Author: Vincent Diepeveen
Date: 13:05:21 10/13/02
Go up one level in this thread
On October 13, 2002 at 15:55:05, Vincent Diepeveen wrote: hello, you can also see it from the other side. suppose the hardware search (which theoretical is impossible) would be a+b. In that case the only way to search is to start with: 1,2,3,4,5,6,7,8,9,10,11,12,13 ply. If you skip an iteration with so many extensions then you get dicked completely. If you look with me in game 1 the first output: 3(4) -25 T=0 Bc8g4 ph2h3 4(5) -25 T=0 Bc8g4 ph2h3 5(5)[Bg4](-22) -22 T=0 Bc8g4 bf1g2 6(5)[Bg4](-23) -23 T=0 Bc8g4 nf3e5 Ng8f6 ne5g4B 7(5) #[Bg4](-18) -18 T=1 Bc8g4 bf1g2 Ng8f6 nf3e5 Pe7e6 8(6) #[Bg4](-21) -21 T=1 Bc8g4 bf1g2 Ng8f6 o-o Pe7e6 ph2h3 9(6) #[Bg4](-12) -12 T=4 Bc8g4 nf3e5 Bg4e6 bf1g2 Pg7g6 o-o 10(6)<ch> 'g3' [0 sec (main.c:6581)] -12 T=13 Bc8g4 nf3e5 Bg4f5 pd2d4 Pe7e6 bc1f4 from 3(4) it jumps to 4(5) that's from 7 to 9 ply. And after a few seconds already it has 10(6) which would be 16 ply?? that's impossible to search in 16 seconds with 100 million nodes a second. Please follow knuth. impossible right. Look at the mainline length. At 5(5) it shows only a 2 ply mainline. If that's 5 ply in total from which say 3 ply in hardware (up to 5 ply maximum, but of course at 5 ply you don't get that) then it is logical to get a 2 ply mainline. If i search 10 ply from which 5 ply in software and i get a 2 ply mainline, then i cry murder, rape and FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF. But i am not happy. It is missing 3 ply then. Well as we know it is not missing 3 ply there. 6(5) Bc8g4 nf3e5 Ng8f6 ne5g4B the ne5g4 gets extended there it is an important capture. 10(6) Bc8g4 nf3e5 Bg4e6 bf1g2 Pg7g6 o-o up to 6 ply in hardware. Suppose you do just 4 ply in software. How do you get 480 processors *ever* to work then? Answer: NOT. A 4 ply tree in mainline is very small. Everything from hashtable. Answer: it was doing 6 ply searches there and had most searches like 4 ply searches in hardware. Makes sense. You can perhaps get a 50 cpu's or so to work with a 6 ply search. *at most*. No way to get 50 cpu's to work with a 4 ply software search. >On October 13, 2002 at 15:47:43, Jeremiah Penery wrote: > >>On October 13, 2002 at 14:14:13, Vincent Diepeveen wrote: >> >>>On October 13, 2002 at 13:50:24, Jeremiah Penery wrote: >>> >>>for any chessprogrammer it is very clear Jeremiah. >>> >>>It's 12 ply and the 6 only means that the hardware search can >>>be up to 6 ply within that 12 ply nominal. >> >>For the millionth time, explain 4(5) then. > >it's a lot easier when you try to imagine how you would start >a search. Look at 4 ply you want to involve other cpu's too. > >You have 480 chess chips. 9 out of 10 is idling anyway >(we know this because everyone says it was capable in theory >to get a billion nodes a second in theory, yet it got only >126 million nodes on average; there is no discussion here) > >you want to use as many cpu's as possible. So at the root you >ask a few chips. Of course you must take into account extensions too. > >If a 4 ply search for this chip is too much, then you split it. and >split it. Giving a 2 ply search to chip A, and a 3 ply search to >chip B. > >The 5 in short has no meaning here. > >You can see that the mainlines at the start of the game (so before >hashtables are filled with a lot of info that extends lines) >that the lines it gives at 4 , 5 , 6 ply is usually 1 ply long.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.