[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Bug-gnubg] Re: gnubg bug? Baseline Information
From: |
Christian Anthon |
Subject: |
Re: [Bug-gnubg] Re: gnubg bug? Baseline Information |
Date: |
Fri, 7 Aug 2009 14:25:34 +0200 |
Hi All,
I can now reproduce the bug on my office system. It is indeed
dependent on the prune and and compiler optimization settings. I
haven't found the root of the problem yet, but I can see that the
compiler optimized prune code plays a number of moves differently in
the 2ply eval.
Christian.
On Wed, Aug 5, 2009 at 9:01 PM, Michael Petch<address@hidden> wrote:
>
> To start things off properly and establish a baseline, I’m going to post my
> gnubgautorc (see attachment). The commands you can paste directly into gnubg
> at the command line to do the tests are:
>
> set lang C
> set evaluation chequerplay evaluation plies 2
> set evaluation chequerplay evaluation cubeful on
> set evaluation chequerplay evaluation prune on
> set evaluation cubedecision evaluation plies 2
> set evaluation cubedecision evaluation cubeful on
> set evaluation cubedecision evaluation prune on
> set gnubgid NwAAgN3MAGgBAA:cAnmAEAAIAAA
> hint 1
>
> set evaluation chequerplay evaluation prune off
> set evaluation cubedecision evaluation prune off
> set gnubgid NwAAgN3MAGgBAA:cAnmAEAAIAAA
> hint 1
>
> Please note, I discovered that if you toggle pruning on/off and then do Hint
> without setting the gnubgid then the position isn’t re-evaluated. Its
> important to so set gnubgid after changing the pruning parameters. If you
> use my gnubgautorc, if it fails loading the MET – just ignore the warning
> message (The MET doesn’t influence the outcome).
>
> Below is a successful test, and an unsuccessful one.
>
> ------------------------
> SUCCESSFUL TEST
> ------------------------
>
> With all that being said here is the expected output (This is on my Intel
> Mac OS/X 10.5 with –g -O2 ):
> mpetchs-intel-mac:Desktop mpetch$ gnubg -t
> GNU Backgammon 0.90.0 Aug 4 2009
> Copyright (C) 1999, 2000, 2001, 2002, 2003, 2004 by Gary Wong.
> Copyright (C) 2009 by Gary Wong and the AUTHORS; for details type `show
> version'.
> This program comes with ABSOLUTELY NO WARRANTY; for details type `show
> warranty'.
> This is free software, and you are welcome to redistribute it under certain
> conditions; type `show copying' for details.
> (No game) set lang C
> Locale is now 'C'(No game) set evaluation chequerplay evaluation plies 2
> `eval' and `hint' chequerplay will use 2 ply evaluation.
> (No game) set evaluation chequerplay evaluation cubeful on
> `eval' and `hint' chequerplay will use cubeful evaluation.
>
> (No game) set evaluation chequerplay evaluation prune on
> `eval' and `hint' chequerplay will use pruning.
>
> (No game) set evaluation cubedecision evaluation plies 2
> `eval' and `hint' cube decisions will use 2 ply evaluation.
> (No game) set evaluation cubedecision evaluation cubeful on
> `eval' and `hint' cube decisions will use cubeful evaluation.
>
> (No game) set evaluation cubedecision evaluation prune on
> `eval' and `hint' cube decisions will use pruning.
>
> (No game) set gnubgid NwAAgN3MAGgBAA:cAnmAEAAIAAA
> The dice have been set to 4 and 1.
