[MITgcm-devel] diagnostics for SI_Tice1, SI_Tice2, SI_Qice1, and SI_Qice2

Dimitris Menemenlis menemenlis at sbcglobal.net
Mon May 5 19:06:59 EDT 2008 

JM, the error message is fine and does not need to be changed.
For some perverse reason, I have always worked with nuperlist=1.
I must have thought a long time ago that it saved me some space
for the hi-res cube integrations.  I will change it back to 50, the  
default.
Thanks, D.

Dimitris Menemenlis <menemenlis at sbcglobal.net>
5056 Oakwood Ave, La Canada, CA 91011-2450
tel/fax: 818-790-6735;  cell: 818-625-6498

On May 5, 2008, at 4:00 PM, Jean-Michel Campin wrote:

> Dimitris,
>
> Again, I don't think it's so strange (all the diagnostics which have a
> "counter" mate behave like this).
> And the "default" DIAGNOSTICS_SIZE.h provides numperlist = 50,
> you may need to increase it (for long list) but there is no
> real reason to decrease it.
> Also, I though the error message was quiet explicit:
>> DIAGNOSTICS_SET_POINTERS: Not enough space for all active diagnostics
>> (from data.diagnostics)
>> DIAGNOSTICS_SET_POINTERS: numperlist=     1 but needs at least     2
>> STOP ABNORMAL END: S/R DIAGNOSTICS_SET_POINTERS statement executed
> but may be it could be improved.
>
> Jean-Michel
>
> On Mon, May 05, 2008 at 03:38:27PM -0700, Dimitris Menemenlis wrote:
>> Jean-Michel, the problem is not memory space.  The problem is  
>> numperlist
>> For some reason, these four diagnostics, on their own, request
>> numperlist=2.
>> It's a bit strange behavior.  First, I don't understand why SI_Tice1
>> should be mated
>> with SI_Fract.  I thought that mating was only needed for vector
>> quantities to take
>> care of edges properly.  Second, even mated quantities, e.g.,  
>> UVELMASS
>> and
>> VVELMASS or SIuice and SIvice, do not require numperlist=2, unless of
>> course
>> I try to save two fields in the same file.
>>
>> This is not a high priority item.  I just thought I would report it  
>> so
>> we have a record
>> and maybe there is a simple explanation that I am not seeing and also
>> because I
>> did end up wasting a few hours tracking it down so others don't
>> stumble on same issue.
>>
>> How is the downslope parameterization coming along?  Will you have  
>> any
>> time
>> to work on it in next few weeks?
>>
>> Cheers, Dimitris
>>
>> Dimitris Menemenlis <menemenlis at jpl.nasa.gov>
>> Jet Propulsion Lab, California Institute of Technology
>> MS 300-323, 4800 Oak Grove Dr, Pasadena CA 91109-8099
>> tel: 818-354-1656;  cell: 818-625-6498;  fax: 818-393-6720
>>
>> On May 5, 2008, at 3:05 PM, Jean-Michel Campin wrote:
>>
>>> Hi Dimitris,
>>>
>>> On Mon, May 05, 2008 at 02:45:50PM -0700, Dimitris Menemenlis wrote:
>>>> A little more info towards solving the puzzler. SI_Fract is defined
>>>> as
>>>> the diagmate of SI_Tice1:
>>>>
>>>>      diagName  = 'SI_Tice1'
>>>>      diagTitle = 'Sea-Ice Temperature, 1srt layer (area weighted)'
>>>>      diagUnits = 'degC            '
>>>>      diagCode  = 'SM  C   M1      '
>>>>      CALL DIAGNOSTICS_ADDTOLIST( diagNum,
>>>>   I     diagName, diagCode, diagUnits, diagTitle, numFract,
>>>> myThid )
>>>>
>>>> but why does pkg/diagnostics request to write the two outputs in  
>>>> the
>>>> same file.  The behavior below
