Discussion:
Stupidest game of the month. 11 move Mate.
(too old to reply)
Sanny
2008-07-12 11:21:40 UTC
Permalink
I played Beginner. I played Handicap Level and got a Knight Extra.

At GetClub with Handicap Level you can turn any level with a Knight/
Bishop/ Rook/ Queen Extra.

I thought that beginer without a Knight will be easy for me to win.

But when I played the game Beginner gave me Check Mate even without a
Knight.

And that too in just 11 Moves.

Rember I was having a Knight Extra. Still Beginner was able to win the
game.

Game Played between sanjay11 and beginner_x_N at GetClub.com

--------------------------------------------------------------------------------
sanjay11: (Black)
beginner_x_N: (White)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM22495&game=Chess
--------------------------------------------------------------------------------

White -- Black
(beginner_x_N) -- (sanjay11)

1. e2-e4{0} e7-e5{10}
2. Nb1-c3{16} Nb8-c6{4}
3. Bf1-c4{8} d7-d6{4}
4. Qd1-h5{12} g7-g6{14}
5. Qh5-f3{6} Ng8-f6{8}
6. Nc3-b5{10} a7-a6{4}
7. Nb5-c7{6} Qd8-c7{8}
8. Qf3-f6{8} Rh8-g8{12}
9. c2-c3{6} Bf8-e7{8}
10. Qf6-f7{6} Ke8-d7{16}
11. Bc4-e6{2}
--------------------------------------------------------------------------------
sanjay11: (Black)
beginner_x_N: (White)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM22495&game=Chess

So you can see even with a Knight down Beginner is tough on me.

Bye
Sanny

Play Chess at: http://www.GetClub.com/Chess.html
tOmmetje
2008-07-12 13:30:12 UTC
Permalink
Post by Sanny
I played Beginner. I played Handicap Level and got a Knight Extra.
At GetClub with Handicap Level you can turn any level with a Knight/
Bishop/ Rook/ Queen Extra.
I thought that beginer without a Knight will be easy for me to win.
But when I played the game Beginner gave me Check Mate even without a
Knight.
And that too in just 11 Moves.
Rember I was having a Knight Extra. Still Beginner was able to win the
game.
Game Played between sanjay11 and beginner_x_N at GetClub.com
--------------------------------------------------------------------------------
sanjay11: (Black)
beginner_x_N: (White)
Game Played at:http://www.GetClub.com/Chess.html
View Recorded Game:http://www.getclub.com/playgame.php?id=DM22495&game=Chess
--------------------------------------------------------------------------------
White -- Black
(beginner_x_N) -- (sanjay11)
1. e2-e4{0} e7-e5{10}
2. Nb1-c3{16} Nb8-c6{4}
3. Bf1-c4{8} d7-d6{4}
4. Qd1-h5{12} g7-g6{14}
5. Qh5-f3{6} Ng8-f6{8}
6. Nc3-b5{10} a7-a6{4}
7. Nb5-c7{6} Qd8-c7{8}
8. Qf3-f6{8} Rh8-g8{12}
9. c2-c3{6} Bf8-e7{8}
10. Qf6-f7{6} Ke8-d7{16}
11. Bc4-e6{2}
--------------------------------------------------------------------------------
sanjay11: (Black)
beginner_x_N: (White)
Game Played at:http://www.GetClub.com/Chess.html
View Recorded Game:http://www.getclub.com/playgame.php?id=DM22495&game=Chess
So you can see even with a Knight down Beginner is tough on me.
Bye
Sanny
Play Chess at:http://www.GetClub.com/Chess.html
Well DUH, of course Beginner easily wins from you even when down a
knight. Your chess skills are even worse than your programming skills.
Sanny
2008-07-12 15:48:34 UTC
Permalink
Post by tOmmetje
Well DUH, of course Beginner easily wins from you even when down a
knight. Yourchessskills are even worse than your programming skills.- Hide quoted text -
There is one Baby level which is even weak than Beginner. I am unable
to win the Baby Level also.

Now, I will try Baby Level with Extra Knight. If I fail I will get
Extra Rook against Baby Level.

Once can take Extra Queen/ Rook/ Bishop / Knight with any of the
levels. Still they are very strong.

Bye
Sanny

Play Chess at:http://www.GetClub.com/Chess.html
Tobias Heidelmann
2008-07-12 16:11:26 UTC
Permalink
Post by Sanny
Post by tOmmetje
Well DUH, of course Beginner easily wins from you even when down a
knight. Yourchessskills are even worse than your programming skills.- Hide quoted text -
There is one Baby level which is even weak than Beginner. I am unable
to win the Baby Level also.
Now, I will try Baby Level with Extra Knight. If I fail I will get
Extra Rook against Baby Level.
Once can take Extra Queen/ Rook/ Bishop / Knight with any of the
levels. Still they are very strong.
Bye
Sanny
Play Chess at:http://www.GetClub.com/Chess.html
Dude, are you serious?
this makes me very curious.. i mean, you openly admit that you have
virtually no understanding of chess whatsoever and YET you are writing
some chess software? how do you do that?
your software must use some algorithm to evaluate a given position. As
you cannot do that yourself,how do you teach a computer? do you use
other existing algorithms?

Tobias
Sanny
2008-07-12 16:47:33 UTC
Permalink
Post by Tobias Heidelmann
Dude, are you serious?
this makes me very curious.. i mean, you openly admit that you have
virtually no understanding ofchesswhatsoever and YET you are writing
somechesssoftware? how do you do that?
your software must use some algorithm to evaluate a given position. As
you cannot do that yourself,how do you teach a computer? do you use
other existing algorithms?
See I know all rules of Chess. But when I play a game I make so many
mistakes that I am never able to win Beginner Level. It takes my
pieces like birthday cake.

I have good understanding of Chess by getting Help Bot suggestions. He
has helped me a lot. Thats the reasion I seek help from people on the
group. Help Bot suggestions and analysis help me understand the
weaknesses in the game.

Now, I do not find any thing wrong with the game. Yes it makes a few
illegal moves that needs to be corrected.

Beginner Level used to play better than me the first day the program
was launched. I was so happy 2 years back but players on this group
beat it like cucumber and sliced and eat it.

I am just 1100 rated as per the players here. I often play using
Rybka help. Then I am able to beat even the Easy and Normal Levels.

Bye
Sanny

Play Chess at: http://www.GetClub.com/Chess.html
Tobias Heidelmann
2008-07-12 17:16:24 UTC
Permalink
Post by Sanny
See I know all rules of Chess. But when I play a game I make so many
mistakes that I am never able to win Beginner Level. It takes my
pieces like birthday cake.
I have good understanding of Chess by getting Help Bot suggestions.
the reason i asked you ist because i am a student of computer science.
game theory in general is of some interest to me. Unfortunately you did
not reply to my other message concerning GetClub and open source. Have
you thought about that option?

And i am still curious. Do you want to reveal some internals of your
engine? The main question is, how do you compare two positions. ( i mean
positions when the engine has left the opening book)
Say, position A and a position B which is a couple of moves away from
position A. Do you compare material? possibilities of tactics (pins,
forks, ets.), and also, how do you weigh the possible moves. Which move
is more probable to occur in the game? how do you answer that question?

did you write your engine from scratch? or did you start with some
existing engines?

Greetings,
Tobias
Guest
2008-07-12 17:25:44 UTC
Permalink
Post by Tobias Heidelmann
Post by Sanny
See I know all rules of Chess. But when I play a game I make so many
mistakes that I am never able to win Beginner Level. It takes my
pieces like birthday cake.
I have good understanding of Chess by getting Help Bot suggestions.
the reason i asked you ist because i am a student of computer science.
game theory in general is of some interest to me. Unfortunately you did
not reply to my other message concerning GetClub and open source. Have you
thought about that option?
I doubt that he has.

He's rather secretive about his program. Not sure if it's shame or
arrogance.
Post by Tobias Heidelmann
And i am still curious. Do you want to reveal some internals of your
engine? The main question is, how do you compare two positions. ( i mean
Many people have asked him about some internal details, and he always
ignores the question.

I think he's so secretive because he believes he's such a good programmer
and his chess program is so good that he's right on the verge becoming
internationally famous and rich. (And unfortunately, I'm not joking....)
Post by Tobias Heidelmann
positions when the engine has left the opening book)
Say, position A and a position B which is a couple of moves away from
position A. Do you compare material? possibilities of tactics (pins,
forks, ets.), and also, how do you weigh the possible moves. Which move is
more probable to occur in the game? how do you answer that question?
Considering Sanny has said he knew nothing about computer chess when he
started, and that he has done no research during this time, it's uncertain
as to what kind of program structure he has. (Do we believe him about his
lack of research.... I think most of us do, else the program would have
progressed much faster.)

I have some suspicioun that it might even be selective search. That would
explain the tactical blunders that people talk about regularly. He's trying
to look at the plasible moves and is missing the tactical stuff because they
don't apear reasonable at first glance.

