<html><head></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; ">Perhaps another option is blaze-lib. It has a nice license.<div><br></div><div><a href="http://code.google.com/p/blaze-lib/wiki/Benchmarks">http://code.google.com/p/blaze-lib/wiki/Benchmarks</a></div><div><br></div><div><br><div apple-content-edited="true">
<div><div><br></div></div>
</div>
<br><div><div>On Aug 5, 2014, at 3:00 PM, <a href="mailto:developers-request@admb-project.org">developers-request@admb-project.org</a> wrote:</div><br class="Apple-interchange-newline"><blockquote type="cite"><div>Send Developers mailing list submissions to<br><span class="Apple-tab-span" style="white-space:pre"> </span><a href="mailto:developers@admb-project.org">developers@admb-project.org</a><br><br>To subscribe or unsubscribe via the World Wide Web, visit<br><span class="Apple-tab-span" style="white-space:pre"> </span>http://lists.admb-project.org/mailman/listinfo/developers<br>or, via email, send a message with subject or body 'help' to<br><span class="Apple-tab-span" style="white-space:pre"> </span>developers-request@admb-project.org<br><br>You can reach the person managing the list at<br><span class="Apple-tab-span" style="white-space:pre"> </span>developers-owner@admb-project.org<br><br>When replying, please edit your Subject line so it is more specific<br>than "Re: Contents of Developers digest..."<br><br><br>Today's Topics:<br><br> 1. Re: interfacing with openblas for large problems (dave fournier)<br><br><br>----------------------------------------------------------------------<br><br>Message: 1<br>Date: Mon, 04 Aug 2014 15:21:41 -0700<br>From: dave fournier <davef@otter-rsch.com><br>To: Jim Ianelli - NOAA Federal <jim.ianelli@noaa.gov><br>Cc: "developers@admb-project.org" <developers@admb-project.org><br>Subject: Re: [Developers] interfacing with openblas for large problems<br>Message-ID: <53E00775.9020609@otter-rsch.com><br>Content-Type: text/plain; charset="utf-8"; Format="flowed"<br><br>On 08/04/2014 02:47 PM, Jim Ianelli - NOAA Federal wrote:<br><br>I found that there is a higher level lapack routine which works just <br>like the ADMB solve.<br><br>You can see how simple the interface with ADMB becomes.<br>In linux all you need is an include file and to link with two libraries.<br><br><br><blockquote type="cite">Seems like a worthwhile enterprise.<br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite">Looking at the site there are a number of interesting aspects of <br></blockquote><blockquote type="cite">OpenBlas that could help with ideas for the Foundation, not least of <br></blockquote><blockquote type="cite">which is their balance sheet and funding model:<br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite">https://docs.google.com/spreadsheet/ccc?key=0AghkTjXe2lDndE1UZml0dGpaUzJmZGhvenBZd1F2R1E&usp=sharing#gid=0<br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite">also hadn't heard of:<br></blockquote><blockquote type="cite">https://www.bountysource.com/faq<br></blockquote><blockquote type="cite">before, looks like something of interest.<br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite">also an aside but a useful git-branching model...: <br></blockquote><blockquote type="cite">http://nvie.com/files/Git-branching-model.pdf<br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite">so back to the question, how can we as a community help with the <br></blockquote><blockquote type="cite">interface work? I suppose steps might be something like:<br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite"> 1. pick some example function sets<br></blockquote><blockquote type="cite"> 2. map the variable types appropriately (so they will work with ADMB<br></blockquote><blockquote type="cite"> 3. write adjoint code for dvar functions (?)<br></blockquote><blockquote type="cite"> 4. test against current ADMB implmentations<br></blockquote><blockquote type="cite"> 5. ...<br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite">thanks for the interest dave.<br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite">jim<br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite">On Mon, Aug 4, 2014 at 12:31 PM, dave fournier <davef@otter-rsch.com <br></blockquote><blockquote type="cite"><mailto:davef@otter-rsch.com>> wrote:<br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite"> I wonder if there is any interest in doing a bit of work to<br></blockquote><blockquote type="cite"> interface some of the common<br></blockquote><blockquote type="cite"> function like ludecomp and choleski decomp with the openblas versions.<br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite"> this is not too difficult once one figures out the boring stuff<br></blockquote><blockquote type="cite"> like the fact that matrices are strored<br></blockquote><blockquote type="cite"> by column in contiguous arrays.<br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite"> we are now looking at ever larger problems. I compared the admb<br></blockquote><blockquote type="cite"> lu decomp of a 2,000x2,000 matrix<br></blockquote><blockquote type="cite"> from numerical recipes with the openblas version.<br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite"> Time for openblas lu decomp 172.05 ms<br></blockquote><blockquote type="cite"> Time for numerical recipes lu decomp 22102.9 ms<br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite"> which is about 125 times faster<br></blockquote><blockquote type="cite"> _______________________________________________<br></blockquote><blockquote type="cite"> Developers mailing list<br></blockquote><blockquote type="cite"> Developers@admb-project.org <mailto:Developers@admb-project.org><br></blockquote><blockquote type="cite"> http://lists.admb-project.org/mailman/listinfo/developers<br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite">-- <br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite">James Ianelli<br></blockquote><blockquote type="cite">NMFS/NOAA Building 4<br></blockquote><blockquote type="cite">7600 Sand Pt Way NE<br></blockquote><blockquote type="cite">Seattle WA 98115<br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite">206 526 6510<br></blockquote><blockquote type="cite"><br></blockquote><blockquote type="cite">Visit the ADMB project http://admb-project.org/<br></blockquote><blockquote type="cite"><br></blockquote><br>-------------- next part --------------<br>An HTML attachment was scrubbed...<br>URL: <http://lists.admb-project.org/pipermail/developers/attachments/20140804/cd263b5f/attachment-0001.html><br>-------------- next part --------------<br>A non-text attachment was scrubbed...<br>Name: newludecomp.cpp<br>Type: text/x-c++src<br>Size: 1476 bytes<br>Desc: not available<br>URL: <http://lists.admb-project.org/pipermail/developers/attachments/20140804/cd263b5f/attachment-0001.cpp><br><br>------------------------------<br><br>_______________________________________________<br>Developers mailing list<br>Developers@admb-project.org<br>http://lists.admb-project.org/mailman/listinfo/developers<br><br><br>End of Developers Digest, Vol 64, Issue 4<br>*****************************************<br></div></blockquote></div><br></div></body></html>