> GNU Backgammon Position ID: 4HPwATDgc/ABMA
> Match ID : cAnmAEAAIAAA
> +13-14-15-16-17-18------19-20-21-22-23-24-+ O: gnubg
> | X O | | O X | 4 points
> | X O | | O X |
> | X O | | O |
> | X | | O |
> | X | | O |
> v| |BAR| | 7 point match (Cube: 1)
> | O | | X |
> | O | | X |
> | O X | | X |
> | O X | | X O | Rolled 41
> | O X | | X O | 4 points
> +12-11-10--9--8--7-------6--5--4--3--2--1-+ X: mpetch
>
> GNU Backgammon Position ID: NwAAgN3MAGgBAA
> Match ID : cAnmAEAAIAAA
> +13-14-15-16-17-18------19-20-21-22-23-24-+ O: gnubg
> | | | X X X O O | OO 4 points
> | | | X O O | OO
> | | | O | OO
> | | | | OO
> | | | | OO
> v| |BAR| | 7 point match (Cube: 1)
> | | | |
> | | | |
> | | | X |
> | X | | X X X X | Rolled 41
> | X | | X X X X | 4 points
> +12-11-10--9--8--7-------6--5--4--3--2--1-+ X: mpetch
>
> Setting GNUBG id NwAAgN3MAGgBAA:cAnmAEAAIAAA
> (mpetch ) hint 1
> 1. Cubeful 2-ply 21/17 19/18 Eq.: -2.729
>
> 0.000 0.000 0.000 - 1.000 1.000 0.085
> 2-ply cubeful prune [world class]
> (mpetch )
> (mpetch ) set evaluation chequerplay evaluation prune off
> `eval' and `hint' chequerplay will not use pruning.
>
> (mpetch ) set evaluation cubedecision evaluation prune off
> `eval' and `hint' cube decisions will not use pruning.
>
> (mpetch ) set gnubgid NwAAgN3MAGgBAA:cAnmAEAAIAAA
> The dice have been set to 4 and 1.
> GNU Backgammon Position ID: NwAAgN3MAGgBAA
> Match ID : cAnmAEAAIAAA
> +13-14-15-16-17-18------19-20-21-22-23-24-+ O: gnubg
> | | | X X X O O | OO 4 points
> | | | X O O | OO
> | | | O | OO
> | | | | OO
> | | | | OO
> v| |BAR| | 7 point match (Cube: 1)
> | | | |
> | | | |
> | | | X |
> | X | | X X X X | Rolled 41
> | X | | X X X X | 4 points
> +12-11-10--9--8--7-------6--5--4--3--2--1-+ X: mpetch
>
> GNU Backgammon Position ID: NwAAgN3MAGgBAA
> Match ID : cAnmAEAAIAAA
> +13-14-15-16-17-18------19-20-21-22-23-24-+ O: gnubg
> | | | X X X O O | OO 4 points
> | | | X O O | OO
> | | | O | OO
> | | | | OO
> | | | | OO
> v| |BAR| | 7 point match (Cube: 1)
> | | | |
> | | | |
> | | | X |
> | X | | X X X X | Rolled 41
> | X | | X X X X | 4 points
> +12-11-10--9--8--7-------6--5--4--3--2--1-+ X: mpetch
>
> GNU Backgammon Position ID: NwAAgN3MAGgBAA
> Match ID : cAnmAEAAIAAA
> +13-14-15-16-17-18------19-20-21-22-23-24-+ O: gnubg
> | | | X X X O O | OO 4 points
> | | | X O O | OO
> | | | O | OO
> | | | | OO
> | | | | OO
> v| |BAR| | 7 point match (Cube: 1)
> | | | |
> | | | |
> | | | X |
> | X | | X X X X | Rolled 41
> | X | | X X X X | 4 points
> +12-11-10--9--8--7-------6--5--4--3--2--1-+ X: mpetch
>
> Setting GNUBG id NwAAgN3MAGgBAA:cAnmAEAAIAAA
> (mpetch ) hint 1
> 1. Cubeful 2-ply 21/17 19/18 Eq.: -2.729
>
> 0.000 0.000 0.000 - 1.000 1.000 0.085
> 2-ply cubeful
> (mpetch )
>
> ----------------------------
> UNSUCCESSFUL TEST
> ----------------------------
>
>
> On a system (Debian Lenny Stable –32 Bit) using –g –O2 as the compiler
> options, we get this unusual output:
>
> GNU Backgammon 0.90.0 Aug 4 2009
> Copyright (C) 1999, 2000, 2001, 2002, 2003, 2004 by Gary Wong.