Unless the program is incredibly buggy (which is possible), a brute force
with just a tiny bit of knowledge would be expected to play substantially
better.
Post by Tobias Heidelmann
did you write your engine from scratch? or did you start with some
existing engines?
Considering the quality of play and that it's still making illegal moves and
tactical blunders, it is safe to say that he started from scratch with
little to no knowldge of chess programming.
Post by Tobias Heidelmann
Greetings,
Tobias
----== Posted via Pronews.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.pronews.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= - Total Privacy via Encryption =---
Sanny
2008-07-12 17:58:34 UTC
Permalink
Post by Guest
I have some suspicioun that it might even be selective search. That would
explain the tactical blunders that people talk about regularly. He's trying
to look at the plasible moves and is missing the tactical stuff because they
don't apear reasonable at first glance.
Unless the program is incredibly buggy (which is possible), a brute force
with just a tiny bit of knowledge would be expected to play substantially
better.
I am very satisfied with the Game GetClub is playing. I only wanted it
for general Players and not to compete with Fritz & Rybka which are
the leaders in Chess.

GetClub is just a penny in front of these Big programs.

GetClub Chess is not for competing with Chess Engines but only for
fun. and people wishing to pass there time playing Chess.

Bye
Sanny

Play Chess at: http://www.GetClub.com/Chess.html
Tobias Heidelmann
2008-07-12 23:29:58 UTC
Permalink
Post by Sanny
Post by Guest
I have some suspicioun that it might even be selective search. That would
explain the tactical blunders that people talk about regularly. He's trying
to look at the plasible moves and is missing the tactical stuff because they
don't apear reasonable at first glance.
Unless the program is incredibly buggy (which is possible), a brute force
with just a tiny bit of knowledge would be expected to play substantially
better.
I am very satisfied with the Game GetClub is playing. I only wanted it
for general Players and not to compete with Fritz & Rybka which are
the leaders in Chess.
GetClub is just a penny in front of these Big programs.
GetClub Chess is not for competing with Chess Engines but only for
fun. and people wishing to pass there time playing Chess.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
instead of simply ignoring my questions concerning the internals of your
program it would be much nicer if you just said something like "no, i do
not reveal anything, my program is my big freaking secret"

greetings
Tobias
Patrick Volk
2008-07-13 04:00:59 UTC
Permalink
On Sun, 13 Jul 2008 01:29:58 +0200, Tobias Heidelmann
Post by Tobias Heidelmann
Post by Sanny
Post by Guest
I have some suspicioun that it might even be selective search. That would
explain the tactical blunders that people talk about regularly. He's trying
to look at the plasible moves and is missing the tactical stuff because they
don't apear reasonable at first glance.
Unless the program is incredibly buggy (which is possible), a brute force
with just a tiny bit of knowledge would be expected to play substantially
better.
I am very satisfied with the Game GetClub is playing. I only wanted it
for general Players and not to compete with Fritz & Rybka which are
the leaders in Chess.
GetClub is just a penny in front of these Big programs.
GetClub Chess is not for competing with Chess Engines but only for
fun. and people wishing to pass there time playing Chess.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
instead of simply ignoring my questions concerning the internals of your
program it would be much nicer if you just said something like "no, i do
not reveal anything, my program is my big freaking secret"
It's in Java, I bet you could find it off his site... If he does
security like he (er, I'm sorry, his *team*) programs.
Post by Tobias Heidelmann
greetings
Tobias
Sanny
2008-07-13 07:06:20 UTC
Permalink
Post by Patrick Volk
Post by Tobias Heidelmann
instead of simply ignoring my questions concerning the internals of your
program it would be much nicer if you just said something like "no, i do
not reveal anything, my program is my big freaking secret"
It's in Java, I bet you could find it off his site... If he does
security like he (er, I'm sorry, his *team*) programs.
In case someone does it I will do all needed to put him 10 years in
Jail (Prision) And ask 1 million Dollar as Fine.

Patrick Volf are you a hacker? ***@comcast.net Your email is with
me If you do any thing wrong I will catch you.

This group is full of Hackers. Save yourself !!!

Bye
Sanny
zzz
2008-07-13 11:32:14 UTC
Permalink
Sanny, you misunderstand. People are asking for your source code so
they can help you make improvements and find bugs, not to steal
anything.
Right now you don't have anything worth stealing anyway. If you let
people help you, maybe some day you will.
Tobias Heidelmann
2008-07-13 16:58:28 UTC
Permalink
sorry, i cannot resist this. Your empty threats annoy me too much.
because you used a similar threat in a reply to me, i will respond to this.

But first, let me say that i do not think Patricks reply was nice. If
Sanny does not want to share, thats ok. It is his perfect right to keep
his code a secret, it is his work, after all. Nobody has any right to
critizise Sanny for that.
Post by Sanny
In case someone does it I will do all needed to put him 10 years in
Jail (Prision) And ask 1 million Dollar as Fine.
look, not everybody is located in the us. good luck in asking a million
dollars and 10 years prison from the jurisdiction of a european country
as a penalty for some guy who hacked a worthless site if no damage was
done. Things dont work in europe that way.
Post by Sanny
me If you do any thing wrong I will catch you.
congrats, you have an ip and email of some guy who wrote something on
usenet.
Even if the same guy hacks your site, he would probably have another ip.
it is called dhcp and is a standard protocol that assigns ips dynamically.

Furhtermore, if some guy actually hacks your site, you have to get THIS
guys ip and perhaps email adress, but in case of the email i cannot see
how you could accomplish this, unless he sends it to you.

so, lets say you manage to get his ip. then you will have to make an
international case, because most likely it will be an ip from some
provider in another country. these things could actually take some time
and money.
And considering the potential damage you might suffer from the attack
(as far as i can see the damage is nil) i think it is very unlikely that
you might actually make a profit or much less put that bad guy in prison.
Post by Sanny
This group is full of Hackers. Save yourself !!!
this is not only annoying, but laughable.
Post by Sanny
Bye
Sanny
ok, thaths it. i am done with this discussion. No hard feelings Sanny,
but making empty threats in reply to offers of help is not a nice way to
talk to people.

Greetings,
Tobias
Patrick Volk
2008-07-13 18:22:31 UTC
Permalink
On Sun, 13 Jul 2008 00:06:20 -0700 (PDT), Sanny
Post by Sanny
Post by Patrick Volk
Post by Tobias Heidelmann
instead of simply ignoring my questions concerning the internals of your
program it would be much nicer if you just said something like "no, i do
not reveal anything, my program is my big freaking secret"
It's in Java, I bet you could find it off his site... If he does
security like he (er, I'm sorry, his *team*) programs.
In case someone does it I will do all needed to put him 10 years in
Jail (Prision) And ask 1 million Dollar as Fine.
Hmm.. You do realize Java code is exported to the target machine and
executed?

So you're willing to put the code on the persons' machine, but they're
not supposed to look at it. Curious.

You have it hosted in Phoenix?
Post by Sanny
me If you do any thing wrong I will catch you.
As I said before, I have better things to do with my time. My job
depends a lot on trust, and I wouldn't compromise that.

I hack when I have to, but I'm much better at reverse engineering.
Listen to people enough, and you can not only figure out what their
program does, but how, and even the flaws, and where they're going to
have them if they're not known.
That is a useful skill, because that ability allows you to pick up
programs where the original person left in the middle. You can
ascertain what they were doing, and fill in their thoughts, in a
manner of speaking.
Post by Sanny
This group is full of Hackers. Save yourself !!!
Find out how to fix your hard disk?
Post by Sanny
Bye
Sanny
Sanny
2008-07-13 07:02:46 UTC
Permalink
Post by Tobias Heidelmann
instead of simply ignoring my questions concerning the internals of your
program it would be much nicer if you just said something like "no, i do
not reveal anything, my program is my big freaking secret"
greetings
Tobias- Hide quoted text -
GetClub program is only for playing not for hacking Extra. If you do
it your IP Address is with me. And also your Email Account. Why will I
give you the code of my program?

Sorry I can't help you.

You can get 100s of Programs with Open Source with all code available.
Go and try there they will teach you how to play Chess and other
stuff.

Bye
Sanny
Tobias Heidelmann
2008-07-13 16:07:37 UTC
Permalink
Post by Sanny
GetClub program is only for playing not for hacking Extra. If you do
it your IP Address is with me. And also your Email Account. Why will I
give you the code of my program?
Dude, CHILL!

I do not want to get your precious source code by any illegal means. Why
should i?!? You admit there are still bugs, the webdesign has room for
improvement, and so on. As you pointed out, there is enough open source
already, most probably there will be an engine or two in my linux
distribution, which i can use and modify legally. If i wanted to start
an online chess site or write a chess program, i would use these. No
need to reinvent the wheel if commercial interests are at stake.

The point is, it might be fun to build an engine from scratch. the point
of open source is, that development can be much faster, dynamic and more
fun because you do it in a community.

Well, as you do not want to share, thats fine. At least you said so
instead of ignoring such questions.

