[MITgcm-devel] Fwd: possible bug - dropped myThid

[Patrick Heimbach]

For the record.
-p.

Begin forwarded message:

> From: Ralf Giering <Ralf.Giering at fastopt.de>
> Date: June 14, 2011 4:55:44 AM EDT
> To: "Cheng, Benny N (3244)" <Benny.N.Cheng at jpl.nasa.gov>, "heimbach at mit.edu" <heimbach at mit.edu>
> Subject: Re: possible bug query
> 
> Hi Benny,
> 
> I have fixed the bug regarding the missing mythid argument.
> The new TAF version (2.1.12) generates code similar to the older versions (< 1.9.81).
> 
> Regards,
> Ralf
> 
> On Sunday 05 June 2011, Cheng, Benny N (3244) wrote:
> > 
> >  Dear FastOpt:
> > 
> >     We would like to report a possible bug in the latest version of TAF (2.1.11) that we are using to
> >  obtain the adjoint of the MITgcm (version c62t_201103) code.
> > 
> >  The problem concerns the dropping of some arguments in subroutines that are essential to the calls within
> >  the subroutine itself. For example, in the following TAF generated code,
> > 
> >       subroutine admom_fluxform( bi, bj, imin, imax, jmin, jmax, k, kup,
> >      $ kdown, kapparu, adkapparu, kapparv, adkapparv, adfveru, adfverv,
> >      $adgudiss, adgvdiss )
> > 
> >      ...
> >      integer mythid
> >      ...
> > 
> >      call mom_calc_visc( bi,bj,k,hfacz,r_hfacz,mythid )
> >      ...
> > 
> > 
> >  We see that the mythid argument is missing, but it is a required argument inside the subroutine. Arbitrary 
> > large values of mythid have been identified to be the cause of code crashes. It also appears that
> > this routine does not need a flow file. 
> > 
> >  Curiously, if we use an older TAF version (1.9.73), the mythid argument is restored, and the code runs fine 
> > (although a manual scan of the adjoint code still show some possible missing required mythid arguments).
> > 
> >  I'm attaching the Makefile, the adjoint input files ad_input_code*.f, the ad_taf_output.f, and taf_ad* (one compressed
> > tar file) so you can take a closer look. If you need more files, let us know.
> > 
> >  Regards,
> >  Benny Cheng, NASA-JPL
> > 
> 
> 
> -- 
> ###################################################
>  Dr. Ralf Giering
>  FastOpt GmbH
>  Lerchenstr. 28a, 22767 Hamburg, Germany
>  Tel.: +49 40 48096347
>  Fax : +49 40 48096357
>  Email: Ralf.Giering at FastOpt.de
>  URL  : http://www.FastOpt.de
>  Handelsregister Amtsgericht Hamburg, HRB 103390
>  Geschäftsführer: Ralf Giering, Thomas Kaminski
> ###################################################

---
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 --------------
An HTML attachment was scrubbed...
URL: <http://mitgcm.org/pipermail/mitgcm-devel/attachments/20110614/82e2ca45/attachment-0001.htm>