RE: [Help-glpk] GLPK re-endrant

[D'Agostino, Larry - TX]

-----Original Message-----
From: help-glpk-bounces+larry.d'address@hidden 
[mailto:help-glpk-bounces+larry.d'address@hidden On Behalf Of Rios, Joseph L. 
(ARC-AFO)
Sent: Thursday, July 02, 2009 11:41 AM
To: address@hidden
Subject: Re: [Help-glpk] GLPK re-endrant

>The reentrancy issue could be obviated by
>running GLPK instances in separate processes.

That's certainly true, but inter-process communication is different than
inter-thread communication.  Being able to use the pthread machinery is
nice.

Also, just to clarify on the patch I posted earlier, I wasn't trying to
imply that I was actually "fixing" glpk.  It really is just a hack if you
REALLY need multithreaded glpk to work.  Hope I didn't claim any more than
that.

Joey




Joey,

I apologize if I'm not up to speed on multi-threaded and re-entrant coding.  
What is the purpose you are trying to accomplish with your patch?  Is it just 
the ability to multithread?  Or is there a possible benchmark improvement to 
GLPK?

The reason I ask is this is something I've been thinking about lately.  I look 
at Hans Mittelmann's benchmark tests and wonder why GLPK can not match the 
performance of the other solvers.  By no means is GLPK inefficient.  Just 
wondering what is lacking to make it on par with CPELX or whatnot.

It would be nice to include benchmarks with your patches Joey.  I know I would 
be interested.

Larry

_______________________________________________
Help-glpk mailing list
address@hidden
http://lists.gnu.org/mailman/listinfo/help-glpk