<html><body>
<p>Alexey,<br>
<br>
I would not think that 128 bits versus 64 bits is the problem if you get a segmentation fault. In fact it is more likely that it is a memory problem, which would get worse with 128 bits. How many constraints and nonzeros does the problem have?<br>
<br>
If you put an example somewhere (easiest is some Watson machine) and tell me where it is (and how big), I can download the problem and see what can be done.<br>
<br>
John Forrest<br>
<br>
<img width="16" height="16" src="cid:1__=0ABBFF27DFC5BEDC8f9e8a93df938@us.ibm.com" border="0" alt="Inactive hide details for Alexey Lvov---05/12/2009 10:09:45 AM---Is it possible to substitute "double" by a 128 bit floating po"><font color="#424282">Alexey Lvov---05/12/2009 10:09:45 AM---Is it possible to substitute "double" by a 128 bit floating point class in CLP?</font><br>
<br>
<table width="100%" border="0" cellspacing="0" cellpadding="0">
<tr valign="top"><td width="1%"><img width="96" height="1" src="cid:2__=0ABBFF27DFC5BEDC8f9e8a93df938@us.ibm.com" border="0" alt=""><br>
<font size="2" color="#5F5F5F">From:</font></td><td width="100%"><img width="1" height="1" src="cid:2__=0ABBFF27DFC5BEDC8f9e8a93df938@us.ibm.com" border="0" alt=""><br>
<font size="2">Alexey Lvov/Watson/IBM@IBMUS</font></td></tr>
<tr valign="top"><td width="1%"><img width="96" height="1" src="cid:2__=0ABBFF27DFC5BEDC8f9e8a93df938@us.ibm.com" border="0" alt=""><br>
<font size="2" color="#5F5F5F">To:</font></td><td width="100%"><img width="1" height="1" src="cid:2__=0ABBFF27DFC5BEDC8f9e8a93df938@us.ibm.com" border="0" alt=""><br>
<font size="2">clp@list.coin-or.org</font></td></tr>
<tr valign="top"><td width="1%"><img width="96" height="1" src="cid:2__=0ABBFF27DFC5BEDC8f9e8a93df938@us.ibm.com" border="0" alt=""><br>
<font size="2" color="#5F5F5F">Date:</font></td><td width="100%"><img width="1" height="1" src="cid:2__=0ABBFF27DFC5BEDC8f9e8a93df938@us.ibm.com" border="0" alt=""><br>
<font size="2">05/12/2009 10:09 AM</font></td></tr>
<tr valign="top"><td width="1%"><img width="96" height="1" src="cid:2__=0ABBFF27DFC5BEDC8f9e8a93df938@us.ibm.com" border="0" alt=""><br>
<font size="2" color="#5F5F5F">Subject:</font></td><td width="100%"><img width="1" height="1" src="cid:2__=0ABBFF27DFC5BEDC8f9e8a93df938@us.ibm.com" border="0" alt=""><br>
<font size="2">[Clp] high precision floating point numbers</font></td></tr>
</table>
<hr width="100%" size="2" align="left" noshade style="color:#8091A5; "><br>
<br>
<br>
<b><font size="5">Is it possible to substitute "double" by a 128 bit floating point class in CLP?</font></b><font size="4"> <br>
<br>
<br>
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). <br>
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. <br>
<br>
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....<br>
<br>
- Alexey<br>
<br>
----------------------------------------------------<br>
Dr. Alexey Lvov<br>
IBM T.J. Watson Research Institute<br>
1101 Kitchawan Rd. (Rt. 134) Office 34-159<br>
Yorktown Heights NY 10598<br>
Tel: (914)-945-3732</font><tt>_______________________________________________<br>
Clp mailing list<br>
Clp@list.coin-or.org<br>
</tt><tt><a href="http://list.coin-or.org/mailman/listinfo/clp">http://list.coin-or.org/mailman/listinfo/clp</a></tt><tt><br>
</tt><br>
<br>
</body></html>