[MITgcm-devel] Fwd: TAF issue for "special" parameter lists

Jean-Michel Campin jmc at ocean.mit.edu
Tue Feb 28 22:01:57 EST 2006


Hi Patrick,

I am about to check-in a cleaner version of gad_calc_rhs.F, with one
more argument.
I removed all the #ifdef ALLOW_ADAMSBASHFORTH_3 
and if this causes problems for the adjoint, we can always 
go back and put all the "IF ( applyAB_onTracer ) THEN" blocks 
within #ifndef ALLOW_AUTODIFF_TAMC (I have already a version
for this, in case).
I run the 3 adm test on faulks, and the number are OK but it seems
to be slower than before.

Cheers,
Jean-Michel

On Mon, Feb 27, 2006 at 02:54:23PM -0500, Patrick Heimbach wrote:
> 
> Hi Jean-Michel,
> 
> that solution works (because it avoids variable 
> aliasing in the parameter list).
> So that's all set.
> 
> Thanks
> -Patrick
> 



More information about the MITgcm-devel mailing list