< Earlier Kibitzing · PAGE 104 OF 203 ·
|Feb-28-08|| ||Peligroso Patzer: <When does round 8 begin?>|
Thursday (28 February) at 10:00 a.m. EST.
|Feb-28-08|| ||hovik2003: This gigantesque numbers and calculation on these previous pages reminds me the legend about inventor of chess game itself.
The story goes after he presented his new designed game to the royal court( could be Lydian or Indian or maybe Persian), the king liked the game so much that he granted him the right to ask for anything he wished for, the inventor humbly asked instead of gold and jewlery to be paid off by wheat grains, in a manner of that in first square of chessboard 2 grains, in the second 4 grains, in the third 8 grains and so on to the last 64th square. After hearing his proposal king and every body in the court started to laugh at him and called him crazy and deranged.
But next day after royal mathematician of those days coudn't even calculate the exact number but estimated that king should give him all the wheat in the silos plus wheat and barley produced in the country for three thousand years thenafter.|
By todays factorial math it is easy to reach the solution:
Say 30 grains=1gm so 1 ton=30,000,000 grains
2^64/3^7 = 615 billion tons of wheat!!
Today's world output of wheat is about 1.3 billion tons annually, so just imagine!, and his royal highness learned a hard lesson about human mind's imagination and fantasy, and only thing funny in that day was their ignorance!
|Feb-28-08|| ||WillC21: <frogbert><pawnsgambit> I think, actually, frogbert's math is correct here, given that he is disregarding pruning efficiencies. Not sure what his final number is(too lazy to go back and look), but it's massively large per my calculations, even for a "monster" like Blue Gene. Let's assume Blue Gene's 135 trillion/sec calculations converts "neatly" to 135 trillion nodes(positions) per second (I know it's not meant for chess, but let's disregard that point and pretend it is). My math:|
Consider that the "average" snapshot of a board during a game may contain, for whomever color's move it is, 20 possible responses (maybe it's higher or lower, but 20 seems fair enough to me).
1) 30 ply = 20^30 total positions to examine.
2) Time in seconds for Blue Gene to process such positions: (20^30)/(135*10^12)
3) Time in years: ((2)Result)/(60*60*24*365)
Result = 2.52208*10^17 years
Or, expressed more plainly, the final result(3) would be over twenty-five sextillion years! Wow!
|Feb-28-08|| ||hovik2003: For your information "light year" is a astromomical unit for "distance" and not time. our whole universe is about 14 billion years old after BigBang, so logically speaking no creation or event could overpass the worped boundries of our space and time behind its limits. |
But any way I think if the software doesn't include written sources for repeative transpositions, and various positional and material algorithms and restrictions, even Blue Gene or any other monster made in next one hundred years will blow its boards to pieces trying to calculate all the positions(nods) arrised in 50 move chess game from start up to the end.
|Feb-28-08|| ||brankat: Most fascinating conversation!
Since I know nothing about algorithms or chess engines, haven't even heard of any such thing as "Blue Gene" until now, I can just wish the human Tournament resumes soon :-)
|Feb-28-08|| ||Thorski: <Pawnsgambit: I also have got Masters in Electrical Engineering from Top US university.>|
And here I was considering pursuing my graduate degree at a US school; surely a terrible mistake.
|Feb-28-08|| ||krivin: I remember reading 22yrs back about the number of games ( total) that could be played which ends in 30 moves in chess as 2 power 100.Funnily it was written that it will take so many yrs ( a very big number this too) to even play out all these games by the population of the world at that time.!!
If this is so with 30 move games.. what to say of those games ending in 40, 50 ..100 and so on?!!!
It looks it is inexhaustible...We may need infinite energy , infinite time to actually see all thses things in reality for one single human being..
An intelligent man will leave the numbers as it is and allow the mind to rest in peace for ever without breaking himself..
All your calculations will have only theoretical value.No single man or even machine will be able to make use of it fully..
