[MITgcm-devel] Problem with recent pkg/seaice changes

[Patrick Heimbach]

Hi Jean-Michel,

don't think ITD was the culprit.
Recomputations came from etaZ calculation.
A simple initialization should fix this
(hope I didn't overlook something here).

p.

On Nov 10, 2012, at 2:46 PM, Jean-Michel Campin <jmc at ocean.mit.edu> wrote:

> Hi Patrick and Martin,
> 
> Apart from the bug in seaice_growth.F (double declaration of msgBuf)
> that has been fixed this morning (revision 1.81),
> there are still few issues for the Adjoint testing:
> 1) experiment lab_sea : the 4 AD runs are taking much longer:
> on baudelaire, gfortran -devel:
> yesterday (Nov 09):
> User:    36.980 sys:  0.410 clock:    74.379 <- lab_sea.evp
> User:    27.780 sys:  0.280 clock:    28.135 <- lab_sea.noseaicedyn
> User:    27.830 sys:  0.380 clock:    54.892 <- lab_sea.noseaice
> User:   100.590 sys:  0.310 clock:   101.245 <- lab_sea
> and this morning (Nov 10):
> User:  6360.020 sys: 14.070 clock:  6422.029 <- lab_sea.evp
> User:  5302.890 sys: 13.750 clock:  5328.328 <- lab_sea.noseaicedyn
> User:  5164.820 sys: 13.600 clock:  5215.376 <- lab_sea.noseaice
> User:  5798.870 sys: 13.930 clock:  5825.861 <- lab_sea
> 
> This is so much slower that I had few AD tests that did not finish in time
> (exceed the batch-queue limit).
> 
> 2) experiment global_ocean.cs32x15 : the 4 AD tests are failling 
> (wong AD grad, 1 or no digits match)
> 
> I am going to check with the version of pkg/seaice just before the 
> latest ITD changes.
> 
> Cheers,
> Jean-Michel
> 
> _______________________________________________
> MITgcm-devel mailing list
> MITgcm-devel at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-devel


---
Patrick Heimbach | heimbach at mit.edu | http://www.mit.edu/~heimbach
MIT | EAPS 54-1420 | 77 Massachusetts Ave | Cambridge MA 02139 USA
FON +1-617-253-5259 | FAX +1-617-253-4464 | SKYPE patrick.heimbach

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 1588 bytes
Desc: not available
URL: <http://mitgcm.org/pipermail/mitgcm-devel/attachments/20121110/a2abb4e0/attachment.p7s>