[ADMB Users] Problems with RE model, using FUNCTIONs for integration

[Jeff Laake]

On Fri, Nov 30, 2012 at 11:29 AM, dave fournier <davef at otter-rsch.com>wrote:

> On 12-11-30 11:15 AM, Jeff Laake wrote:
>
> Well I didn't say to remove the Initialization section unless you provide
> a pin file.
> However it may work anyway with the default values.  It appears to.
>
Yes I was just taking a shortcut on that.

>
> Those are not errors they are warnings.  I ignore all warnings about const
> stuff. In this
> case they are not a problem.   This is just a junk model to check that the
> adromb stuff
> works.  There is probably something else wrong with your model. Almost for
> sure
> you are walking out of an array. Are you using the safe library?
>
>
> I'm using ADMB11 and I thought I read that it does safe mode by default.
 I have seen the array out of bounds issue before and I don't think that
was the issue then. Now that I have a working tpl using other code for the
integral, I'll go back and investigate whether the adromb code works.
 Thanks for the verification that the warnings can be ignored in this case.

--jeff

>
>  So when I remove the INITIALIZATION_SECTION out of t.tpl,below is what I
>> get when I compile which looks very much like what I got when I did it with
>> my tpl file.  I think Mark experienced something similar. It creates a
>> t.exe but when I tried with my example the code didn't work but maybe it
>> was another issue.
>>
>> When I plow ahead and run your example t without initial values I get.
>> Model file: t
>> Negative log-likelihood: 629.367
>> Coefficients:
>>    log_tau     log_nu       beta
>>  1.7637522 -0.4109849  0.5500000
>>
>> So maybe it is working and the compile errors are unimportant. Can you
>> verify that?
>>
>> thanks --jeff
>>
>> > compile_admb("t",re=T,verbose=**T)
>> compiling with args: ' -r ' ...
>> compile output:
>>   *** tpl2rem t xxglobal.tmp xxhtop.tmp header.tmp xxalloc1.tmp
>> xxalloc2.tmp xxalloc3.tmp xxalloc4.tmp xxtopm.tmp xxalloc6.tmp         1
>> file(s) copied. tfile1 tfile2 tfile3 tfile4         1 file(s) copied.  ***
>> adcomp -r t g++ -c -O3 -Wno-deprecated -D__GNUDOS__ -Dlinux -DOPT_LIB
>> -DUSE_LAPLACE -fpermissive -I. -I"c:/admb\include" -I"c:/admb\contrib" -o
>> t.obj t.cpp  *** adlink -r t g++ -s -static -L"c:/admb\lib"
>> -L"c:/admb\contrib"  t.obj -ldf1b2o -ladmod -lcontribo -ladt -lado -ldf1b2o
>> -ladmod -ladt -lcontribo -lado -lcontribo -o t  Done
>> compile log:
>> t.cpp: In member function 'dvariable model_parameters::trapzd(**double,
>> double, int, const dvariable&)':
>> t.cpp:102:35: warning: passing 'const dvariable' as 'this' argument of
>> 'dvariable& dvariable::operator=(const prevariable&)' discards qualifiers
>> [-fpermissive]
>> t.cpp:109:36: warning: passing 'const dvariable' as 'this' argument of
>> 'dvariable& dvariable::operator=(const prevariable&)' discards qualifiers
>> [-fpermissive]
>> t.cpp: In member function 'df1b2variable df1b2_parameters::trapzd(**double,
>> double, int, const df1b2variable&)':
>> t.cpp:254:35: warning: passing 'const df1b2variable' as 'this' argument
>> of 'df1b2variable& df1b2variable::operator=(const df1b2variable&)' discards
>> qualifiers [-fpermissive]
>> t.cpp:261:36: warning: passing 'const df1b2variable' as 'this' argument
>> of 'df1b2variable& df1b2variable::operator=(const df1b2variable&)' discards
>> qualifiers [-fpermissive]
>> >
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.admb-project.org/pipermail/users/attachments/20121130/15f0a4be/attachment.html>