Greetings,
Tobias
Frisco Del Rosario
2008-07-12 21:08:04 UTC
Permalink
Post by Guest
I think he's so secretive because he believes he's such a good programmer
and his chess program is so good that he's right on the verge becoming
internationally famous and rich. (And unfortunately, I'm not joking....)
*That's* the thing that so amazes me, that Sanny really believes that
his ugly, unusable, unfriendly, insufficiently tested software is a few
steps from the big time.

But his planewreck piece of shit GetClub Chess did make me realize that
I've been too relaxed about the design stage in my own web development
project, so thanks for that.
Guest
2008-07-13 14:51:47 UTC
Permalink
Post by Frisco Del Rosario
Post by Guest
I think he's so secretive because he believes he's such a good programmer
and his chess program is so good that he's right on the verge becoming
internationally famous and rich. (And unfortunately, I'm not joking....)
*That's* the thing that so amazes me, that Sanny really believes that
his ugly, unusable, unfriendly, insufficiently tested software is a few
steps from the big time.
Look at his reply today. Threatening people with lawsuits and jail because
they are curious why his program has been around so long but yet still so
bad. And he probably thinks he can actually do it, too.

Nevermind the fact that he uploads his program to everybody that wants to
play, so he has no way of knowing who might take it and try to disassemble
it.

Never mind that actually disassembling it into something readable &
understandable is a major undertaking. One that few people would think
about considering the quality of GC's play compared to 500 other freely
available chess programs.


I really don't think he's grasped just how insignificant his program is.
There are already lots of open source programs (many public domain) that are
almost certainly better than his. Even other Java programs.

I can understand putting a lot of work into your program and being proud of
it. I can understand having some pride in developing it yourself, with
little or no outside help. But he takes it a bit to far considering the end
result.


The only questions people have about his program involve why it still plays
so poorly. (Or so I've heard. I've never played it and have no intention
of doing so. But if Glaurung can beat it with a 1800 time difference, then
something is obviously not right.)

People wonder what kind of chess program structure could he have developed
on his own that still plays so poorly after all this time compared to the
simple methods people have developed over the past 40 years that are proven
to play pretty well with minimal effort.


I wonder what's the lowest, simplest program you can find that can beat
GC... One of Muller's MicroMax programs, could probably do it. The
strongest is under 2k of C source (excluding comments & whitespace.) A
simpler one is under 1k of C, I think.

I think the ChessProgrammingWiki chess engine is about done. I think it
still lacks decent time control and repetition and stuff.

I suppose a basic "Tech" style program would be a good test. Material
balance only. But they tend to be fast and have a deep qsearch, so are
tactically very strong. But positionally weak.


I wonder about the JavaScript chess programs? I wonder if those could beat
it. There were a couple that were supposed to be surprisingly decent.

http://p4wn.sourceforge.net/

Not sure which one is the strongest, but probably
http://p4wn.sourceforge.net/main-branch/

This is back from 2005, so there might be newer, better ones on the web
somewhere.

(Admittedly the javascript program is pretty limited... So who knows, GC
might be able to beat it.)
Post by Frisco Del Rosario
But his planewreck piece of shit GetClub Chess did make me realize that
I've been too relaxed about the design stage in my own web development
project, so thanks for that.
----== Posted via Pronews.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.pronews.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= - Total Privacy via Encryption =---
Peter Osterlund
2008-07-13 19:57:57 UTC
Permalink
Post by Guest
I wonder what's the lowest, simplest program you can find that can beat
GC... One of Muller's MicroMax programs, could probably do it. The
strongest is under 2k of C source (excluding comments & whitespace.) A
simpler one is under 1k of C, I think.
I played a game between Micro-Max 4.0 and getclub beginner level.
Micro-Max consists of 131 lines of C code and can be downloaded from
this web page:

http://home.hccnet.nl/h.g.muller/progress.html

The average thinking time of Micro-Max was 1.77 seconds. Here is the
game:

[Date "2008.07.13"]
[White "GetClub beginner level"]
[Black "Micro-Max 4.0"]
[Result "0-1"]

1. e4 e5 2. d4 exd4 3. Qxd4 Nc6 4. Qd3 Nf6 5. Bf4 d5 6. exd5 Nxd5 7. Ne2 g5
8. Be3 Ne5 9. Qe4 Bd6 10. Bxg5 Qxg5 11. Qxd5 Qg6 12. Nd4 O-O 13. Nc3 Re8
14. O-O-O Ng4 15. Bd3 Qh6+ 16. Kb1 Nxf2 17. Rde1 Be6 18. Nxe6 Nxh1 19. g3
Nxg3 20. hxg3 fxe6 21. Qxb7 Rab8 22. Qe4 Qg7 23. Rf1 Be5 24. Kc1 Qh6+ 25.
Rf4 Bxf4+ 26. gxf4 c5 27. Bc4 Rbd8 28. Nb5 Kf8 29. a3 Re7 30. b3 e5 31.
Qf5+ Kg7 32. Qg5+ Qxg5 33. fxg5 Kg6 34. c3 a6 35. a4 axb5 36. Bxb5 Kxg5 37.
Bc6 Rf7 38. a5 Ra7 39. Ba4 Rxa5 40. c4 Rxa4 41. bxa4 e4 42. a5 e3 43. Kb2
e2 44. a6 e1=Q
{White resigns} 0-1
--
Peter Osterlund - ***@telia.com
http://web.telia.com/~u89404340
Guest
2008-07-13 20:31:20 UTC
Permalink
Post by Peter Osterlund
Post by Guest
I wonder what's the lowest, simplest program you can find that can beat
GC... One of Muller's MicroMax programs, could probably do it. The
strongest is under 2k of C source (excluding comments & whitespace.) A
simpler one is under 1k of C, I think.
I played a game between Micro-Max 4.0 and getclub beginner level.
Micro-Max consists of 131 lines of C code and can be downloaded from
http://home.hccnet.nl/h.g.muller/progress.html
The average thinking time of Micro-Max was 1.77 seconds. Here is the
Why am I not really surprised....

Incidentally, his full download page is a little hidden.

http://home.hccnet.nl/h.g.muller/dwnldpage.html

It has several versions, including v4.8 which can be done via winboard.
(I've heard that v4.8 is a popular choice for beginning chess programmers
because if your program can't beat v48, then you still have quite a bit of
work to do.)

v4.8 is a little bigger at 1953 characters (excluding whitespace & comments)

He's done a heck of a job with his MicroMax programs.


Thanks for running that.

MicroMax is about as simplistic as you can reasonably expect. And you doing
v4.0 (instead of the slightly larger and stronger 4.8) was a good choice.
It gives Sanny something to aim for.



For those of you who don't remember H. G. Muller, way back in the 80's, he
was the guy who showed up at a computer chess tournament with a chess
computer in a matchbox. Literally a matchbox.

It had a cpu, a memory chip, a battery, an LED, and a couple staples to use
for move input.










----== Posted via Pronews.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.pronews.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= - Total Privacy via Encryption =---
Sanny
2008-07-14 07:59:42 UTC
Permalink
I played a game between Micro-Max 4.0 andgetclubbeginner level.
Micro-Max consists of 131 lines of C code and can be downloaded from
The average thinking time of Micro-Max was 1.77 seconds. Here is the
Do you gave you Core 2 Duo to GetClub. I find you give weaker computer
to GetClub. Just Change that give Core2 Duo to GetClub Chess and then
tell me the result.

And do not use multiple processors As GetClub only uses single
processor.

Bye
Sanny

Play Chess at: http://www.GetClub.com/Chess.html
Peter Osterlund
2008-07-14 09:25:10 UTC
Permalink
Post by Sanny
Post by Peter Osterlund
I played a game between Micro-Max 4.0 and getclub beginner level.
Micro-Max consists of 131 lines of C code and can be downloaded from
The average thinking time of Micro-Max was 1.77 seconds. Here is the
Do you gave you Core 2 Duo to GetClub. I find you give weaker computer
to GetClub. Just Change that give Core2 Duo to GetClub Chess and then
tell me the result.
And do not use multiple processors As GetClub only uses single
processor.
Micro-Max is a deterministic program, so it always plays the same move
in the same position. The only difference when changing computer is
how long it takes to compute the move. I replayed the game with
Micro-Max on the P4, and the average thinking time was then about
2.73s instead of 1.77s on the Core2. 2.73s is still less time than
what GetClub was using on average in that game.

As for GetClub, I don't know how to measure its search speed, so I
don't know how much faster it would get if I ran it on the Core2 CPU.
A rough guess is that it gets the same speed-up as Micro-Max, ie 1.5x
faster.
(I haven't tested because I run Fedora 8 64bit on the Core2, and for
some reason the java implementation doesn't work with GetClub chess.)

Does GetClub chess use 64-bit datatypes? If it does, then it might get
another 2x speed-up if you run it on a 64-bit computer, like the
Core2. (Crafty seems to get about 2x faster on a 64-bit-computer
compared to the same computer running in 32-bit mode.)
--
Peter Osterlund - ***@telia.com
http://web.telia.com/~u89404340
Sanny
2008-07-14 09:40:02 UTC
Permalink
Post by Peter Osterlund
Micro-Max is a deterministic program, so it always plays the same move
in the same position. The only difference when changing computer is
how long it takes to compute the move. I replayed the game with
Micro-Max on the P4, and the average thinking time was then about
2.73s instead of 1.77s on the Core2. 2.73s is still less time than
whatGetClubwas using on average in that game.
Download Firefox browser. Its free to download. and it will play
GetClub Chess if you have Java installed.

http://www.mozilla.com/en-US/firefox/all.html

Gore2duo is 10 times faster than the Pentium Processor.

