[ADMB Users] Help test admb-11.2

[Tom Wainwright]

Johnoel,

You also asked about runtime speeds, so I did a quick comparison on a
random-effects model I'm currently working on. I have three versions of
ADMB installed as of today: 11.1 installed from source, 11.2pre installed
from the intel 64bit binary, and 11.2pre installed from source.  I built
and ran the model 3 times with each compiler, with these results:

1) ADMB 11.1 built from source (g++ 4.4.7):

BUILD:
$ /usr/bin/time admb -r euly
0.67user 0.03system 0:00.73elapsed
0.67user 0.05system 0:00.75elapsed
0.64user 0.03system 0:00.71elapsed

RUN:
$ /usr/bin/time ./euly
7.23user 1.17system 0:08.41elapsed
7.25user 1.17system 0:08.43elapsed
7.29user 1.14system 0:08.44elapsed

2) ADMB 11.2pre prebuilt binary admb-11.2pre-centos6-intel2013-64bit.zip

BUILD:
$ /usr/bin/time admb11.2 -r euly
0.66user 0.05system 0:00.74elapsed 96%CPU
0.66user 0.05system 0:00.74elapsed
0.64user 0.04system 0:00.71elapsed

RUN:
$ /usr/bin/time ./euly
7.60user 1.19system 0:08.79elapsed
7.60user 1.16system 0:08.77elapsed
7.28user 1.17system 0:08.45elapsed

3) ADMB 11.2pre built from source (g++ 4.4.7):

BUILD:
$ /usr/bin/time ~/tmp/admb/admb -r euly
0.64user 0.03system 0:00.70elapsed
0.65user 0.04system 0:00.72elapsed
0.64user 0.05system 0:00.72elapsed

RUN:
$ /usr/bin/time ./euly
6.84user 1.22system 0:08.07elapsed
6.89user 1.20system 0:08.10elapsed
7.22user 1.20system 0:08.43elapsed

All three versions are comparable on build times, but 11.2 built from
source seems a bit faster in execution than 11.1 (Hooray!). 11.2pre from
binary is slowest.

System:
Linux 2.6.32-504.1.3.el6.centos.plus.x86_64
Intel Xeon E3-1240 v3 (3.40 GHz)
g++ (GCC) 4.4.7 20120313 (Red Hat 4.4.7-11)

  Tom Wainwright

On Wed, Dec 17, 2014 at 2:26 PM, Tom Wainwright <thomas.wainwright at noaa.gov>
wrote:
>
> Tested on my CentOS system:
>   Linux 2.6.32-504.1.3.el6.centos.plus.x86_64
>   g++ (GCC) 4.4.7 20120313 (Red Hat 4.4.7-11)
>
> 1) admb-11.2pre-centos6-intel2013-64bit.zip installed and cleanly built a
> few examples and recent projects.
>
> 2) I had a boring conference call this afternoon, so also did a source
> install. Built cleanly (except for a few type-conversion warnings), built
> and ran "simple" cleanly.
>
> Seems fine. Didn't test any new features.
>
>   Tom Wainwright
>
> On Fri, Dec 12, 2014 at 4:18 PM, Johnoel Ancheta <johnoel at hawaii.edu>
> wrote:
>>
>> Hi all,
>>
>> I need to get some feedback.  Can someone test out the
>> prerelease?
>>
>> http://www.admb-project.org/downloads/admb-11.2pre
>>
>> If there is no issues, the admb-11.2 will release next week.
>>
>> Johnoel
>>
>> _______________________________________________
>> Users mailing list
>> Users at admb-project.org
>> http://lists.admb-project.org/mailman/listinfo/users
>>
>>
>
> --
>
> Tom Wainwright
>
> Research Fisheries Biologist
>
> Newport Research Station
>
> Northwest Fisheries Science Center
>
> National Oceanic and Atmospheric Administration
>
> www.nwfsc.noaa.gov
> <http://www.nwfsc.noaa.gov/research/divisions/fed/climatechange.cfm>
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> 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

Research Fisheries Biologist

Newport Research Station

Northwest Fisheries Science Center

National Oceanic and Atmospheric Administration

www.nwfsc.noaa.gov
<http://www.nwfsc.noaa.gov/research/divisions/fed/climatechange.cfm>
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
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.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.admb-project.org/pipermail/users/attachments/20141217/c2bf5e01/attachment-0001.html>