[Clp] high precision floating point numbers
Hok Go
Hok.Go at qrm.com
Tue May 12 11:59:28 EDT 2009
Sounds pretty cool! Do you think it'd be more useful to go the decomposition route instead to avoid seg-faults?
On a related note, have you ever noticed memory corruption on small (10 variable) QPs?
From: clp-bounces at list.coin-or.org [mailto:clp-bounces at list.coin-or.org] On Behalf Of Alexey Lvov
Sent: Tuesday, May 12, 2009 9:05 AM
To: clp at list.coin-or.org
Subject: [Clp] high precision floating point numbers
Is it possible to substitute "double" by a 128 bit floating point class in CLP?
I really admire how CLP can solve 50000 variable linear systems having just 15 decimal digits available (my own "brute force" simplex can not handle a system with more than 300 variables with the standard 64-bit double type).
But when it comes to 600,000 variables, handling it with just 15 decimal digits, I think, becomes impossible... I tried different combinations of tolerances in both primal and dual simplex of CLP but always got a segmentation fault after 24-48 hours.
If someone knows the answer, please let me know... If it is not implemented, I would volunteer to write a 128 bit floating point class (unless it already exists in COIN).... Of course it will be much slower than "double" but for someone who really needs a solution of a large problem and has a lot of patience this would help....
- Alexey
----------------------------------------------------
Dr. Alexey Lvov
IBM T.J. Watson Research Institute
1101 Kitchawan Rd. (Rt. 134) Office 34-159
Yorktown Heights NY 10598
Tel: (914)-945-3732
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/clp/attachments/20090512/48bd34ba/attachment.html>
More information about the Clp
mailing list