2x due to dual core
2x due to 64 bit
2x due to faster CPU. (moores law)

So total it is 2x2x2 = 8 times faster than your old Pentium.

So multiply by 10 to the time taken on core2duo processor.

But when GetClub will play on it GetClub will not take advantage of
dual core. As it only uses 1 core.

So GetClub will run -5 times faster on your Core2Duo.

Play Chess at: http://www.GetClub.com/Chess.html
Post by Peter Osterlund
(I haven't tested because I run Fedora 8 64bit on the Core2, and for
some reason the java implementation doesn't work withGetClubchess.)
Download Firefox browser. Its free to download. and it will play
GetClub Chess if you have Java installed.

http://www.mozilla.com/en-US/firefox/all.html
Post by Peter Osterlund
DoesGetClubchessuse 64-bit datatypes? If it does, then it might get
another 2x speed-up if you run it on a 64-bit computer, like the
Core2. (Crafty seems to get about 2x faster on a 64-bit-computer
compared to the same computer running in 32-bit mode.)
Try and find that I have not yet tested if the game is 2x faster on
dual core. I think may be if Browser supports 64 bit computing. All
depends on Browser.

Bye
Sanny

Play Chess at: http://www.GetClub.com/Chess.html
Peter Osterlund
2008-07-14 11:19:50 UTC
Permalink
Post by Sanny
Post by Peter Osterlund
Micro-Max is a deterministic program, so it always plays the same move
in the same position. The only difference when changing computer is
how long it takes to compute the move. I replayed the game with
Micro-Max on the P4, and the average thinking time was then about
2.73s instead of 1.77s on the Core2. 2.73s is still less time than
whatGetClubwas using on average in that game.
Download Firefox browser. Its free to download. and it will play
GetClub Chess if you have Java installed.
I already have firefox 2.0.0.15 installed and java version
java-1.7.0-icedtea. According to this test page

http://www.java.com/en/download/help/testvm.xml

the JVM is working. However, when I go to the GetClub chess page, it
says "Java Not Enabled Please Download Java Virtual Machine." If I
follow the download link to www.java.com and click on "Do I have
Java?" I get a response that says:

Congratulations!
You have the recommended Java installed (1.7.0).
Post by Sanny
Gore2duo is 10 times faster than the Pentium Processor.
It depends on what program you run. As I wrote before, for Micro-Max
the difference was 1.5x, not 10x. This is on a 2.6GHz Core2 and a
3.1GHz P4.
Post by Sanny
Post by Peter Osterlund
DoesGetClubchessuse 64-bit datatypes? If it does, then it might get
another 2x speed-up if you run it on a 64-bit computer, like the
Core2. (Crafty seems to get about 2x faster on a 64-bit-computer
compared to the same computer running in 32-bit mode.)
Try and find that I have not yet tested if the game is 2x faster on
dual core. I think may be if Browser supports 64 bit computing. All
depends on Browser.
I think it depends on the program too. If your chess program doesn't
use any 64-bit data types, I don't think it will run 2x faster just by
running it in a 64-bit environment.
--
Peter Osterlund - ***@telia.com
http://web.telia.com/~u89404340
Sanny
2008-07-14 11:46:50 UTC
Permalink
the JVM is working. However, when I go to theGetClubchesspage, it
says "Java Not Enabled Please Download Java Virtual Machine." If I
follow the download link towww.java.comand click on "Do I have
        Congratulations!
        You have the recommended Java installed (1.7.0).
I feel after you install java you need to restart the Computer. Or re-
download the Java Plugin.

Please see if your Browser allows Java in "Internet Options" of your
Browser.

When it works on Pentium 3 Ghz, I don't think it will not on your Core
2 Duo.

On my Computer the program works on Firefox.

Bye
Sanny

Play Chess at: http://www.GetClub.com/Chess.html
Peter Osterlund
2008-07-14 12:13:55 UTC
Permalink
Post by Sanny
the JVM is working. However, when I go to theGetClubchesspage, it
says "Java Not Enabled Please Download Java Virtual Machine." If I
follow the download link towww.java.comand click on "Do I have
        Congratulations!
        You have the recommended Java installed (1.7.0).
I feel after you install java you need to restart the Computer. Or re-
download the Java Plugin.
I installed java four days ago. The computer has been restarted
multiple times since then. I can play other java games, for example
this reversi game:

http://www.darkfish.com/turncoat/index.html
Post by Sanny
Please see if your Browser allows Java in "Internet Options" of your
Browser.
There are no "internet options" in firefox, and internet explorer is
not available in linux.
Post by Sanny
When it works on Pentium 3 Ghz, I don't think it will not on your Core
2 Duo.
On my Computer the program works on Firefox.
Probably because you are running Windows, while my Core2 machine is
running linux. My Pentium is also running windows and there the
GetClub chess program works.
--
Peter Osterlund - ***@telia.com
http://web.telia.com/~u89404340
Sanny
2008-07-14 13:59:31 UTC
Permalink
Post by Peter Osterlund
Post by Sanny
On my Computer the program works on Firefox.
Probably because you are running Windows, while my Core2 machine is
running linux. My Pentium is also running windows and there the
GetClub chess program works.
1. When do you last tried. Try again and let me know what you find at
Display area?

2. In FireFox There is a Check Box : Enable Java. Please see that
Check Box is enabled.

3. What is the error you get? Login & let me know what happened.

You are not facing any problem on Pentium 3.0. So I think it will work
on Linux too.

Bye
Sanny

Play Chess at: http://www.GetClub.com/Chess.html
Peter Osterlund
2008-07-14 14:27:43 UTC
Permalink
Post by Sanny
Post by Peter Osterlund
Post by Sanny
On my Computer the program works on Firefox.
Probably because you are running Windows, while my Core2 machine is
running linux. My Pentium is also running windows and there the
GetClub chess program works.
1. When do you last tried. Try again and let me know what you find at
Display area?
2. In FireFox There is a Check Box : Enable Java. Please see that
Check Box is enabled.
3. What is the error you get? Login & let me know what happened.
OK, I tried again, and it kind of works, but with a few quirks. First
it says:

Java Not Enabled Please Download Java Virtual Machine.
You must have JAVA to Play Chess at GetClub.com
Please, Download Java. From here.
(Restart Computer) when Java Installed.
Also Check Browser Settings that it allows Java.

This is displayed in the box where the chess program should be.
However, I found that I can scroll that box, and below the message the
java applet is running. It seems to work, except that when I'm
supposed to choose the level, I don't see any text. I can still click
on an empty line and then the chess game is starting. I've started a
new game with Micro-Max 4.0. It's not finished yet though.

I noticed another interesting thing. When I run GetClub chess on my
Core 2 Duo, the java applet is using both CPUs. The "top" program
shows that there are 4 java threads running that each get 50% of one
CPU. Since you have said that the chess algorithm only uses one core,
I think the three other threads are busy waiting for something. If the
same problem exists on windows single-cpu machines, you could speed up
the program by a factor of 4 by fixing these busy-wait problems.
--
Peter Osterlund - ***@telia.com
http://web.telia.com/~u89404340
Guest
2008-07-14 14:37:16 UTC
Permalink
Post by Peter Osterlund
This is displayed in the box where the chess program should be.
However, I found that I can scroll that box, and below the message the
java applet is running. It seems to work, except that when I'm
supposed to choose the level, I don't see any text. I can still click
on an empty line and then the chess game is starting. I've started a
new game with Micro-Max 4.0. It's not finished yet though.
I'm surprised you didn't go ahead and go up to v4.8 since it's stronger.
Just to make it tougher for Sanny's 3000 rated chess program...
Post by Peter Osterlund
I noticed another interesting thing. When I run GetClub chess on my
Core 2 Duo, the java applet is using both CPUs. The "top" program
shows that there are 4 java threads running that each get 50% of one
CPU. Since you have said that the chess algorithm only uses one core,
I think the three other threads are busy waiting for something. If the
same problem exists on windows single-cpu machines, you could speed up
the program by a factor of 4 by fixing these busy-wait problems.
Those are probably Java threads, and aren't releated to the chess program.

Since Java is an interpretive / JIT language, the Java VM can take a little
advantage of multiple cores (compiling a little faster in the background,
etc.), but once all that gets done, the cpu usage should drop back down to a
single core chess program running as expected.

In other words, although the extra cores may help GC in the first few
seconds, as the java bytecodes get converted into native code by the JIT,
after the first move or so, all that will be done and it'll be GetClub on
just a single core, with no performance improvement over a single core even
if you had 4 cores.


I could be wrong, though. Nobody knows how Sanny programmed his chess
program.

