Re: [rules-users] Problems Extending the Golfer example - Cartesian joins ??

Hi Mark, Any news on this one? I know its just a puzzle, but I would have hoped drools would cope with it. I was hoping to use the golfer example as a training aid to people in work. Regards Simon On 09/08/07, *Simon French* <french.simon(a)gmail.com <mailto:french.simon@gmail.com>> wrote: Hi Mark, I gave the latest version a go but still having the same problem. It appears Drools doesn't like it when the last name is asserted into working memory. The memory usage at this stage is just under 900meg. After that drools gradually eats up all the available memory but no more assertions are made. Regards Simon On 03/08/07, *Mark Proctor * <mproctor(a)codehaus.org <mailto:mproctor@codehaus.org>> wrote: I must admit I didn't check against this example. But we did find a memory leak. Each activation remembers its parent activation, this chain was never getting nulled, which meant for recursive algorithms it went out of memory very quickly - this is now fixed. Maybe you can check trunk and make sure that fixes our problem? Trunk is broken at the moment, while we do MVEL updates, but should be fixed in a few hours. I'll message the list when it is - you won't need to build anything yourself as you will then be able to get the latest snapshot to try from here http://cruisecontrol.jboss.com/cc/artifacts/jboss-rules. Mark Simon French wrote: > Hi Mark, > After all the hard work you guys have put into drools > recently I hope you had a good holiday. > Has anybody had a chance to look at this? > > Simon > > On 20/07/07, * Mark Proctor* <mproctor(a)codehaus.org > <mailto:mproctor@codehaus.org>> wrote: > > we'll look into this to see where the memory is going. > but won't be for another week or so - I'm currently on > holiday. > > Mark > Simon French wrote: >> Hi, >> After attending a session with Mark Proctor in London I >> was inspired to play around with Drools in a slightly >> different way than we use it for in the workplace, and >> try and solve a puzzle. >> Unfortunately I couldn't get Drools to solve it as I >> kept getting memory problems (I believe are cartesian >> joins being the problem), >> so I decided to extend the golfer example to see if I >> got the same problem >> I added two new names and a new integer property, club >> >> // create all possible Golfer objects >> String[] names = new String[] { "Fred", >> "Joe", "Bob", "Tom", "Des", "Terry" }; >> String[] colors = new String[] { "red", >> "blue", "plaid", "orange","black", "white" }; >> int[] positions = new int[] { 1, 2, 3, >> 4, 5, 6 }; >> int[] clubs = new int[] {9,8,7,4,5,5}; >> >> for ( int n = 0; n < names.length; n++ ) { >> for ( int c = 0; c < colors.length; >> c++ ) { >> for ( int p = 0; p < >> positions.length; p++ ) { >> for ( int q = 0; q < >> clubs.length; q++ ) { >> session.insert(new >> Golfer( names[n], colors[c], positions[p], clubs[q]) ); >> >> .... >> >> >> Then changed the golder.drl to: >> >> package com.sample >> >> import com.sample.Golfer; >> >> >> rule "Golfer Riddle" >> when >> // A golfer named Fred, >> >> Golfer( name == "Fred", >> $fredsPosition : position, $fredsColor : >> color, $fredsClub : club ) >> >> // Der Golfer hinter Fred trägt blau >> Golfer( $unknownsName : name != "Fred", >> $unknownsPosition : position == ( >> $fredsPosition + 1 ), >> $unknownsColor : color == "blue", color >> != $fredsColor, >> $unknownsClub : club == 5 ) >> >> // Joe steht an zweiter Stelle >> Golfer( name == "Joe", $joesPosition : position >> == 2, >> position != $fredsPosition, >> $joesColor : color != $fredsColor, >> $joesClub : club == 5 ) >> >> // Bob traegt Karo >> Golfer( name == "Bob", >> name != $unknownsName, >> $bobsPosition : position != $fredsPosition, >> position != $unknownsPosition, position >> != >> $joesPosition, >> >> $bobsColor : color == "plaid", >> color != $fredsColor, color != $joesColor, >> color != $unknownsColor, >> $bobsClub : club < $joesClub ) >> >> // Tom ist nicht 1. oder 4., traegt kein Orange >> Golfer( $tomsName : name == "Tom", >> $tomsPosition : position != 1, position >> != 4, >> position != $fredsPosition, position != >> $joesPosition, >> position != >> $bobsPosition, >> $tomsColor : color != "orange", color != >> "blue", >> color != $fredsColor, color != $joesColor, >> color != $bobsColor ) >> >> Golfer ( $des : name == "Des", $desPosition : >> position < $fredsPosition, $desColor : color != "blue", >> color != $fredsColor, color != $joesColor, >> color != $bobsColor ) >> Golfer ( $terry : name == "Terry", >> $terryPosition : position < $desPosition, $terryColor : >> color != "blue", >> color != $fredsColor, color != $joesColor, >> color != $bobsColor ) >> then >> System.out.println( "Fred " + $fredsPosition + " >> " + $fredsColor ); >> System.out.println( "Joe " + $joesPosition + " " >> + $joesColor ); >> System.out.println( "Bob " + $bobsPosition + " " >> + $bobsColor ); >> System.out.println( "Tom " + $tomsPosition + " " >> + $tomsColor ); >> System.out.println ( "Des " + $desPosition + " " >> + $desColor ); >> System.out.println( "terry " + $terryPosition + >> " " + $terryColor ); >> end >> >> >> After a couple of minutes of processing I got:- >> >> Exception in thread "main" java.lang.OutOfMemoryError: >> Java heap space >> >> I also changed java memory allocation with -Xms1640M >> -Xmx1640M (The most I can allocate on my PC), and tried >> running it as a stateless session too. >> >> Is there anything else I can do, or is this simply >> pushing drools too far (surely not!!) ? >> >> I'm very keen on pushing this technology as much as >> possible in our workplace and I'd really appreciate any >> help. >> >> Thanks in advance >> >> Simon >> >> ------------------------------------------------------------------------ >> >> _______________________________________________ >> rules-users mailing list >> rules-users(a)lists.jboss.org >> >> >> <mailto:rules-users@lists.jboss.org> >> https://lists.jboss.org/mailman/listinfo/rules-users >> > > > _______________________________________________ > rules-users mailing list > rules-users(a)lists.jboss.org > <mailto:rules-users@lists.jboss.org> > https://lists.jboss.org/mailman/listinfo/rules-users > > > ------------------------------------------------------------------------ > > _______________________________________________ > rules-users mailing list > rules-users(a)lists.jboss.org > > <mailto:rules-users@lists.jboss.org> > https://lists.jboss.org/mailman/listinfo/rules-users > _______________________________________________ rules-users mailing list rules-users(a)lists.jboss.org <mailto:rules-users@lists.jboss.org> https://lists.jboss.org/mailman/listinfo/rules-users ------------------------------------------------------------------------ _______________________________________________ rules-users mailing list rules-users(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users