Author: Robert Hyatt
Date: 21:30:51 04/25/01
Go up one level in this thread
On April 25, 2001 at 19:18:18, Vincent Diepeveen wrote: > >it shows regurarly 12(6) which would mean it has finished 18 ply fullwidht >according to Bob. > No, according to Hsu/Campbell/etc. >That means even with a perfect branching factor > >40 (first ply) ==> not taken into account though >4 (factor 4 for overhead because of using big qsearch) >let's take an optimal branching factor for deep blue with SE in software >and also captures added. > >optimal branching factor fullwidth 4.5 (it's way more without any extensions >my b.f. is even bigger fullwidth). >added to that 0.5 for SE where do you get that number? You don't do SE as defined by Hsu... IE PV-singular and FH-singular... >added to that 1.0 for capture extensions What is that? I don't do capture extensions. DB didn't do them. _nowhere_ will you find the concept "capture extension" mentioned in _any_ deep blue description, nor in any deep-thought description. This is simply fabrication. >total b.f. = 6.0 Just take the search times in the above output from DB and compute the_real_ effective branching factor rather than manufacturing a number that means nothing at all. Their branching factor can be computed using the search times they provided. I did it last year myself as I was interested... > >6.0 ^ 18 * 4 = 406239826673664 nodes needed. > >Ok let's forget even the factor 4 overhead for qsearch >then it's 101559956668416 > >Bit big number? I can make up numbers much larger... but they aren't relevant to the DB discussion... > >Oh well now let's take the number of nodes explored by deep blue. >200M nodes a second. Of that 20% was effective according to Hsu (see IEEE99 >publication). > >So compared to a single cpu that would be 40M nodes a second. >Let's even be MILD here for Hsu and ignore the fact and just use 200M >a second, because Bob is going to complain. > >So even when taking the defense of DB into account the >b.f. was absolutely better as > >18th square from 200M x 3 minutes = 18th sq from 36B nodes. >that's about 3.86 > >Now that's *impossible* without nullmove everywhere or very dubious >forward pruning. Note the example of Hsu he takes a 'typical 12 ply search' >as example. I love it when someone says "that is impossible". Someone said that to NASA in the 1960's... Someone said it to Lindberg, someone said it to Hillary before he went up Everest. Etc. They were _all_ wrong. It would be more correct to say "I don't know how to do that"... > >More likely this 8(6) says nothing and it's simply 8 plies in total from >which 6 in hardware. To solve many tricks he did the junior/belle trick >and he just extends in software about every tactical move he sees. Sorry... but speculation gets you nowhere. DB's search was well-described in two JICCA papers plus in the various talks and articles written/given since 1997. They didn't just extend every tactical move. > >That solves of course testsets bigtime. > >Also on the internet Campbell was recently searching during the >world championship FIDE with deep blue junior (i do not know how >many processors 30 or so?). >After many minutes of search he reported his variation was >a 12 ply search. So? Murray and Hsu almost always reported the software depth because that is the part of their search that is "interesting" and produces PVs. Hsu started doing the 12(6) with late deep thought and into deep blue to provide more information because early versions did a fixed search in hardware while later versions varied the hardware depth within certain bounds to balance the hardware speeds. > >Now that was like after a lot of time. And with the huge b.f. of >Deep Blue at those depths (as SE and recapture extensions >go take their toll on bigger depths in software), it's very likely >that Deep Blue only searched 1 ply deeper in a bit shorter time! > >So there is a source confirming this indirectly! Where is this "source"? Comes from the same mentality that once insisted the world is flat... because they could not see far enough to see the curvature.. > >He told this to channel 211 too, so there are more witnesses if they >still remember the questions i asked regarding search depth! > >Best regards, >Vincent He wasn't aware of the controversy that 12(6) was causing with a "few". If you think (like they did) that the hardware search was really just a fancier than usual 'evaluation' then just reporting the first number makes perfect sense. At least to me... Reporting it as 18 would be misleading since the last 6 had no SE. Reporting it as 12 would be misleading since there were another 6 plies done in the hardware. The only reasonable solution was 12(6).
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.