at ucar.edu
Thu Feb 22 17:44:35 MST 2018
Hi Peter,
I'm sorry I didn't respond sooner. I had started looking into this right
away, and then I didn't get back to it until just now.
What I tried to do was trim down the curvilinear_to_SCRIP to its bare
minimum required steps so that maybe the weird errors go away.
Given the size of the grid, I've added
:
opt at NetCDFType = "netcdf4"
Can you run your script with the new attached version of curv_to_scrip.ncl
and see how far it gets before it bombs? I've inserted a lot of print
statements. If it still bombs, please send me all the output.
Meanwhile, I'm running the script now, but it's going to take awhile and I
have to run some errands. Here's the size of the source and destination
files that get created, and you'll see the source one is over 3 GB:
-rw-rw-r-- 1 haley CIT\Domain Users 117M Feb 22 17:34
destination_grid_file.nc
-rw-rw-r-- 1 haley CIT\Domain Users 3.5G Feb 22 17:34
source_grid_file.nc
--Mary
On Thu, Feb 15, 2018 at 2:21 AM, Kuipers Munneke, P. (Peter) <
P.KuipersMunneke at uu.nl> wrote:
> Hi,
>
> Today I have tried the solution below. Unfortunately, the customized
> script didn’t provide a solution. The behavior in the customized script is
> totally erratic and variables get undefined within the script. It’s not
> just the cts_opt variable, but when I comment some lines that try to read
> the cts_opt variable, it starts to complain that other variables are
> unknown, e.g. lat2d, or FTitle. I have tried to generate local variables
> and write the variables into these but they suffer from the same problem. I
> suspect that it is some sort of memory issue where bits and bytes are
> altered or overwritten as soon as a variable is read.
>
> For example, I inserted a printVarSummary statement at many places in the
> file attached. All the way down this message is the result. The procedure
> crashes between lines 42 and 49 in the attached file. Which seems
> completely random. If I comment lines 44-47, the procedure crashes at line
> 60. If I comment lines 51-58, I get:
>
> fatal:Variable (FTitle) is undefined
> fatal:["Execute.c":8640]:Execute: Error occurred at or near line 99 in
> file ./curv_to_scrip.ncl
>
> So all of a sudden it’s no longer cts_opt that is missing, but another
> variable.
>
> I suppose there is not really an easy solution to this problem. As said, I
> think it is memory-related.
>
> Best,
> Peter
>
>
>
> ==================
>
> Copyright (C) 1995-2017 - All Rights Reserved
> University Corporation for Atmospheric Research
> NCAR Command Language Version 6.4.0
> The use of this software is governed by a License Agreement.
> See http://www.ncl.ucar.edu/ for more details.
> (0) test
>
>
> Variable: opt
> Type: logical
> Total Size: 4 bytes
> 1 values
> Number of Dimensions: 1
> Dimensions and sizes: [1]
> Coordinates:
> Number Of Attributes: 12
> NetCDFType : netcdf4
> Debug : False
> PrintTimings : False
> ForceOverwrite : True
> InterpMethod : bilinear
> DstRegional : True
> SrcRegional : True
> DstGridLon : <ARRAY of 1464100 elements>
> DstGridLat : <ARRAY of 1464100 elements>
> SrcGridLon : <ARRAY of 44448889 elements>
> SrcGridLat : <ARRAY of 44448889 elements>
> WgtFileName : ./Bedmap_on_MODIS_grid.nc
> (0) True
>
> Variable: cts_opt (parameter)
> Type: logical
> Total Size: 4 bytes
> 1 values
> Number of Dimensions: 1
> Dimensions and sizes: [1]
> Coordinates:
> Number Of Attributes: 4
> Debug : False
> ForceOverwrite : True
> PrintTimings : False
> NetCDFType : netcdf4
>
> Variable: cts_opt (parameter)
> Type: logical
> Total Size: 4 bytes
> 1 values
> Number of Dimensions: 1
> Dimensions and sizes: [1]
> Coordinates:
> Number Of Attributes: 4
> Debug : False
> ForceOverwrite : True
> PrintTimings : False
> NetCDFType : netcdf4
>
> Variable: cts_opt (parameter)
> Type: logical
> Total Size: 4 bytes
> 1 values
> Number of Dimensions: 1
> Dimensions and sizes: [1]
> Coordinates:
> Number Of Attributes: 4
> Debug : False
> ForceOverwrite : True
> PrintTimings : False
> NetCDFType : netcdf4
> fatal:Variable (cts_opt) is undefined
> fatal:["Execute.c":8640]:Execute: Error occurred at or near line 49 in
> file ./curv_to_scrip.ncl
>
> fatal:["Execute.c":8640]:Execute: Error occurred at or near line 3912 in
> file $NCARG_ROOT/lib/ncarg/nclscripts/esmf/ESMF_regridding.ncl
>
> fatal:["Execute.c":8640]:Execute: Error occurred at or near line 4042 in
> file $NCARG_ROOT/lib/ncarg/nclscripts/esmf/ESMF_regridding.ncl
>
> fatal:["Execute.c":8640]:Execute: Error occurred at or near line 73 in
> file make_topogrid.ncl
>
>
>
>
>
> > On Feb 14, 2018, at 0:22, Mary Haley <haley at ucar.edu> wrote:
> >
> > Hi Peter,
> >
> > Thanks for providing good details about the problem and the files.
> >
> > I'm running your script right now with 6.4.0, but haven't seen the
> problem yet. The script is taking awhile to run, so I'll post back here
> when it's done. So far, it has printed this output:
> >
> > (0) get_src_grid_info: source lat dims = (6667,6667)
> > (0) get_src_grid_info: source lon dims = (6667,6667)
> > (0) get_src_grid_info: source grid type is 'curvilinear'
> > (0) curvilinear_to_SCRIP: calculating grid corners...
