[Developers] Revert of ludcmp to version from ADMB-9
davef at otter-rsch.com
Tue Sep 4 22:28:58 PDT 2012
On 12-08-31 09:14 AM, Mollie Brooks wrote:
Well it is Dave's ludcmp function and only partially Dave's solve function.
The latter only appeared because of open source crap.
> Holy Cow!
> Running my pond model with Dave's solve function and version 10 code
> took 2 hours and 15 minutes. With the code in this branch
> (reverted solve function and reverted ludcmp) my model only took 35
> minutes! I checked and the results are the same. I'm glad this change
> happened before I submit it for publication.
> Mollie Brooks
> Postdoctoral Researcher, Ponciano Lab
> Biology Department, University of Florida
> On 30 Aug 2012, at 6:19 PM, Johnoel Ancheta wrote:
>> Hi all,
>> A branch has been created to revert LU Decomposition routines to code
>> from ADMB-9.
>> The functions affected are the internal solve and the det code.
>> Please checkout
>> the branch and test with your existing tpl files. Compare speed and
>> the .par outputs
>> with the currents ADMB-11 release.
>> $ svn http://www.admb-project.org/svn/branches/revert-ludcmp
>> Thanks in advance for providing any feedback,
>> Developers mailing list
>> Developers at admb-project.org <mailto:Developers at admb-project.org>
> Developers mailing list
> Developers at admb-project.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Developers