It is actually possible those are 4 search threads to take advantage of 4
core cpu's. Who knows.




----== Posted via Pronews.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.pronews.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= - Total Privacy via Encryption =---
Peter Osterlund
2008-07-14 15:01:18 UTC
Permalink
Post by Guest
I'm surprised you didn't go ahead and go up to v4.8 since it's stronger.
Just to make it tougher for Sanny's 3000 rated chess program...
It seems pointless to try stronger programs until I find one that is
able to lose against GetClub ;)
Post by Guest
Post by Peter Osterlund
I noticed another interesting thing. When I run GetClub chess on my
Core 2 Duo, the java applet is using both CPUs. The "top" program
shows that there are 4 java threads running that each get 50% of one
CPU. Since you have said that the chess algorithm only uses one core,
I think the three other threads are busy waiting for something. If the
same problem exists on windows single-cpu machines, you could speed up
the program by a factor of 4 by fixing these busy-wait problems.
Those are probably Java threads, and aren't releated to the chess program.
Since Java is an interpretive / JIT language, the Java VM can take a little
advantage of multiple cores (compiling a little faster in the background,
etc.), but once all that gets done, the cpu usage should drop back down to a
single core chess program running as expected.
In other words, although the extra cores may help GC in the first few
seconds, as the java bytecodes get converted into native code by the JIT,
after the first move or so, all that will be done and it'll be GetClub on
just a single core, with no performance improvement over a single core even
if you had 4 cores.
Well, the threads have been running all the time since I wrote my
previous post, so I doubt it's JIT. I'll try to see if I can duplicate
this problem in windows too. (I just need to figure out how to see
individual thread CPU usage in windows xp first.)
Post by Guest
I could be wrong, though. Nobody knows how Sanny programmed his chess
program.
It is actually possible those are 4 search threads to take advantage of 4
core cpu's. Who knows.
Yes, it's possible, but since he has explicitly said a number of times
that GetClub doesn't take advantage of multiple CPUs, I doubt it.

On the other hand, having three dummy threads that eat all CPU time
they can get is a simple way of making the program scale linearly with
the number of CPUs, at least up to 4 CPUs. ;)
--
Peter Osterlund - ***@telia.com
http://web.telia.com/~u89404340
Guest
2008-07-14 15:31:42 UTC
Permalink
Post by Peter Osterlund
Post by Guest
I'm surprised you didn't go ahead and go up to v4.8 since it's stronger.
Just to make it tougher for Sanny's 3000 rated chess program...
It seems pointless to try stronger programs until I find one that is
able to lose against GetClub ;)
(laugh) I guess you are right... Save the rest for next week when Sanny
says the program's strength has double 5 more times.
Post by Peter Osterlund
Well, the threads have been running all the time since I wrote my
previous post, so I doubt it's JIT. I'll try to see if I can duplicate
this problem in windows too. (I just need to figure out how to see
individual thread CPU usage in windows xp first.)
If they are running constantly, then either there is an issue with the Java
interpreter itself, or it's something in Sanny's program.
Post by Peter Osterlund
Post by Guest
I could be wrong, though. Nobody knows how Sanny programmed his chess
program.
It is actually possible those are 4 search threads to take advantage of 4
core cpu's. Who knows.
Yes, it's possible, but since he has explicitly said a number of times
that GetClub doesn't take advantage of multiple CPUs, I doubt it.
I missed him saying that, so I didn't know.

I've never programmed in Java, but are there any tasks that can be ordered?
Perhaps those are running constantly in the background just because they
never got shut off or put down to a lower priority?

I think in Java it's easy to spawn off extra threads / processes / tasks, so
maybe he did that for something and it's just stuck in an idle loop.


Oh well... I'm sure Sanny will figure it out. Then he'll say it plays 4
times better than before. (Although just what '4 times better' actually
means, he never says. 4 times fewer tactical mistakes? 4 times fewer
blunders? 4 times the speed? 4 times the ratings?)
Post by Peter Osterlund
On the other hand, having three dummy threads that eat all CPU time
they can get is a simple way of making the program scale linearly with
the number of CPUs, at least up to 4 CPUs. ;)
There you go! You solved it! (laugh)

Seriously though, if it is his program that has 4 threads going, and three
of them are wasted (stuck in an idle loop or some other useless task), then
that means his chess thread would be getting only 25% of the cpu. Removing /
fixing those other 3 threads would indeed speed up his program.

He'd probably see a performance increase of 50-75 points. Real increase,
not a Sanny increase.

(Caveat... that'd be for a single core cpu. For dual core systems, less
than 50 points. For quad core systems, no increase at all. This is due to
the distribution of those idle threads on multiple cores.)

But that assumes those three threads are indeed wasted and aren't doing
anything useful.





----== Posted via Pronews.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.pronews.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= - Total Privacy via Encryption =---
Peter Osterlund
2008-07-14 15:26:32 UTC
Permalink
I've started a new game with Micro-Max 4.0. It's not finished yet
though.
The game is finished now, and it was a draw. Although GetClub cheated
as usual and declared the draw one move too early, in this case it
didn't matter, because Micro-Max wouldn't have made any attempt to
avoid the draw anyway.

Here is the game. Micro-Max made multiple severe tactical mistakes,
but it was lucky and GetClub didn't find the correct moves to exploit
the mistakes.

[Date "2008.07.11"]
[White "Micro-Max 4.0"]
[Black "GetClub easy level"]
[Result "1/2-1/2"]

1. e3 e5 2. Nf3 e4 3. Ng1 d5 4. d3 Bd6 5. Nc3 Nf6 6. dxe4 dxe4 7. f3 exf3
8. Nxf3 O-O 9. Be2 Nc6 10. O-O Re8 11. Bb5 Bd7 12. Bd2 Ng4 13. Nd5 Bf5 14.
h3 Be6 15. hxg4 Bxd5 16. c4 Be4 17. Kf2 Be5 18. Bxc6 Bxc6 19. Nxe5 Rxe5 20.
Kg3 Qd6 21. Kh3 Qh6+ 22. Kg3 Be4 23. Qe1 Rae8 24. Bc3 R5e7 25. Bb4 Re5 26.
Bc3 R5e7 27. Bb4 Re5 28. Bc3
{Draw by repetition} 1/2-1/2

(I added the last Bc3 move myself to get a correct draw annotation.
GetClub incorrectly thought it had the right to claim a draw after
27... Re5.)
--
Peter Osterlund - ***@telia.com
http://web.telia.com/~u89404340
Sanny
2008-07-14 16:06:00 UTC
Permalink
Post by Peter Osterlund
The game is finished now, and it was a draw. Although GetClub cheated
as usual and declared the draw one move too early, in this case it
didn't matter, because Micro-Max wouldn't have made any attempt to
avoid the draw anyway.
So Stronger Processor made it draw. How much time each program was
taking?

I am very thankful to you that you found 4 threads running
simulataniously. So GetClub is 4 times slow.

Can you find out whether those 4 threads are part of Java Applet or
they are else where?

Are these threads taking 50% of each Processors? Can you get the name
of these threads or any other details of these thread?

Because my program works as a single thread.

It may happen the JVM is distributing the code to 4 processors.

Please tell me more details abt the threads.

Bye
Sanny

Play Chess at: http://www.GetClub.com/Chess.html
Guest
2008-07-14 16:50:01 UTC
Permalink
Post by Sanny
Post by Peter Osterlund
The game is finished now, and it was a draw. Although GetClub cheated
as usual and declared the draw one move too early, in this case it
didn't matter, because Micro-Max wouldn't have made any attempt to
avoid the draw anyway.
So Stronger Processor made it draw. How much time each program was
taking?
Sanny, we've warned you before about jumping to conclusions of a single
game.

Occasionally even a very weak program will beat a very strong program simply
due to the weak program accidently making a strong move for the wrong
reasons.

Several games need to be done under similar situations before you should
even think about drawing any conclusions.
Post by Sanny
I am very thankful to you that you found 4 threads running
simulataniously. So GetClub is 4 times slow.
*ONLY* on a single core system. You might get 50-75 points increase. (But
how many people still have a single core system?)

On a dual core system, it's only 2 times slow. You might get 25-40 points
increas.

On a quad core system, it's running at full speed and you wont get any
increase.

Those points increase will only happen if you can get rid of those extra
threads (or drastically reduce their cpu usage) and increase the cpu usage
for the main chess search.

These will be *real* ratings increases. Remember the Technology Curve I
mentioned before? Even though those are just estimates, they will still be
real. Not some imaginary number you picked out of thin air.
Post by Sanny
Can you find out whether those 4 threads are part of Java Applet or
they are else where?
Are these threads taking 50% of each Processors? Can you get the name
of these threads or any other details of these thread?
Because my program works as a single thread.
It may happen the JVM is distributing the code to 4 processors.
(First a clairification. He didn't say 4 processors. He said 4 processes.
Or 4 threads. Which could be done to 4 processors, but he has a dual core
system, not a quad core system.)

Not distributing *your* code to 4 processes.

