[ADMB Users] vector vs loop and admb speed

[Jeff Laake]

Steve-

You should get the same results.  I'd be concerned if you didn't.  The
models and data are the same. They only differ in vectorized versus loop
implementation.

--jeff


On Tue, Jan 15, 2013 at 10:54 AM, Martell, Steven <
s.martell at fisheries.ubc.ca> wrote:

> Hi Jeff,
> I compiled and ran you code and Jess's code and can confirm your results
> on run time. I also checked to make sure you were getting the same answers
> in both cases and I do get the same results (see below). Not sure why?
>
> bash-3.2$ head multistate_randeff_fast.par
> # Number of parameters = 5  Objective function value = 620.912  Maximum
> gradient component = 4.45222e-06
> # Here:
> 1.85484902146
> # Hol:
> 0.959450135883
> # Die:
> 0.0109419856083
> # meanobs:
> 1.69515728550
> # sdindiv_tr:
> bash-3.2$ head multistate_randeff_slow.par
> # Number of parameters = 5  Objective function value = 620.912  Maximum
> gradient component = 4.45224e-06
> # Here:
> 1.85484902146
> # Hol:
> 0.959450135883
> # Die:
> 0.0109419856083
> # meanobs:
> 1.69515728550
> # sdindiv_tr:
>
> On 2013-01-14, at 9:22 AM, Jeff Laake <jefflaake at gmail.com>
>  wrote:
>
> >
> <multistate_randeff_slow.tpl><multistate_randeff_slow.dat><multistate_randeff_fast.tpl><multistate_randeff_fast.dat>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.admb-project.org/pipermail/users/attachments/20130115/a101e81d/attachment.html>