> Copyright (C) 2009 by Gary Wong and the AUTHORS; for details type `show
> version'.
> This program comes with ABSOLUTELY NO WARRANTY; for details type `show
> warranty'.
> This is free software, and you are welcome to redistribute it under certain
> conditions; type `show copying' for details.
>
> (No game) set lang C
> Locale is now 'C'(No game) set evaluation chequerplay evaluation plies 2
> `eval' and `hint' chequerplay will use 2 ply evaluation.
> (No game) set evaluation chequerplay evaluation cubeful on
> `eval' and `hint' chequerplay will use cubeful evaluation.
>
> (No game) set evaluation chequerplay evaluation prune on
> `eval' and `hint' chequerplay will use pruning.
>
> (No game) set evaluation cubedecision evaluation plies 2
> `eval' and `hint' cube decisions will use 2 ply evaluation.
> (No game) set evaluation cubedecision evaluation cubeful on
> `eval' and `hint' cube decisions will use cubeful evaluation.
>
> (No game) set evaluation cubedecision evaluation prune on
> `eval' and `hint' cube decisions will use pruning.
>
> (No game) set gnubgid NwAAgN3MAGgBAA:cAnmAEAAIAAA
> The dice have been set to 4 and 1.
> GNU Backgammon Position ID: 4HPwATDgc/ABMA
> Match ID : cAnmAEAAIAAA
> +13-14-15-16-17-18------19-20-21-22-23-24-+ O: gnubg
> | X O | | O X | 4 points
> | X O | | O X |
> | X O | | O |
> | X | | O |
> | X | | O |
> v| |BAR| | 7 point match (Cube: 1)
> | O | | X |
> | O | | X |
> | O X | | X |
> | O X | | X O | Rolled 41
> | O X | | X O | 4 points
> +12-11-10--9--8--7-------6--5--4--3--2--1-+ X: mpetch
>
> GNU Backgammon Position ID: NwAAgN3MAGgBAA
> Match ID : cAnmAEAAIAAA
> +13-14-15-16-17-18------19-20-21-22-23-24-+ O: gnubg
> | | | X X X O O | OO 4 points
> | | | X O O | OO
> | | | O | OO
> | | | | OO
> | | | | OO
> v| |BAR| | 7 point match (Cube: 1)
> | | | |
> | | | |
> | | | X |
> | X | | X X X X | Rolled 41
> | X | | X X X X | 4 points
> +12-11-10--9--8--7-------6--5--4--3--2--1-+ X: mpetch
>
> Setting GNUBG id NwAAgN3MAGgBAA:cAnmAEAAIAAA
> (mpetch ) hint 1
> 1. Cubeful 2-ply 20/16 19/18 Eq.: -2.828
>
> 0.000 0.000 0.000 - 1.000 1.000 0.124
> 2-ply cubeful prune [world class]
> (mpetch )
> (mpetch ) set evaluation chequerplay evaluation prune off
> `eval' and `hint' chequerplay will not use pruning.
>
> (mpetch ) set evaluation cubedecision evaluation prune off
> `eval' and `hint' cube decisions will not use pruning.
>
> (mpetch ) set gnubgid NwAAgN3MAGgBAA:cAnmAEAAIAAA
> The dice have been set to 4 and 1.
