[MITgcm-devel] additional high-precision output for testreport

[Patrick Heimbach]

Ed,

I've updated the gradient check package.
Standard output now has lines that look like
(PID.TID 0000.0001)       precision_grdchk_result ADM     1.84044933456284E+14     1.74865283966910E+08
(PID.TID 0000.0001)       precision_grdchk_result ADM     1.84044933456284E+14     1.31884844820653E+08
(PID.TID 0000.0001)       precision_grdchk_result ADM     1.84044933456284E+14     1.96420903150567E+07
(PID.TID 0000.0001)       precision_grdchk_result ADM     1.84044933456284E+14    -9.99869528291716E+06

The second-last column is the reference cost function,
the last column is the ADM (or TLM) gradient.
Precision for both numbers should be compared,
but pass/fail should be based on the latter (the gradient).

In a next step we may think about also
grep-ing the taf_ad.log for
'RECOMPUTATION WARNING' and 'EXTENSIVE RECOMPUTATION WARNING'
(those indicate that we may have broken the code in certain ways).

-Patrick



-------- Original Message --------
Subject: [MITgcm-cvs] MITgcm/pkg/grdchk CVS Commit
Date: Tue, 4 Nov 2003 15:47:43 -0500
From: Patrick Heimbach <heimbach at mitgcm.org>
Reply-To: MITgcm-cvs at mitgcm.org
To: mitgcm-cvs at mitgcm.org

Update of /u/gcmpack/MITgcm/pkg/grdchk
In directory forge:/tmp/cvs-serv6017

Modified Files:
	grdchk.h grdchk_loc.F grdchk_main.F
Log Message:
o merged from ecco-branch
   (remaining bug fixes for obcs gradient checks)
o additional high-precision output for testreport
   (grep for precision_grdchk_result)

_______________________________________________
MITgcm-cvs mailing list
MITgcm-cvs at mitgcm.org
http://dev.mitgcm.org/mailman/listinfo/mitgcm-cvs

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