[Mitgcm-support] Re: barrier barrier

[mitgcm-support at dev.mitgcm.org]

Either, you've included a subroutine in the original code with
a wrong parameter list (leading to a wrong value of myThid),
or you've added a call to a S/R in the adjoint code.

In the latter case it may seem to you that mythid is
available, but often myThid is not passed as an argument
but re-defined locally and set to zero, because TAMC/TAF
has determined that it's not needed.
So the problem is the call statement that you've added
into the adjoint code "by hand".

Does any of this apply to your code?

P.

Chris Hill wrote:
> 
> Jake,
> 
>  This sounds familiar, but I can't remeber what causes it.
> I am cc'ing to support.
> Anyone else know?
> 
> Chris
> ----- Original Message -----
> From: "Geoffrey GEBBIE" <gebbie at MIT.EDU>
> To: <cnh at plume.mit.edu>
> Sent: Monday, December 10, 2001 1:17 PM
> Subject: barrier barrier
> 
> >
> > hi Chris: I'm running the global passive tracer run with the "sucking"
> > term. Forward model run is fine. The first barrier in the adjoint
> > computation part of the code hits the first !!!PANIC!!! block inside
> > s/r barrier. It passes the argument myThid=1 but it becomes myThid=0
> > immediately inside the s/r. Any ideas?
> >
> > Jake
> > p.s. i'm at woods hole today and tomorrow.
> >
> >
> >

-- 
_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/_/

Patrick Heimbach ........................ FON: +1/617/253-5259
Massachusetts Institute of Technology ... FAX: +1/617/253-4464 
EAPS, Room 54-1518 ...........................................
77 Massachusetts Avenue .............. mailto:heimbach at mit.edu
Cambridge MA 02139, U.S.A. ..... http://www.mit.edu/~heimbach/