> GNU Backgammon Position ID: NwAAgN3MAGgBAA
> Match ID : cAnmAEAAIAAA
> +13-14-15-16-17-18------19-20-21-22-23-24-+ O: gnubg
> | | | X X X O O | OO 4 points
> | | | X O O | OO
> | | | O | OO
> | | | | OO
> | | | | OO
> v| |BAR| | 7 point match (Cube: 1)
> | | | |
> | | | |
> | | | X |
> | X | | X X X X | Rolled 41
> | X | | X X X X | 4 points
> +12-11-10--9--8--7-------6--5--4--3--2--1-+ X: mpetch
>
> GNU Backgammon Position ID: NwAAgN3MAGgBAA
> Match ID : cAnmAEAAIAAA
> +13-14-15-16-17-18------19-20-21-22-23-24-+ O: gnubg
> | | | X X X O O | OO 4 points
> | | | X O O | OO
> | | | O | OO
> | | | | OO
> | | | | OO
> v| |BAR| | 7 point match (Cube: 1)
> | | | |
> | | | |
> | | | X |
> | X | | X X X X | Rolled 41
> | X | | X X X X | 4 points
> +12-11-10--9--8--7-------6--5--4--3--2--1-+ X: mpetch
>
> GNU Backgammon Position ID: NwAAgN3MAGgBAA
> Match ID : cAnmAEAAIAAA
> +13-14-15-16-17-18------19-20-21-22-23-24-+ O: gnubg
> | | | X X X O O | OO 4 points
> | | | X O O | OO
> | | | O | OO
> | | | | OO
> | | | | OO
> v| |BAR| | 7 point match (Cube: 1)
> | | | |
> | | | |
> | | | X |
> | X | | X X X X | Rolled 41
> | X | | X X X X | 4 points
> +12-11-10--9--8--7-------6--5--4--3--2--1-+ X: mpetch
>
> Setting GNUBG id NwAAgN3MAGgBAA:cAnmAEAAIAAA
> (mpetch ) hint 1
> 1. Cubeful 2-ply 21/17 19/18 Eq.: -2.729
>
> 0.000 0.000 0.000 - 1.000 1.000 0.085
> 2-ply cubeful
> (mpetch )
>
>
>
> _______________________________________________
> Bug-gnubg mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/bug-gnubg
>
>
- Re: [Bug-gnubg] Re: Getting gnubg to use all available cores, (continued)
- [Bug-gnubg] Re: Getting gnubg to use all available cores, Louis Zulli, 2009/08/06
- RE: [Bug-gnubg] Re: Getting gnubg to use all available cores, Ingo Macherius, 2009/08/06
- Re: [Bug-gnubg] Re: Getting gnubg to use all available cores, Louis Zulli, 2009/08/06
- RE: [Bug-gnubg] Re: Getting gnubg to use all available cores, Ingo Macherius, 2009/08/06
- Re: [Bug-gnubg] Re: Getting gnubg to use all available cores, Michael Petch, 2009/08/07
- Re: [Bug-gnubg] Re: Getting gnubg to use all available cores, Christian Anthon, 2009/08/07
- Re: [Bug-gnubg] Re: Getting gnubg to use all available cores, Michael Petch, 2009/08/07
- Re: [Bug-gnubg] Re: Getting gnubg to use all available cores, Louis Zulli, 2009/08/07
- Re: [Bug-gnubg] Re: gnubg bug? Baseline Information,
Christian Anthon <=
- Re: [Bug-gnubg] Re: gnubg bug? Baseline Information, Christian Anthon, 2009/08/08
- Re: [Bug-gnubg] Re: gnubg bug? Baseline Information, Philippe Michel, 2009/08/08
- Re: [Bug-gnubg] Re: gnubg bug? Baseline Information, Michael Petch, 2009/08/08
- Re: [Bug-gnubg] Re: gnubg bug? Baseline Information, Michael Petch, 2009/08/08
- Re : [Bug-gnubg] Re: gnubg bug? Baseline Information, Massimiliano Maini, 2009/08/09
- Re: [Bug-gnubg] Re: gnubg bug? Baseline Information, Christian Anthon, 2009/08/11