You can't take a single threaded program and magically make it suitable for
multi-core.

There is a lot of hype about various pre-processors or libraries etc. etc.
that can take a regular program and make it multi-threaded / multi-process
suitable for a multi-core system.

Unfortunately, they don't work nearly as well as they claim.

Plus, chess (and other tree searching applications) are rather specialized
and don't respond to those 'magical' methods. You have to explicitly do the
work and the tuning. The hard way.



Perhaps you are starting some java task / background process to do something
and it's not ending or reducing its priority?

Or it's some background java task (garbage collection?) that is being
heavily used and is therefore taking such drastic cpu.

Or it could be some issue with the version of Java that he uses.

About anything is possible.

Have you checked your own system to see how many java threads are running?
That would answer several questions.

Perhaps the GUI is one task and the chess engine is another task? That
would account for two of the threads.

You are in the best situation to figure out what's going on. It's your code
and nobody else knows anything about it.
Post by Sanny
Please tell me more details abt the threads.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
----== Posted via Pronews.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.pronews.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= - Total Privacy via Encryption =---
Peter Osterlund
2008-07-14 18:28:44 UTC
Permalink
Post by Sanny
Post by Peter Osterlund
The game is finished now, and it was a draw. Although GetClub cheated
as usual and declared the draw one move too early, in this case it
didn't matter, because Micro-Max wouldn't have made any attempt to
avoid the draw anyway.
So Stronger Processor made it draw. How much time each program was
taking?
I didn't measure exactly but I think Micro-Max was taking
approximately 5 seconds per move. However, it was slowed down by the 4
java threads that were running on the same machine. GetClub was at
easy level and I guess it was thinking about 40 seconds per move on
average.
Post by Sanny
I am very thankful to you that you found 4 threads running
simulataniously. So GetClub is 4 times slow.
Can you find out whether those 4 threads are part of Java Applet or
they are else where?
Are these threads taking 50% of each Processors? Can you get the name
of these threads or any other details of these thread?
They are all listed with the same name, "pluginappletviewer". They
took about 50% CPU time each. After the game ended with a draw, three
of the threads were still running (taking about 67% each). When I
closed the GetClub page in the browser, those three threads
disappeared too.
Post by Sanny
Because my program works as a single thread.
It may happen the JVM is distributing the code to 4 processors.
Please tell me more details abt the threads.
When I run GetClub chess in internet explorer in windows xp, all CPU
time is taken up by the iexplore.exe program. Explorer says that it's
using "Java(TM) SE Runtime Environment 6 Update 5". I downloaded
procexp.exe, which says that almost all CPU time is consumed by a
single thread in IE. However, there could still be several java
threads in the program that all share the same OS thread. I guess you
need some kind of java debugger to tell if that is the case. (I'm not
a java programmer, so I don't know.)
--
Peter Osterlund - ***@telia.com
http://web.telia.com/~u89404340
Sanny
2008-07-14 18:37:59 UTC
Permalink
Post by Peter Osterlund
They are all listed with the same name, "pluginappletviewer". They
took about 50% CPU time each. After the game ended with a draw, three
of the threads were still running (taking about 67% each). When I
closed the GetClub page in the browser, those three threads
disappeared too.
A few things were corrected. Play another game and let me know if
these 4 threads again popup?

And also if the threads are taking same 50% of the time?

Bye
Sanny

Play Chess at: http://www.GetClub.com/Chess.html
Peter Osterlund
2008-07-14 19:22:35 UTC
Permalink
Post by Sanny
Post by Peter Osterlund
They are all listed with the same name, "pluginappletviewer". They
took about 50% CPU time each. After the game ended with a draw, three
of the threads were still running (taking about 67% each). When I
closed the GetClub page in the browser, those three threads
disappeared too.
A few things were corrected. Play another game and let me know if
these 4 threads again popup?
And also if the threads are taking same 50% of the time?
I started another game, and now there is only one thread which takes
100% of one CPU, so whatever you did it appears that it works.
--
Peter Osterlund - ***@telia.com
http://web.telia.com/~u89404340
Sanny
2008-07-15 06:49:06 UTC
Permalink
Post by Peter Osterlund
Post by Sanny
A few things were corrected. Play another game and let me know if
these 4 threads again popup?
And also if the threads are taking same 50% of the time?
I started another game, and now there is only one thread which takes
100% of one CPU, so whatever you did it appears that it works.
A few more things were corrected today, Play on your dual core and let
me know if you find any speed improvements.

Do you see the flickering of pieces removed?

Bye
Sanny

Play Chess at: http://www.GetClub.com/Chess.html
Peter Osterlund
2008-07-15 09:27:59 UTC
Permalink
Post by Sanny
Post by Peter Osterlund
Post by Sanny
A few things were corrected. Play another game and let me know if
these 4 threads again popup?
And also if the threads are taking same 50% of the time?
I started another game, and now there is only one thread which takes
100% of one CPU, so whatever you did it appears that it works.
A few more things were corrected today, Play on your dual core and let
me know if you find any speed improvements.
No improvements. Now there are 4 threads again that each get about 50%
of the CPU. The previous version was better.
Post by Sanny
Do you see the flickering of pieces removed?
The pieces don't flicker, but when you move them the update frequency
is very low, maybe two times per second. So it doesn't really look
like they are moved, more like they occasionally jump to the current
mouse pointer position.
--
Peter Osterlund - ***@telia.com
http://web.telia.com/~u89404340
Peter Osterlund
2008-07-17 23:52:12 UTC
Permalink
Post by Peter Osterlund
Post by Sanny
Post by Peter Osterlund
Post by Sanny
A few things were corrected. Play another game and let me know if
these 4 threads again popup?
And also if the threads are taking same 50% of the time?
I started another game, and now there is only one thread which takes
100% of one CPU, so whatever you did it appears that it works.
A few more things were corrected today, Play on your dual core and let
me know if you find any speed improvements.
No improvements. Now there are 4 threads again that each get about 50%
of the CPU. The previous version was better.
I take that back. Further testing indicates that the problem lies
within the java implementation (icedtea) that comes with Fedora 8.
When playing sounds, a separate thread is started and that thread then
sometimes gets stuck in an infinite loop using all CPU time it can
get. Some googling revealed that the problem is known but a solution
doesn't appear to exist yet.

The problem that the web page says that java is not installed is also
a bug in the browser and/or java plugin. For some reason, javascript
thinks that java is disabled, even though in reality java is enabled.
I haven't found a solution for that yet.

So those problems were not caused by GetClub chess. However, I think
there is one problem caused by GetClub chess:

The font used by GetClub chess is larger in linux than in windows. My
guess is that you are using some non-standard font that is not
available in linux. The effect is that some text falls outside
buttons/windows, and thus becomes unreadable.

Btw, when I was investigating the above problems, I looked at some
backtraces from the GetClub chess program. I got the feeling that the
Q-search is rather inefficient. Maybe you need to implement MVV/LVA or
SEE move ordering.
--
Peter Osterlund - ***@telia.com
http://web.telia.com/~u89404340
Guest
2008-07-14 14:30:20 UTC
Permalink
Post by Sanny
Gore2duo is 10 times faster than the Pentium Processor.
2x due to dual core
2x due to 64 bit
2x due to faster CPU. (moores law)
So total it is 2x2x2 = 8 times faster than your old Pentium.
Doesn't work like that Sanny.

First, a single core program will get ZERO performance boost from multiple
cores. MicroMax is a single core program. For a program, any program, to
get a performance boost from multiple cores require they be designed for
multiple cores. For chess that means parallel searching. It doesn't happen
automatically. In fact, a lot of people have put a lot of effort into
researching parallel searching and trying to get the efficiency up. For 2
cores, you can expect maybe a 1.5x performance boost. For 4 cores, maybe
3x. It varies, but it's not 1:1 performance growth.

Second, 64 bits wont give you 2x performance. If you deal with lots of 64
bit data, you will see a performance boost if you have a 64 bit cpu and a 64
bit OS. However, the benefits to chess (bitboards) usually minimal from
that. Less than 5% because a program will usually be doing things other
than 64 bit bitboards.

The other part of the performance boost comes from the 64 bit CPU having
more registers, so there is less register shuffling.

All total, the performance boost you will like see (for a 64 bit program
with 64 bit data on a 64 bit cpu with a 64 bit OS) will be between 10%-25%,
usually.

And if you have bad luck (misaligned data, overloaded caches because of
larger data, etc.) you can actually see a performance drop. It's possible
for it to actually be slower.


Third, saying 2x due to faster cpu, would depend a heck of a lot of what
cpu's you
are actually comparing.....

For a single core / thread / process program, going from a 3.0 ghz Pentium 4
to a 2.0 ghz core2duo would not be too different in performance. (Not
exact. I'm just saying no major performance difference for a simple,
typical chess program.)

For a faster core2, you'd see a higher performance difference. But still
nothing drastic. Under a factor of 1.5 due to the limitations of clock rate
for the available core2's.