>>>> remains even when diagnostic output for SI_Fract is specified in a
>>>> separate output file.
>>>>
>>>> D.
>>>
>>> I don't think you need to output also SI_Fract when you select
>>> SI_Tice1 ; but you do need to provide enough (memory) space:
>>>>> (PID.TID 0000.0001) *** ERROR *** DIAGNOSTICS_SET_POINTERS: Not
>>>>> enough space for all active diagnostics (from data.diagnostics)
>>>
>>> Jean-Michel
>>>
>>>>
>>>> Dimitris Menemenlis <menemenlis at jpl.nasa.gov>
>>>> Jet Propulsion Lab, California Institute of Technology
>>>> MS 300-323, 4800 Oak Grove Dr, Pasadena CA 91109-8099
>>>> tel: 818-354-1656;  cell: 818-625-6498;  fax: 818-393-6720
>>>>
>>>> On May 5, 2008, at 2:15 PM, Dimitris Menemenlis wrote:
>>>>
>>>>> There is some strange behavior in requesting diagnostics for
>>>>> SI_Tice1, SI_Tice2, SI_Qice1, and SI_Qice2.
>>>>> For example if I put the following in the data.diagnostics file:
>>>>>
>>>>> &diagnostics_list
>>>>> frequency(1) = 2635200.,
>>>>> fields(1,1) = 'SI_Tice1',
>>>>> filename(1) = 'SI_Tice1',
>>>>> &
>>>>>
>>>>> The model stops and I get the following error message:
>>>>>
>>>>> (PID.TID 0000.0001) DIAGNOSTICS_SET_LEVELS: done
>>>>> (PID.TID 0000.0001)  Total Nb of available Diagnostics: ndiagt=
>>>>> 273
>>>>> (PID.TID 0000.0001) SETDIAG: Allocate  1 x  1 Levels for  
>>>>> Diagnostic
>>>>> #   205 SI_Tice1
>>>>> (PID.TID 0000.0001) SETDIAG: Allocate  1 x  1 Levels for  
>>>>> Count.Diag
>>>>> #   201 SI_Fract
>>>>> DIAGNOSTICS_SET_POINTERS: Not enough space for all active
>>>>> diagnostics (from data.diagnostics)
>>>>> DIAGNOSTICS_SET_POINTERS: numperlist=     1 but needs at  
>>>>> least     2
>>>>> STOP ABNORMAL END: S/R DIAGNOSTICS_SET_POINTERS statement executed
>>>>>
>>>>> (PID.TID 0000.0001) *** ERROR *** DIAGNOSTICS_SET_POINTERS: Not
>>>>> enough space for all active diagnostics (from data.diagnostics)
>>>>> (PID.TID 0000.0001) *** ERROR *** DIAGNOSTICS_SET_POINTERS:
>>>>> numperlist=     1 but needs at least     2
>>>>>
>>>>> It's as if I was also requesting diagnostics for SI_Fract in the  
>>>>> sam
>>>>> file, but I am not.  Same thing happens for
>>>>> SI_Tice2, SI_Qice1, and SI_Qice2.
>>>>>
>>>>> Is this normal-known behavior or is this a bug?
>>>>>
>>>>> Dimitris
>>>>>
>>>>> Dimitris Menemenlis <menemenlis at jpl.nasa.gov>
>>>>> Jet Propulsion Lab, California Institute of Technology
>>>>> MS 300-323, 4800 Oak Grove Dr, Pasadena CA 91109-8099
>>>>> tel: 818-354-1656;  cell: 818-625-6498;  fax: 818-393-6720
>>>>>
>>>>> _______________________________________________
>>>>> MITgcm-devel mailing list
>>>>> MITgcm-devel at mitgcm.org
>>>>> http://mitgcm.org/mailman/listinfo/mitgcm-devel
>>>>
>>>> _______________________________________________
>>>> MITgcm-devel mailing list
>>>> MITgcm-devel at mitgcm.org
>>>> http://mitgcm.org/mailman/listinfo/mitgcm-devel
>>> _______________________________________________
>>> MITgcm-devel mailing list
>>> MITgcm-devel at mitgcm.org
>>> http://mitgcm.org/mailman/listinfo/mitgcm-devel
>>
>> _______________________________________________
>> MITgcm-devel mailing list
>> MITgcm-devel at mitgcm.org
>> http://mitgcm.org/mailman/listinfo/mitgcm-devel
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel

More information about the MITgcm-devel mailing list