> > (0) curvilinear_to_SCRIP: one or more lat values are at the
> > (0) poles, so calculating grid corners using
> > (0) calc_SCRIP_corners_boundaries...
> > (0) calc_SCRIP_corners_boundaries
> > (0) min/max original lat: -90/-48.4644
> > (0) min/max original lon: -179.983/180
> >
> > When you run the script, do you see the error message before this output
> or after?
> >
> >
> > We've seen this problem before with other users and are completely
> mystified about where it's happening. You are correct that it appears to
> be coming from ESMF_regridding.ncl, although we've looked at the code
> multiple times and don't see the source of the issue. This has actually
> happened with earlier versions of NCL (up to 2 years ago), but it doesn't
> happen consistently.
> >
> > What I did for another user that had this problem was to take the
> "curvilinear_to_SCRIP" procedure that's in ESMF_regridding.ncl, put it in
> it's own xxxx.ncl script, and then changed all the "Opt" references to
> something else, like "cts_opt", just to see if the problem persists.
> >
> > Since I can't reproduce your problem yet, maybe you can try this
> yourself. Save the attached curv_to_scrip.ncl file, and then load it from
> the make_topogrid_debug.ncl script as follows:
> >
> > load "$NCARG_ROOT/lib/ncarg/nclscripts/esmf/ESMF_regridding.ncl"
> > load "./curv_to_scrip.ncl"
> >
> > What will happen is that the "curvilinear_to_SCRIP" in the second NCL
> script will override the one in ESMF_regridding.ncl, because I made sure
> there was an "undef" command before it.
> >
> > [As I've been typing this email and trying things out, your script is
> still running without any errors.]
> >
> > Let ncl-talk know if you still see the problem with the new script.
> >
> > Thanks,
> >
> > --Mary
> >
> >
> >
> > On Tue, Feb 13, 2018 at 8:20 AM, Kuipers Munneke, P. (Peter) <
> P.KuipersMunneke at uu.nl> wrote:
> > Dear all,
> >
> > As a regular NCL user, I stumbled upon an issue with the ESMF regridding
> software, which I have just started to use. I suspect that the Attached you
> find the cleanest version of the code that reproduces the error, in
> make_topogrid_debug.ncl. The structure of the code is an exact replica of
> https://www.ncl.ucar.edu/Applications/Scripts/ESMF_regrid_16.ncl.
> >
> > The undefined variable “opt” is not the “opt” in the
> make_topogrid_debug.ncl file. If I rename it to something else, the error
> message continues to mention the “Opt” variable instead of the renamed one.
> This makes me suspect that the error is generated in ESMF_regridding.ncl
> itself.
> >
> > I would be very grateful for your help.
> >
> > With best regards,
> > Peter
> >
> > ============
> >
> >
> > - NCL version 6.4.0
> >
> > - Data files are available at http://www.staff.science.uu.
> nl/~kuipe117/ncl_debug
> >
> > - The exact error message is
> >
> > fatal:Variable (Opt) is undefined
> > fatal:["Execute.c":8640]:Execute: Error occurred at or near line 1613
> in file $NCARG_ROOT/lib/ncarg/nclscripts/esmf/ESMF_regridding.ncl
> >
> > fatal:["Execute.c":8640]:Execute: Error occurred at or near line 3912
> in file $NCARG_ROOT/lib/ncarg/nclscripts/esmf/ESMF_regridding.ncl
> >
> > fatal:["Execute.c":8640]:Execute: Error occurred at or near line 4042
> in file $NCARG_ROOT/lib/ncarg/nclscripts/esmf/ESMF_regridding.ncl
> >
> > fatal:["Execute.c":8640]:Execute: Error occurred at or near line 51 in
> file make_topogrid_debug.ncl
> >
> > - The system is Mac OS High Sierra 10.13.3
> >
> > uname -a gives:
> >
> > Darwin mbpkuipsmunneke.soliscom.uu.nl 17.4.0 Darwin Kernel Version
> 17.4.0: Sun Dec 17 09:19:54 PST 2017; root:xnu-4570.41.2~1/RELEASE_X86_64
> x86_64
> >
> > gcc —version gives:
> >
> > Configured with: --prefix=/Library/Developer/CommandLineTools/usr
> --with-gxx-include-dir=/usr/include/c++/4.2.1
> > Apple LLVM version 9.0.0 (clang-900.0.39.2)
> > Target: x86_64-apple-darwin17.4.0
> > Thread model: posix
> > InstalledDir: /Library/Developer/CommandLineTools/usr/bin
> >
> > _______________________________________________
> > ncl-talk mailing list
> > ncl-talk at ucar.edu
> > List instructions, subscriber options, unsubscribe:
> > http://mailman.ucar.edu/mailman/listinfo/ncl-talk
> >
> >
> > <curv_to_scrip.ncl>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.ucar.edu/pipermail/ncl-talk/attachments/20180222/13c2304f/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: curv_to_scrip.ncl
Type: application/octet-stream
Size: 6522 bytes
Desc: not available
URL: <http://mailman.ucar.edu/pipermail/ncl-talk/attachments/20180222/13c2304f/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: make_topogrid_debug_mod.ncl
Type: application/octet-stream
Size: 1340 bytes
Desc: not available
URL: <http://mailman.ucar.edu/pipermail/ncl-talk/attachments/20180222/13c2304f/attachment-0001.obj>
More information about the ncl-talk
mailing list