----== Posted via Pronews.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.pronews.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= - Total Privacy via Encryption =---
Simon Krahnke
2008-07-15 12:26:21 UTC
Permalink
Post by Guest
First, a single core program will get ZERO performance boost from multiple
cores.
That's not exactly true. There is an OS and usually several background
processes which can use the other core.
Post by Guest
MicroMax is a single core program. For a program, any program, to
get a performance boost from multiple cores require they be designed for
multiple cores.
That's more important, however.
Post by Guest
Second, 64 bits wont give you 2x performance. If you deal with lots of 64
bit data, you will see a performance boost if you have a 64 bit cpu and a 64
bit OS. However, the benefits to chess (bitboards) usually minimal from
that. Less than 5% because a program will usually be doing things other
than 64 bit bitboards.
I think you are underestimating the usage of 64 bit values in a
bitboard oriented chess program. move generator, move doer, SEE and
the evaluation function are as inner loop as it gets, and all about
manipulating bitboards.
Post by Guest
The other part of the performance boost comes from the 64 bit CPU having
more registers, so there is less register shuffling.
All total, the performance boost you will like see (for a 64 bit program
with 64 bit data on a 64 bit cpu with a 64 bit OS) will be between 10%-25%,
usually.
Die Rybka FAQ gives 60% performance from 64 bits. Thats not 100% though.

mfg, simon .... l
Guest
2008-07-15 14:37:36 UTC
Permalink
Post by Simon Krahnke
Post by Guest
First, a single core program will get ZERO performance boost from multiple
cores.
That's not exactly true. There is an OS and usually several background
processes which can use the other core.
We are talking user program, not OS.

And the OS's overhead is usually close to zero unless you have background
tasks running (such as a defrag, etc.)
Post by Simon Krahnke
Post by Guest
Second, 64 bits wont give you 2x performance. If you deal with lots of 64
bit data, you will see a performance boost if you have a 64 bit cpu and a 64
bit OS. However, the benefits to chess (bitboards) usually minimal from
that. Less than 5% because a program will usually be doing things other
than 64 bit bitboards.
I think you are underestimating the usage of 64 bit values in a
bitboard oriented chess program. move generator, move doer, SEE and
the evaluation function are as inner loop as it gets, and all about
manipulating bitboards.
No, I've read quite a few people's comments about when they go from 32 to 64
bit OS's.

And I've written bitboard programs myself. (Although I don't normally use a
64 bit OS.)

Most of a bitboard program is not actually 64 bit operations. Most of it is
plain ordinary kind of stuff that is just as happy on a 32 bit cpu. That
part wont see any improvements from going to 64 bit.

Only the actual 64 bit data (bitboard & hash values) will see any
performance boost.

Doing the actual WRBoard ^= Sqr kind of stuff certainly double. (Unless you
do it as Board[Piece] ^= Sqr and then you'll have to deal with the indexing
too.)

But all the house keeping stuff, the function calls, the loops, etc. etc.
wont see anything. And those make up the majority of the program.


The bitscans and popcounts certainly get a performance boost. As do the new
'Magic" and Kindergarten bitboard attack getters that have been developed.

But those are such a small part of the whole program they don't
significantly effect things when going from 32 to 64 bits.
Post by Simon Krahnke
Post by Guest
The other part of the performance boost comes from the 64 bit CPU having
more registers, so there is less register shuffling.
All total, the performance boost you will like see (for a 64 bit program
with 64 bit data on a 64 bit cpu with a 64 bit OS) will be between 10%-25%,
usually.
Die Rybka FAQ gives 60% performance from 64 bits. Thats not 100% though.
I'm not in a position to say whether that's accurate or not. Whether that
was due to just a recompile or some other changes.

But there is a wide variation. Especially if you can make good use of the
extra registers.

The more advanced your program is, and the more it requires 64 bit stuff and
the more registers it requires, the better the performance boost will be.

However, the majority of programs when going from a 32 bit CPU to a 64 bit
CPU will not see that kind of performance boost. There is too much non-64
bit stuff going on.

On the register poor x86, the big area is the extra registers. You often
get more from that than the 64 bit data. Is some cases, you can take a
mailbox program and recompile it and get a performance boost solely because
of having the extra cpu registers for the compiler to use. (In some
cases.... I never could, but others have done that and reported
improvements. YMMV.)

Also, I think some x86 compilers change they data is done in 64 bit mode.
The code no longer likes global vars but prefers pointers etc. I'm not an
assembly code expert, so I can't go into details, but I've repeatedly heard
that some data structures that work great on the 32 bit model don't work
nearly as well on the 64 bit model. (This isn't about alignment or data
holes, etc.)


I've even tried it myself when I played with a 64 bit Vista. (Too many
other things didn't work right, so I stayed with 32 bit.)

I've compiled my own and a few others and many bitboard programs wont see
much improvement when going to 64 bit. There just isn't enough there to
really need the 64 bit cpu & OS.



And then, of course, there's the Java issue for Sanny's program. That
definetly gets in the way of any sort of 64 bit optimizations that a
compiled program could do.


Still, for the sake of argument, I'll change my statements.


1) A single core program will see ZERO improvement when going to a
multi-core cpu (this excludes any OS overhead.)

2) Going to a 64 bit cpu will often see little improvement because often,
most of the operations aren't based on 64 bit data. Often, normal data
operations, loops, indexing etc. will be the majority of your program. You
may see as little as 5% improvement. If your program is very heavily 64 bit
data, or you have key areas that are heavily used (the 80/20 rule) you may
see more.

3) For a typical, average bitboard chess program, your performance boost may
be around 25% or less. If your program is heavily optimized for a 64 bit
system, taking advantage of the extra registers and the compiler and the 64
bit nature, you might see 50%. Under rare occasions, highly tuned programs
might see a little more.









----== Posted via Pronews.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.pronews.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= - Total Privacy via Encryption =---
Guest
2008-07-12 17:14:33 UTC
Permalink
Post by Sanny
Post by Tobias Heidelmann
Dude, are you serious?
this makes me very curious.. i mean, you openly admit that you have
virtually no understanding ofchesswhatsoever and YET you are writing
somechesssoftware? how do you do that?
your software must use some algorithm to evaluate a given position. As
you cannot do that yourself,how do you teach a computer? do you use
other existing algorithms?
See I know all rules of Chess. But when I play a game I make so many
mistakes that I am never able to win Beginner Level. It takes my
pieces like birthday cake.
I have good understanding of Chess by getting Help Bot suggestions. He
has helped me a lot. Thats the reasion I seek help from people on the
group. Help Bot suggestions and analysis help me understand the
weaknesses in the game.
Now, I do not find any thing wrong with the game.
Yes it makes a few illegal moves that needs to be corrected.
Stunned amazement..... ROFL!!!

Man, that problem should not exist.

If you are *still* having illegal move problems after all this time (2
years?), you have got to be one BAD programmer.

The rules of chess are not that difficult to program.

Did it ever occur to you that you should make sure it plays legal chess
before trying to make the play better? That's one of the most basic things
you do.

Did it ever occur to you that even an *average* programmer could write a
chess program from scratch within a month, and it play fully legal chess.
And that even with material & mobility (which is trivial to program), I
suspect it would be at least as strong as yours is now? (If Glaurung is
beating your program with a 1sec vs. 30min thinking difference, then you
have *serious* problems.)

You really don't need to know much about chess to write a basic chess
program that is very tactically strong. Just be even a mediocre programmer
with some basic knowledge of chess and a chess programming.

Sheesh.... "Yes it makes a few illegal moves that needs to be corrected."
What an admission of incompetence. I can't believe you admitted that.


There are a lot of references on computer chess programming. You have to go
out of your way to not see them.

Maybe it's time you stop avoiding them?? Actually learn a few things about
what you are trying to program.

Maybe it's time for you to throw away your first program and start over and
concentrate on the basics? The basics would probably give you a real,
verifiable 2000+ rating on today's hardware. (Most beginning chess
programmers go through at least a couple 'learning' programs before they
learn enough to write a real one.)
Post by Sanny
Beginner Level used to play better than me the first day the program
was launched. I was so happy 2 years back but players on this group
beat it like cucumber and sliced and eat it.
I am just 1100 rated as per the players here. I often play using
Rybka help. Then I am able to beat even the Easy and Normal Levels.
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
----== Posted via Pronews.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.pronews.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= - Total Privacy via Encryption =---
Sanny
2008-07-12 18:00:38 UTC
Permalink
The rules ofchessare not that difficult to program.
Did it ever occur to you that you should make sure it plays legalchess
before trying to make the play better?  That's one of the most basic things
you do.
See, that illegal move is made when it finds Mate in 2. It just plays
King in Check. This actually never hamper any game.

Bye
Sanny

Play Chess at: http://www.GetClub.com/Chess.html
Sanny
2008-07-12 17:03:46 UTC
Permalink
I was able to win the Baby Level with 1 Rook extra.

