[Clp] CLP with warmstart running too long
John Forrest
john.forrest at fastercoin.com
Thu May 29 03:16:38 EDT 2014
Robert,
Obviously something is wrong. If you are getting output then cycling,
if no output then some data may have got corrupted and in a infinite
"while" loop or similar.
Difficult for me to debug as unlikely to be able to get exact compiler
results. I might be able to if your system Linux.
If you can attach a debugger then you could see where it is and maybe
set handler_->logLevel_=63 to get more output.
John Forrest
On 26/05/14 22:10, Lacroix, Robert wrote:
>
> I am running around 5,000 instances of an LP problem of size 66112
> rows by 104643 columns programmed through the OSI interface with
> approximately normal distribution of CPU time peaking around 3
> minutes, with about 1000 cases taking between 10 and 36 minutes. They
> all finished except one, which is still running after 67 hours. These
> runs are all provided a warmstart that conditions the solution.
> Without a warmstart, this exceptional case completes in under 4
> minutes. Is it possible that CLP is cycling and will never find a
> solution?
>
>
> Robert Lacroix
>
> Manitoba Hydro
>
>
>
> _______________________________________________
> 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/20140529/ae934b4b/attachment.html>
More information about the Clp
mailing list