Date: Wed, 31 Jan 1996 11:08:45 -0800
From: gery@atmos.phy.pdx.edu (Gerhard Gross)
Message-Id: <9601311908.AA16858@atmos.phy.pdx.edu>
To: ccm-users@ncar.ucar.edu
Greetings CCM2 users.
A couple of weeks ago I started work on trying to get CCM2 to run on my
RS 6000 (actually it belongs to the school). The UNIX script that came
with the NCAR package call "runibm.sh" is the one I'm running. It
runs an R15 case with mostly default values. This compiles
and executes properly, giving me normal looking output files,
with the following exception. But when I examined the output file
"bchr15.out" I found that some of the numbers were replaced by NaNQ.
This only occured in three areas of the file, while the other numbers
came out to be exactly what they are supposed to be compared with the
NCAR package file "RESOLUTION.TEXT". The numbers missing in my output
file are of the same range of value as those appearing in the file so
overflow doesn't seem to be likely. The three areas giving me NaNQ are
included below, in case it will help someone help me with my problem.
The compile statement used in the run script is:
xlf -c -qautodbl=dblpad -NS2000 $CASE.f >&! listing
If I leave out the -qautodbl=dblpad option, the program exits pre-
maturely and gives me a runtime error of No Convergence after 10
iterations. So that looks OK.
If anyone can give me some ideas it would be greatly appreciated.
INIDAT: DRY MASS OF INITIAL DATA BEFORE CORRECTION = NaNQ
MASS WILL BE HELD = 9.8222000000E+04
MASS OF MOISTURE AFTER REMOVAL OF NEGATIVES = NaNQ