<html>
<head>
<style>
.hmmessage P
{
margin:0px;
padding:0px
}
body.hmmessage
{
font-size: 10pt;
font-family:Verdana
}
</style>
</head>
<body class='hmmessage'>
Thanks, <br><br>If I use SYMPHONY solver to solve the same mILP models on the same hardware platform,<br>how much speedup I can get ?<br><br>For example, if I solve the same MILP model with one CPU, and then I solve it with <br>10 CPUs, how much the average speedup I can get ?<br><br>"Average" means that the average speedup of 10 or 20 MILP models.<br><br>Are there some papers that use SYMPHONY as a paralle solver and show some results ? <br><br>I need to know these results before I decide whether I need parallel SYMPHONY to solve <br>large milp models that may take more than 15 minutes.<br><br>thanks<br><br>Jack<br><br>April 22 2009 <br><br>> Date: Wed, 22 Apr 2009 10:16:48 +0200<br>> From: achterberg@zib.de<br>> To: dtustudy68@hotmail.com<br>> CC: coin-discuss@list.coin-or.org<br>> Subject: Re: [Coin-discuss] SYMPHONY speedup<br>> <br>> Hi Jack,<br>> <br>> for MILP, your question cannot be answered (LP is a different story). Parallel speedup and<br>> performance difference accross different solvers depend on the model.<br>> <br>> For example, there are models that can be solved by solver A in less than a second but<br>> that take forever by another solver B, just because solver B is lacking the important<br>> feature (like a cutting plane or a presolving reduction) to solve this particular model.<br>> And there may be other models for which the situation is reversed. Finally, there are<br>> models that can be solved in roughly the same time by all branch-and-cut MILP solvers.<br>> <br>> Overall, my experience is that the algorithmic quality of a solver is usually much more<br>> important than parallelization. This means, that often solvers like CBC, SCIP, Gurobi,<br>> XPress, or CPLEX will be faster than Symphony, even if you use 1000 CPUs for Symphony.<br>> But, as I said, some models do not require the sophisticated machinery that these solvers<br>> provide and parallelize well. In this setting, Symphony can be really great if you have<br>> access to a massively parallel computing environment.<br>> <br>> So, the short answer is: you just need to test it for your models. You cannot say in advance.<br>> <br>> <br>> Cheers,<br>> <br>> Tobias<br>> <br>> Jack Bryan wrote:<br>> > Hi ,<br>> > <br>> > I am a new user of SYMPHONY .<br>> > <br>> > I need to use it to solve very large mixed integer linear programming <br>> > models.<br>> > <br>> > So, I need to use PVM for SYMPHONY .<br>> > <br>> > But, I do not know how much speed up I can get from the parallel SYMPHONY ?<br>> > <br>> > For example, if it takes 100 seconds to solve a MILP model in CPLEX, how<br>> > long to solve it in SYMPHONY ?<br>> > <br>> > <br>> > thanks<br>> > <br>> > Jack<br>> > <br>> > April 21 2009<br>> > <br>> > ------------------------------------------------------------------------<br>> > Windows Live™ Hotmail®:…more than just e-mail. Check it out.<br>> > <http://windowslive.com/online/hotmail?ocid=TXT_TAGLM_WL_HM_more_042009><br>> > <br>> > <br>> > ------------------------------------------------------------------------<br>> > <br>> > _______________________________________________<br>> > Coin-discuss mailing list<br>> > Coin-discuss@list.coin-or.org<br>> > http://list.coin-or.org/mailman/listinfo/coin-discuss<br><br /><hr />Windows Live™ Hotmail®:…more than just e-mail. <a href='http://windowslive.com/online/hotmail?ocid=TXT_TAGLM_WL_HM_more_042009' target='_new'>Check it out.</a></body>
</html>