Game Played between sanjay11 and baby_x_R at GetClub.com

--------------------------------------------------------------------------------
sanjay11: (White)
baby_x_R: (Black)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM22549&game=Chess
--------------------------------------------------------------------------------

White -- Black
(sanjay11) -- (baby_x_R)

1. e2-e4{4} e7-e5{0}
2. d2-d3{2} d7-d5{0}
3. e4-d5{2} Qd8-d5{2}
4. Nb1-c3{2} Bf8-b4{2}
5. Bc1-d2{4} Bb4-c3{2}
6. Bd2-c3{2} Ng8-f6{2}
7. Ng1-f3{2} Nb8-c6{2}
8. h2-h3{2} Ke8-g8{2}
9. a2-a3{4} Bc8-d7{4}
10. Bf1-e2{4} Ra8-d8{8}
11. Ke1-g1{4} Bd7-f5{4}
12. b2-b4{4} b7-b6{2}
13. Qd1-d2{10} h7-h6{4}
14. Rf1-e1{6} Rd8-d6{4}
15. Be2-f1{14} Rd6-e6{10}
16. Qd2-e3{4} e5-e4{2}
17. d3-e4{4} Re6-e4{2}
18. Qe3-d2{8} Re4-e1{26}
19. Ra1-e1{4} Kg8-h8{6}
20. Nf3-e5{2} Nc6-e5{2}
21. Qd2-d5{6} Nf6-d5{2}
22. Re1-e5{2} Bf5-e6{4}
23. Bc3-b2{8} c7-c6{2}
24. Bf1-c4{4} b6-b5{2}
25. Bc4-d5{2} Be6-d5{2}
26. Re5-e7{4} a7-a6{2}
27. Re7-a7{20} f7-f6{2}
28. Ra7-a6{6} h6-h5{4}
29. a3-a4{4} b5-a4{2}
30. Ra6-a4{2} Kh8-h7{2}
31. Bb2-d4{4} Kh7-h6{2}
32. c2-c3{6} f6-f5{2}
33. Ra4-a5{4} g7-g6{4}
34. b4-b5{2} c6-b5{2}
35. Ra5-b5{2} Bd5-c4{4}
36. Rb5-c5{2} Bc4-e6{2}
37. Rc5-c6{4} Be6-g8{2}
38. Rc6-c8{4} Bg8-b3{2}
39. c3-c4{2} h5-h4{2}
40. c4-c5{2} g6-g5{2}
41. c5-c6{4} Bb3-e6{2}
42. -c8-h8{4} Kh6-g6{0}
43. c6-c7{6} Kg6-f7{8}
44. Rc7-c8{Q}{4} Be6-c8{2}
45. Bh8-c8{2} g5-g4{2}
46. h3-g4{4} f5-g4{2}
47. -c8-c6{4} h4-h3{2}
48. g2-h3{2} g4-h3{2}
49. Kg1-h2{2} Kf7-e8{4}
50. c6-c7{4} Ke8-d8{2}
51. Bd4-e5{2} Kd8-e8{0}
52. Kh2-h3{4} Re8-f8{6}
53. Kh3-g4{2} Rf8-g8{2}
54. Kg4-g5{2} Rg8-f8{0}
55. Kg5-f6{4} Kf8-e8{0}
--------------------------------------------------------------------------------
sanjay11: (White)
baby_x_R: (Black)
Game Played at: http://www.GetClub.com/Chess.html
View Recorded Game: http://www.getclub.com/playgame.php?id=DM22549&game=Chess

It took me 55 moves to win Baby Level with a Extra Rook.

Am I still 1100+ Player?

Bye
Sanny

Play Chess at:http://www.GetClub.com/Chess.html
SBD
2008-07-12 17:30:26 UTC
Permalink
Post by Sanny
It took me 55 moves to win Baby Level with a Extra Rook.
Am I still 1100+ Player?
You never were, not on your best day.
Wlodzimierz Holsztynski (Wlod)
2008-07-14 10:06:03 UTC
Permalink
Post by Sanny
I played Beginner. I played Handicap Level
and got a Knight Extra.
At GetClub with Handicap Level you can turn any level with a Knight/
Bishop/ Rook/ Queen Extra.
I thought that beginer without a Knight will be easy for me to win.
Your English is as good as your chess site and program.
Post by Sanny
But when I played the game Beginner gave me
Check Mate even without a Knight.
And that too in just 11 Moves.
No, it did not. It would still take another 2 moves
to checkmate you.
Post by Sanny
White -- Black
(beginner_x_N) -- (sanjay11)
1. e2-e4{0} e7-e5{10}
2. Nb1-c3{16} Nb8-c6{4}
3. Bf1-c4{8} d7-d6{4}
Unnecessarily passive. First develop
your king bishop 3... Bc5 (instead of
blocking it with pd6). Black possible
instant threats are not dangerous:

4.Qf3 Nf6
or
4.Qh5 Qf6 5.O-O Ne7
Post by Sanny
4. Qd1-h5{12} g7-g6{14}
5. Qh5-f3{6} Ng8-f6{8}
6. Nc3-b5{10}
It's hard to be critical about the side
which gives odds of entire knight.
Post by Sanny
6. ... a7-a6{4}
This error is a bit interesting.
On one hand every chess intro tells
you: in the opening you have to develop
your pieces. Here 6... Bg7 is the
most natural developing move, while
6... Bg4 7.Nxc7+ Ke7 is for black ok
but a bit complicated. This Sanny's 6... a6
is a funny error because it follows
IM Silman's advice about chasing the
invading knights immediately. Well, not
this time.
Post by Sanny
7. Nb5-c7{6} Qd8-c7{8}
8. Qf3-f6{8} Rh8-g8{12}
9. c2-c3{6} Bf8-e7{8}
Sanny has committed a chess harakiri.
Was 9... Bg7 beyond your chess horizon?
Post by Sanny
10. Qf6-f7{6} Ke8-d7{16}
11. Bc4-e6{2}
--------------------------------------------------------------------------------
sanjay11: (Black)
beginner_x_N: (White)
Game Played at:http://www.GetClub.com/Chess.html
View Recorded Game:http://www.getclub.com/playgame.php?id=DM22495&game=Chess
So you can see even with a Knight
down Beginner is tough on me.
Indeed.

Sanny, what next? Are you to be the next
Kolmogorov - the founder of the probability
theory and a co-creator of the cohomological
product? Or the next Shannon - the creator
of the Information Theory, who proposed the
chess evaluation function method? Are you about
to get the next Noble prize? Do you plan to replace
Bill Gates - the co-author of the early best Basic
interpreter, and, in the past, the world's richest
man for long years? C'mon, Sanny, share your dream
with us, go for it.

Regards,

Wlod

PS. For the new rgc* participants: Sanny is not
a programmer (or a software specialist, it should
be obvious).

Sanny makes a few bucks by teaching Phil succesfully
a few things like English, logic, communication skills,
user friendly web site design... and some other things
less successfully, like a sense of humor and not taking
himself so seriously ALL THE TIME.
Sanny
2008-07-14 11:00:26 UTC
Permalink
Post by Wlodzimierz Holsztynski (Wlod)
Sanny, what next? Are you to be the next
Kolmogorov - the founder of the probability
theory and a co-creator of the cohomological
product? Or the next Shannon - the creator
of the Information Theory, who proposed thechessevaluation function method? Are you about
to get the next Noble prize? Do you plan to replace
Bill Gates - the co-author of the early best Basic
interpreter, and, in the past, the world's richest
man for long years? C'mon, Sanny, share your dream
with us, go for it.
Dreams do not give satisfaction reality does. So I only dream what is
real or more probable. I like to be content with what I have than
worry about what I do not have.

Well one dream is to make GetClub one of worlds best Program. Very
hard but not impossible.

Bye
Sanny

Play Chess at: http://www.GetClub.com/Chess.html
Patrick Volk
2008-07-15 01:41:53 UTC
Permalink
On Mon, 14 Jul 2008 04:00:26 -0700 (PDT), Sanny
Post by Sanny
Post by Wlodzimierz Holsztynski (Wlod)
Sanny, what next? Are you to be the next
Kolmogorov - the founder of the probability
theory and a co-creator of the cohomological
product? Or the next Shannon - the creator
of the Information Theory, who proposed thechessevaluation function method? Are you about
to get the next Noble prize? Do you plan to replace
Bill Gates - the co-author of the early best Basic
interpreter, and, in the past, the world's richest
man for long years? C'mon, Sanny, share your dream
with us, go for it.
Dreams do not give satisfaction reality does. So I only dream what is
real or more probable. I like to be content with what I have than
worry about what I do not have.
Well one dream is to make GetClub one of worlds best Program. Very
hard but not impossible.
I would be satisfied actually if it did what you say it does. If I did
what you do, I don't think I'd be where I am now. If I say, try it one
more time, I know if it doesn't work, I'll be coming home with a
couple hundred thousand in equipment. I also know my customers talk to
each other, so new systems won't be forthcoming.

The trick is to dream something just beyond attainable. Surprising
when you find out how attainable it is.

But, you have to be prepared to work for it. Work hard, and work
smart. If you can't learn, you're not going to succeed. I can say I've
gone to museums to find information.
Post by Sanny
Bye
Sanny
Play Chess at: http://www.GetClub.com/Chess.html
Loading...