[MITgcm-support] Re: GAD_SMOLARKIEWICZ_HACK and adjoint code

[Patrick Heimbach]

Hi Suneet,

I haven't adjointed this advection scheme yet.
Very likely you're facing extensive recomputations
which is causing the excessive time.
Will try to take a look.
Also, sorry I haven't gotten back to you on earlier mail,
am scrambling to catch up.

-Patrick

On Jul 22, 2009, at 1:51 AM, Suneet Dwivedi wrote:

> Hi Patrick,
> When i use "GAD_SMOLARKIEWICZ_HACK" in gad_options.h to achieve  
> strict positivity of ptracer values, the adjoint integration is  
> taking forever (almost 10 times) to complete. When i undef the  
> above option, integration finishes normally. Any ideas why this is  
> happening? How i can speed up the adjoint integration with above  
> option defined?
> Thanks,
> Suneet

---
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