[ADMB Users] problems with an old model in new version of ADMB

[Tom Wainwright]

Dear ADMB friends,

I'm trying to update an older ADMB model, and am having some problems getting it
to run under the latest version.

OLD VERSION:
$ ./BaySAM -?
AD Model Builder version 4.5.2 copyright (c) 1993 2000 Otter Research Ltd

NEW VERSION:
$ ../BaySAM -?
AD Model Builder version 9.0.x
Copyright (c) 2008, 2009 Regents of the University of California

(New version ADMB is a fresh install from admb-9.1-linux-gcc4.4-64bit.zip on a
Linux x64 [Fedora 11] system with gcc 4.4.1. Running 'make all' from the
top-level admb directory runs to completion with only a couple errors, but these
don't seem to relate to my problem.)

THE PROBLEM:

The old version of the code still runs fine on my system, but when I build and
run the model in the new version of ADMB, I get two errors:

FIRST ERROR:

Output starts with an error:
"Error trying to open data input file /home/twain/tmp/admbtest/baysam/baysam.dat"

The file path should end in "/BaySAM/BaySAM.dat". I see in the source code for
model7.cpp (line 124) that the model name is converted to all lower case, which
didn't used to happen. This isn't a big problem here, because I can work around
it using the '-ind' option, and the model works without that empty file anyway.

SECOND ERROR:

However, after running the estimation, the output ends with a second terminal error:
"Error in dvar_matrix write"

(This appears where "Start of REPORT_SECTION" appeared in the old version's
output.) I suspect this may also be a file path problem, but I can't track it down.

Does anyone have suggestions? I've attached the full output from the run.

Thanks!
Tom Wainwright

--
Tom Wainwright
NOAA Northwest Fisheries Science Center
Newport, Oregon
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The contents of this message are mine personally and do not necessarily
reflect any position of the Government or the National Oceanic and
Atmospheric Administration.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

[Tom Wainwright]

In my previous message, I mentioned a couple errors (each repeated twice) when I
run "make all" on a fresh installation. I just want to check if these are
anything I should worry about (i.e. installation problems).

System: Linux x64 (Fedora 11) with gcc 4.4.1. ADMB installed from
admb-9.1-linux-gcc4.4-64bit.zip.

Error 1: (Probably trivial.) The 'forest' model compiles and runs, but output
ends with these two lines:
The number created was 1770 The number destroyed was 1769
make: [forest] Error 1 (ignored)

(This looks like the code for some reason has a non-zero return.)

Error 2: (Might be important?) The 'liver_gamma' model results in the following:
*** tpl2rem liver_gamma
cat: xxalloc4.tmp: No such file or directory
cat: xxalloc5.tmp: No such file or directory
Error executing command cat xxglobal.tmp xxhtop.tmp header.tmp
xxalloc1.tmp xxalloc2.tmp xxalloc3.tmp xxalloc4.tmp
xxalloc5.tmp xxtopm.tmp xxalloc6.tmp > liver_gamma.cpp
However, the model still builds and runs. Can't tell if it runs correctly,
though. This may be the same problem discussed in the thread '"xxglobal.tmp"
error with tpl2rem' started by Bolker last August.

Are these typical problems with 'make all', or should I check my installation?

--
Tom Wainwright
NOAA Northwest Fisheries Science Center
Newport, Oregon
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The contents of this message are mine personally and do not necessarily
reflect any position of the Government or the National Oceanic and
Atmospheric Administration.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

_______________________________________________
Users mailing list
Us...@admb-project.org
http://lists.admb-project.org/mailman/listinfo/users

[Johnoel Ancheta]

It's a permission issue in the new distributions which has been
corrected but not yet released.  The way to fix the problem is to run

$ chmod -R u+w examples

in the admb folder, then run

$ make simple

Let me know if it works,
Johnoel

> necessarilyreflect any position of the Government or the

[Tom Wainwright]

Thanks Johnoel, but 'make simple' already ran fine. It's 'make all' that
generates errors, and only in two of the many examples.

Tom