So Chess is Infinite in Nature ..like Nature itself.|
|Feb-28-08|| ||Strongest Force: Any top GM or even ordinary master who has deeply studied the optimum ways in which a human can beat a top program clearly knows that a Rybca is most vulnerable when faced with a closed-position and pawns and pieces are rushing-towards the program's king. Many programers, like angry fathers, have hired top GMs to figure-out ways to keep their "children" from getting beat. The way to keep the comp-beaters at bay was to rotate different opening systems as much as possibleso as to keep the human off-balanced as much as possible.|
|Feb-28-08|| ||jedrus07: By the way do You know that PlayStation 3 runs at 2 teraFLOPS as wikipedia claims? I think Random visitor has roughly the same power (propably greater) in his pc and he gets 30ply in week... That means to get 60ply in week you need (2 teraFLOPS)^2 power... So... It is... 4000000000000 teraFlops... A bit higher then BlueGene so far... If You see any mistake, feel free to criticize.|
And now lets hope that second part of the tournament starts soon so we can discuss new games!
|Feb-28-08|| ||badest: The numbers are fascinating! However, in order to get somewhere one needs to have "better" algorithms (better being faster). This is to get to 30ply and beyond...|
Until P=NP is cracked, maybe one will settle with some type of probabilistic algorithms (in order to subdue the complexity of the problem).
|Feb-28-08|| ||minasina: <Peligroso Patzer: <When does round 8 begin?>
Thursday (28 February) at 10:00 a.m. EST.> |
Is it really at 10? I thought it should have been at 9:30 AM EST (GMT -5)? Well...maybe it is the time to start Chessdom commentaries: http://www.chessdom.com/morelia-lin...
|Feb-28-08|| ||squlpt: It is impossible these days to reach 30 ply in a relatively short period of time (week) without a decent amount of pruning. Take away the pruning and the fastest super pc today would not reach it before the sun burns up the earth in 4-5 billion years or so|
|Feb-28-08|| ||erniecohen: I wouldn't get too excited about quantum computing for chess. First, there are efficient quantum algorithms only for a few problems (factoring, linear search); for general NP problems, the best speedup now is only quadratic (which means that the best you can hope for is a doubling of ply, and that only in comparison to a program that doesn't consider transpositions). Second, the largest credible working quantum computer is only about 6 bits.|
I would wager that before quantum computers can compete with classical computers in chess, chess will have been weakly solved (i.e., we'll have a program that can provably force a draw with black), like checkers has. This might well happen in our lifetimes.
|Feb-28-08|| ||robertgida: Where is the official website of Linares part?|
|Feb-28-08|| ||vlado23: There's a good intro to the techniques chess engines use here: http://www.frayn.net/beowulf/theory...|
<Pawnsgambit: According to your math, it should take zillions of years to do a simple 30 ply analysis. That is unrealistic and impossible.>
Its not unrealistic, as others have pointed out a full-width of all positions to 30-ply would take a v.v.v. long time.
No chess engine does full-width searches beyond a few ply.
|Feb-28-08|| ||bgkuzzy: Now that we discovered water on Pluto, can someone tell me the live page?|
|Feb-28-08|| ||minasina: <bgkuzzy: Now that we discovered water on Pluto, can someone tell me the live page?>|
Sorry, I can't - yet? At the moment I only know these Chessdom commentaries:
with IM Miodrag Perunovic and Goran Urosevic
with GM Vladimir Dimitrov
|Feb-28-08|| ||child in time: <hovik2003> Just a small correction. He should've earned a bit more since he would've got the grains from the 63rd square and so on... Nevertheless, he wasn't humble in his request...|
|Feb-28-08|| ||jedrus07: I really like Shirov, but thats why I fell sorry for him already... I belive he would be able to beat Anand, but maybe its some kind of psychological barier that makes him his punchbag... Lets hope he can overcome that while playing in his country! I would like to see e4 e5 in this game and then sth different from sicilian and petrof. Do You have any predictions?|
|Feb-28-08|| ||zduric: <child in time>
1+2+4+...+2^63 = (2^64)-1
|Feb-28-08|| ||slomarko: <Strongest Force: Any top GM or even ordinary master who has deeply studied the optimum ways in which a human can beat a top program clearly knows that a Rybca is most vulnerable when faced with a closed-position and pawns and pieces are rushing-towards the program's king> its called Rybka not Rybca. btw <strongest force> how did you enjoy the Aeroflot tourney especially Caruana's "performances"? (that is losing to unknown IMs)|
|Feb-28-08|| ||0817: Does anyone know any site where one can watch all the games live? Thanks|
|Feb-28-08|| ||tamar: http://www.ajedrezmorelialinares.co...|
|Feb-28-08|| ||Joshka: What are the pairings??..who does Magnus play today?...do they just continue the rotation from the last 7 rounds?..thanks in advance|
|Feb-28-08|| ||minasina: Round 8 pairings:
Vishwanatan Anand - Alexei Shirov
Vassily Ivanchuk - Magnus Carlsen
Teimour Radjabov - Peter Leko
Levon Aronian - Veselin Topalov
< Earlier Kibitzing · PAGE 104 OF 203 ·