|Re: [eigen] Fwd: Re: Student contribution|
[ Thread Index |
| More lists.tuxfamily.org/eigen Archives
- To: eigen <eigen@xxxxxxxxxxxxxxxxxxx>
- Subject: Re: [eigen] Fwd: Re: Student contribution
- From: Gael Guennebaud <gael.guennebaud@xxxxxxxxx>
- Date: Mon, 17 Jun 2013 13:24:32 +0200
- Cc: "simon.courtemanche" <simon.courtemanche@xxxxxxxx>, Pierre ZOPPITELLI <Pierre.Zoppitelli@xxxxxxxxxxxxxxx>, Nicolas Carré <carren@xxxxxxxxxxxxxx>, Jean CECCATO <Jean.Ceccato@xxxxxxxxxxxxxxx>, "matthieu.moy" <Matthieu.Moy@xxxxxxxxxxxxxxx>
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:in-reply-to:references:from:date:message-id:subject:to :cc:content-type; bh=SL7319c9lNBlG5GspUIiZnxTgDwlK8KhVaKgcVq1PxA=; b=q9QwKy0SWhRBOZ73bMsQl1tN/OnZxy3AZ8vNa0/MDAfzGtxugUYewfp0cPFzVwP4Sz sOHAiOxVWSAbLS+LNAU1IKkJvWcmUzZ/E+WSUzK4hRJ8YLAU8RM2CmwZXGI6pkVOu1Nk gNr3vQ015Jvv/nmUrwVZ31B3DyJLLZW7xXfwRg1OZVfVA0HcyFpYuT/BMaPpqQzAgsxg BEIHqi/emBccxmJquW9z/Nsu7cEWEPtRTeYeXHiR5+scYcaPtZgcMnBKhxPaKoQSFz/y ibsuJr5X+ai0sBIyOsjvf4BwJHSC3KNwG2Amn53k15JW8nyU7RFQcjJQN/8N0Mk68Yim pcvg==
The most important is that your code can be continued by someone else.
Given the high interest in having such a feature in Eigen, I've no
doubt that someone will step up.
In particular, it would be great if you could provide as much details
as possible on the status of the "special SVD" code and what remains
to be done to get it working. Maybe you also have some ideas about
what should/could be done to improve the current implementation with
respect to speed, accuracy, or even flexibility of use.
This bug entry is the right place to put these general information:
thank you again for the hard work.
On Mon, Jun 17, 2013 at 12:48 PM, BRUN Gauthier <brunga@xxxxxxxxxxxxxx> wrote:
> We are proud to announce that thanks to Gael's fix, our algorithm now works
> fine with any dynamic matrices, including complexes one. It passes the same
> 'is_Approx' tests as Jacobi and that is not surprising since the loss of
> precision is caused by the deflation, actually functional but commented, and
> the third part of the algorithm that is yet to be implemented. We are now
> focusing on optimizing the existing code, cleaning some of the functions,
> and making the work of further developers easier. That includes sources for
> the next step, especially a paper to solve the Particle simulation problem
> and code comment to explain where to add the last part of the algorithm.
> Do you have any particular thing you want us to change in the BDCSVD ?
> Best regards,
> Team GL27
> Jean Ceccato, Pierre Zoppitelli, Gauthier Brun, Nicolas Carre