Re: NCL error message

Hi Oli,

We actually do have several test cases that we are able to use for
reproducing this problem. However, it doesn't hurt to have more so
when/if we fix this bug, we can make sure it fixes the problem
everywhere.

On Mar 26, 2010, at 9:56 AM, <Oliver.Fuhrer@meteoswiss.ch> <Oliver.Fuhrer@meteoswiss.ch
> wrote:

> Mary,
>
> I have to take back on unreproducible. Since the same script crashed
> only very rarely, I assumed the problem was unreproducible. But when
> we use the exact same input data, the script crashes reproducibly.
> That probably makes sense, as it is linked to the contouring
> algorithm and only happens for certain data. Would you be interested
> to have a test case? But since you know the problem, you probably
> already have one I assume.
>
>> Are you offering to host?! :-) We certainly love coming to Europe,
>> but don't currently have any plans for a workshop there.
>> It's not out of the question.
>> We are tentatively planning a workshop in the fall somewhere,
>> so we'd have to work around that.
>> Would you like to propose a location, date, etc? Or are you talking
>> about MPI again?
>
> Ok, if it would be Switzerland, the host would probably be ETH and
> not MeteoSwiss. From our side, we would at least 2 people interested
> in participating, also depending on the location. I try to ask
> around (Bjoern but also at ETH) if there is a critical mass of
> people interested.
>
> Could you give me an idea of the budget that would be required for
> the workshop?

I'll reply shortly offline on this question.

--Mary

>
> Cheers,
> Oli
>
>
> ________________________________________
>
> Oliver Fuhrer
> Numerical Models
>
> Federal Departement of Home Affairs FDHA
> Federal Office of Meteorology and Climatology MeteoSwiss
>
> Kraehbuehlstrasse 58, P.O. Box 514, CH-8044 Zurich, Switzerland
>
> Tel. +41 44 256 93 59
> Fax +41 44 256 92 78
> oliver.fuhrer@meteoswiss.ch
> www.meteoswiss.ch - First-hand information
>
>
>
>> -----Original Message-----
>> From: Mary Haley [mailto:haley@ucar.edu]
>> Sent: Freitag, 26. März 2010 14:53
>> To: Fuhrer Oliver
>> Cc: ncl-talk@ucar.edu; Walser Andre
>> Subject: Re: NCL error message
>>
>> Hi Oli,
>>
>> I'm afraid you've run into one of our most infamous and frustrating
>> errors. It's very high on our list to fix, but it's unfortunately a
>> difficult problem to debug.
>>
>> We have an FAQ on this. Please take a look and see if it helps:
>>
>> http://www.ncl.ucar.edu/FAQ/#err_msgs_006
>>
>> --Mary
>>
>> On Mar 26, 2010, at 5:19 AM, <Oliver.Fuhrer@meteoswiss.ch>
>> <Oliver.Fuhrer@meteoswiss.ch> wrote:
>>
>>> Hi all,
>>>
>>> Has anyone encountered this error message using NCL? We are running
>>> quite a lot of NCL jobs every day and this error message occurrs
>>> sporadically und unreproducibly...
>>>
>>> (0) jmb_getvar: reading HSURF named HH_GDS10_SFC in file
>>> fatal:ContourPlotDraw: ARSCAM/ARPRAM - ALGORITHM FAILURE
>>> fatal:ContourPlotDraw: draw error
>>> warning:WorkstationDeactivate: workstation not active or not opened
>>> (0) jmb_getvar: reading TOT_PREC named TOT_PREC_GDS10_SFC_acc3h in
>>> file
>>>
>>> Cheers,
>>> Oli
>>>
>>> The script where the error occurs is the following...
>>>
>>> ; setup plot
>>> name = "${model}_${NAME}_${start}.gin"
>>> format = "png"
>>> xsize = 500.0
>>> ysize = 400.0
>>> shift =392
>>>
>>> ; set plot domain
>>> xs = 400.0
>>> xe = 900.0
>>> ys = 0.0
>>> ye = 400.0
>>>
>>> ; open files
>>> c = addfile("lfff00000000c.grb", "r")
>>> f = addfile("_NCLINP_${start}_delta03h_c.grb", "r")
>>>
>>> ; read data
>>> d = jmb_getvar(f, "TOT_PREC") ; reads a variable from
>> the grib
>>> file
>>>
>>> ; geo-reference
>>> jmb_getgrid(f, c, d) ; attaches
>> geo-referencing data to
>>> the field d
>>>
>>> ; compute swissgrid coordinates
>>> x = latlon2swissx(d@lat2d, d@lon2d)
>>> y = latlon2swissy(d@lat2d, d@lon2d)
>>>
>>> ; remove geo-referencing
>>> jmb_rmgrid(d) ; removed
>> geo-referencing data from
>>> field d
>>>
>>> ; set plot size in inch
>>> dpi = ceil(xsize/8.5/72.0)*72.0
>>> xinch = xsize/dpi
>>> yinch = ysize/dpi
>>>
>>> ; open graphic port
>>> r = "ps"
>>> r@wkDeviceLowerX = 0.0
>>> r@wkDeviceUpperX = xinch*72.0
>>> r@wkDeviceLowerY = 0.0
>>> r@wkDeviceUpperY = yinch*72.0
>>> wks = gsn_open_wks(r, name)
>>> delete(r)
>>> rct = jmb_set_ct(wks,"rr_ninjo_21lev",False)
>>>
>>> ; setup plot
>>> r = rct
>>> r@gsnMaximize = False
>>> r@gsnTickMarksOn = False
>>> r@vpWidthF = 1.0
>>> r@vpHeightF = 1.0
>>> r@vpXF = 0.0
>>> r@vpYF = 1.0
>>> r@vpKeepAspect = True
>>> r@sfXArray = x
>>> r@sfYArray = y
>>> r@trGridType = "TriangularMesh"
>>> r@trXMinF = xs
>>> r@trYMinF = ys
>>> r@trXMaxF = xe
>>> r@trYMaxF = ye
>>> r@cnFillOn = True
>>> r@cnLinesOn = False
>>> r@cnLineLabelsOn = False
>>> r@cnInfoLabelOn = False
>>>
>>> ; make plot
>>> cn = gsn_contour(wks, d, r)
>>> delete(r)
>>>
>>> ; close graphic port
>>> delete(wks)
>>>
>>>
>>>
>>> ________________________________________
>>>
>>> Oliver Fuhrer
>>> Numerical Models
>>>
>>> Federal Departement of Home Affairs FDHA
>>> Federal Office of Meteorology and Climatology MeteoSwiss
>>>
>>> Kraehbuehlstrasse 58, P.O. Box 514, CH-8044 Zurich, Switzerland
>>>
>>> Tel. +41 44 256 93 59
>>> Fax +41 44 256 92 78
>>> oliver.fuhrer@meteoswiss.ch
>>> www.meteoswiss.ch - First-hand information
>>>
>>> _______________________________________________
>>> ncl-talk mailing list
>>> List instructions, subscriber options, unsubscribe:
>>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>
>>
> _______________________________________________
> ncl-talk mailing list
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk

_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Sun Mar 28 07:13:39 2010