[MITgcm-devel] Re: it's probably my faults again, but ...

[Martin Losch]

Hi there,

I have finally checked in my little test. you'll find it in  
MITgcm_contrib/mlosch/cs32test.

Martin
On 9 Mar 2007, at 17:24, Jean-Michel Campin wrote:

> Martin,
>
> Could it be that you are on a platform that do special
> (or don't do) regarding initialisation ? And that this
> kind of problem is only there with imissing tiles (="hole") ?
> I managed to pass a full test last month on weddell, with
> Constantinos magic option to trap un-initialized (local)
> variables. But this missing tiles is not part of the test-list.
> Would be a good idea to check-in in Contrib the whole stuff,
> as Chris suggested.
> Jean-Michel
>
> On Fri, Mar 09, 2007 at 04:25:15PM +0100, Martin Losch wrote:
>> Yes, it points to a problem in EXCH2_UV.
>> Which variables should I check to rule out an EXCH2_Z problem?
>> Again, maybe I am just making a stupid mistake.
>>
>> Martin
>>
>> On 9 Mar 2007, at 16:16, Jean-Michel Campin wrote:
>>
>>> Hi Martin,
>>>
>>> Just to be sure:
>>> Do you have the impression (or more than just an impression)
>>> that the Pb is in the EXCH_UV, and not in the other EXCH ?
>>> And is EXCH_Z out of the list of suspect ?
>>>> From the picture you sent, it seems to me that there is a clear Pb
>>> (not a rotation issue, in this sense I don't follow Patrick here).
>>> But not sure which EXCH is to blame for.
>>>
>>> Jean-Michel
>>>
>>> On Fri, Mar 09, 2007 at 10:04:33AM -0500, Patrick Heimbach wrote:
>>>>
>>>> Hi Martin,
>>>>
>>>> not sure whether you're aware that what you plotted and labeled
>>>> 'taux' and 'EXFuwind' is not the zonal component on all faces.
>>>> It is actually tauy and EXFvwind on some faces
>>>> (due to cs topology).
>>>>
>>>> I wonder whether what you're seeing is either
>>>> * an artifact of the fact that you're looking at rotated fields
>>>>  on some faces
>>>> * an artifact of diagnostics in plotting vector plots on the cube
>>>>
>>>> Evidence for this is that you confirm that the scalar fields
>>>> do not show the holes.
>>>>
>>>> -p.
>>>>
>>>>
>>>>
>>>> On Mar 9, 2007, at 3:05 AM, Martin Losch wrote:
>>>>
>>>>> Well, tell Patrick, that the scalar hflux (diagnostics EXFqnet,
>>>>> EXFempmr) does not have these problems, also TFLUX and SFLUX  
>>>>> appear
>>>>> to be unaffected (although SFLUX does look funny, but not near the
>>>>> edges of the tiles).
>>>>>
>>>>> Martin
>>>>>
>>>>> PS I am not aware of any recent studies on poppy seed bagels and
>>>>> Apple computers.
>>>>> On 8 Mar 2007, at 20:31, Dimitris Menemenlis wrote:
>>>>>
>>>>>> Martin, Patrick asks whether there is a problem with any of the
>>>>>> scalar surface flux diagnostics?  Patrick specifically asks
>>>>>> whether you can reproduc your plot with hflux.
>>>>>>
>>>>>> (His brand new MacPro is broken, so I have to write this from  
>>>>>> my 4-
>>>>>> year old PC.)
>>>>>>
>>>>>> D.
>>>>>> _______________________________________________
>>>>>> 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
>>>>
>>>> ---
>>>> Dr Patrick Heimbach | heimbach at mit.edu | http://www.mit.edu/ 
>>>> ~heimbach
>>>> MIT | EAPS, 54-1518 | 77 Massachusetts Ave | Cambridge, MA  
>>>> 02139, USA
>>>> FON: +1-617-253-5259 | FAX: +1-617-253-4464 | SKYPE:  
>>>> patrick.heimbach
>>>>
>>>>
>>>> _______________________________________________
>>>> 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