[Clp] Clp - Under new management

[armen nahapetyan]

Hello Mr Julian Hall,

We are pleased to see you on this position.

Your ideas about Clp acceleration on GPGPU's seem challenging and interesting.
But lack of documentation makes it very difficult and slow to be
integrated into the development process. And the source code itself is
hard to understand.

I have got experience of parallel systems implementation and GPGPU programming and I would like join to CLP acceleration.
I used to try to do some experiments by myself, but the lack of documentation describing mathematical details and source code organization did not allow to make CLP to run really in parallel, I implemented on GPGPU just some low-level functions with matrix factorizations.

I have got the documentation which is refereed in the Clp source code:
 
     Forrest and Tomlin, Implementing the simplex method for
       the Optimization Subroutine Library - IBM Systems Journal 1992

     Gill, Murray, Saunders, Wright A Practical Anti-Cycling
       Procedure for Linear and Nonlinear Programming SOL report 1988

But that documentation mainly  discourse numerical issues and corner cases.

So if some development specification or other document about the source code organization would be available, I believe many strong guys would be interested to become a part of  Clp acceleration process.

Best Regards,
Armen Nahapetyan



________________________________
From: Julian Hall <J.A.J.Hall at ed.ac.uk>
To: clp at list.coin-or.org
Sent: Sat, February 6, 2010 3:27:57 AM
Subject: [Clp] Clp - Under new management

Dear All,

As John announced earlier in the week, I have taken over as Project
Manager for Clp. This is an exciting opportunity to contribute to the  
premier open-source LP solver.

First of all I'd like to express my appreciation for all that John has  
contributed to computational linear optimization in general, and  
COIN-OR in particular.

Taking over a necessarily large and complex code such as Clp is a  
significant challenge and, until I get a reasonable mastery of it, I  
cannot pretend that (any) bugs will be fixed with John's alacrity. In  
this, and other aspects of Clp management, I will be aided by Lou  
Hafer and Matthew Saltzman.

As for code development, in the short to medium term, my aim will be  
to work with Ted Ralphs and the Cbc team to achieve a clear  
distinction between Cbc and Clp. This will allow the two projects to  
be managed with greater independence, but with no loss in performance.

Beyond that, given that I have my own simplex solver that is  
competitive with Clp, there may well be some computational techniques  
that I can introduce to Clp to improve its performance.

In the longer term, I would expect to develop Clp so that it can  
exploit multi-core and GPGPU architectures - but don't expect linear  
speed-up!

These are merely my ideas: I'll be keen to hear from users of  
developments that they would like to see.

Since I'm new to the COIN community (and am doubtless unknown to most
recipients of this message) let me close by introducing myself!

Having been in Edinburgh since 1990, when I completed my PhD at Dundee  
under Roger Fletcher, I've worked almost exclusively on computational  
techniques for
the simplex method. I have developed my own high performance serial  
industrial-strength simplex solver that has been the basis of  
consultancy work and underpinned some very labour-intensive work on  
parallel simplex. These are some of the reasons why I haven't  
published much---although I did win the 2005 COAP prize for a paper on  
"hyper-sparsity". Note that the Edinburgh group also includes  
computational IPM experts Jacek Gondzio and Andreas Grothey, so we  
cover high performance solution of LP/QP problems pretty well!

Have a good weekend,

Julian Hall
--
Dr. J. A. Julian Hall, Senior Lecturer, School of Mathematics,
University of Edinburgh, JCMB, King's Buildings, EDINBURGH, EH9 3JZ, UK.
Room: 6221   Phone: [+44](131) 650 5075   Email: J.A.J.Hall at ed.ac.uk
Fax: [+44](131) 650 6553   Web: http://www.maths.ed.ac.uk/hall

The University of Edinburgh is a charitable body, registered in
Scotland, with
registration number SC005336.






-- 
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.



_______________________________________________
Clp mailing list
Clp at list.coin-or.org
http://list.coin-or.org/mailman/listinfo/clp



      
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/clp/attachments/20100209/9e26eb2b/attachment.html>