Author: Robert Hyatt
Date: 14:48:08 06/26/98
Go up one level in this thread
On June 26, 1998 at 06:19:01, Vincent Diepeveen wrote: > > > >On June 25, 1998 at 21:33:33, Robert Hyatt wrote: > >>On June 25, 1998 at 20:52:40, Vincent Diepeveen wrote: >> >>> >>>On June 25, 1998 at 12:15:09, Robert Hyatt wrote: >>> >>>>On June 25, 1998 at 08:07:53, Roberto Waldteufel wrote: >>>> >>>>> >>>>>On June 25, 1998 at 06:35:40, Vincent Diepeveen wrote: >>>>> >>>>>> >>>>>>On June 25, 1998 at 04:54:02, Roberto Waldteufel wrote: >>>>>> >>>>>>> >>>>>>>Here's another question about Hsu's chess chip. I seem to recall reading some >>>>>>>time ago that Hsu was considering a commercial release of his chip. Does anyone >>>>>>>know anything more about this? If the chip were to become available, how could I >>>>>>>use it in conjunction with a PC? would the fixed depth not be "out of sync" for >>>>>>>the speed of, eg a Pentium 333Mhz if it was designed to work with a >>>>>>>supercomputer, or can the fixed depth be adjusted to redo the balancing act in >>>>>>>the new environment? If it were possible, I would be very interested in >>>>>>>experimenting with this sort of hardware coupling. I assume that it would extend >>>>>>>the depth to which a program could search by something like 4 extra plies within >>>>>>>the same time. This would surely improve the strength of the PC ches programs >>>>>>>quite a lot! >>>>>>> >>>>>>>Roberto >>>>>> >>>>>>You see it wrong. If you do 4 ply searches without hash etc, then >>>>>>2.5 million drops quickly to say 300k nodes a second. >>>>>> >>>>>>So in fact you're playing against a kind of fritz5, which DOES search >>>>>>all leafs fullwidth, which gives you some extra tactics, so commercial against >>>>>>programs which are only tested at the same hardware and are only >>>>>>busy with outbooking and trying to finish the game by means of tactics, >>>>>>you beat with big numbers then, but it will play horrible. >>>>>> >>>>>>Vincent >>>>> >>>>>Hi Vincent, >>>>> >>>>>I'm not sure I understand - can you explain what yo mean. I do use hash tables - >>>>>would not the Chip have access to my system's RAM? And When you say it would >>>>>play horribly, why so? how could the chip make it play any worse than without >>>>>the chip? >>>>> >>>>>Best wishes, >>>>> >>>>>Roberto >>>> >>>> >>>>you have to ignore part of what vincent writes, because he is an "anti-deepblue" >>>>person from way back. A couple of key points: >>>> >>>>1. no a chess processor could not see your RAM. But they do have their own >>>>hash table memory. >>>> >>>>2. a PC program with one of these chips attached would be far and away stronger >>>>than any existing computer chess program running on a PC. A chess board that >>>>they used had 8 processors n it. A PC could use the same approach, and search >>>>at about 20 million nodes per second. >>>> >>>>I pointed out to Vincent last week here that he was not doing well playing >>>>against Crafty. He pointed out (factually, by the way) that Crafty has a >>>>hardware advantage with the 4 processor ALR. But he's going to have to take >>>>a stand on one side of the fence or the other. Either speed is important, and >>>>his note about me having faster hardware does explain why crafty wins most of >>>>the games played on ICC, or speed is not important and doesn't explain crafty's >>>>win/lose ratio vs the various DiepX programs on ICC. >>>> >>>>I believe in a combination of speed and intelligent. DB has that. A PC-card >>>>with one or more DB processors would *still* have that. And it would produce >>>>unheard-of performance figures for a PC-based chess program. Ala' Deep Blue >>>>Junior, which is basically a PC with a chess board (actually an IBM single-cpu >>>>workstation with a VME-bus chess card, so don't start hoping to buy one of those >>>>and plug it into your PCI bus. :) >>>> >>>>Bob >>> >>>You have to ignore most of what Bob writes about Deep Blue, as he is a >>>known pro-deep-blue authority. >>> >>>As i pointed out a 'smart' program like deep blue can never do much >>>knowledge within 10 clocks. This means that total depth of everything >>>is 10 clocks. >>> >>>Also first Bob wrote that Deep Blue had 1000 adjustable parameters. >>>As i pointed out 1000 adjustable parameters are hardly more than >>>piece square tables (already taking 12*64=768 adjustable parameters). >>> >> >>lets get the numbers right. Hsu has reported, repeatedly, that they have >>8,000 (roughly) positional "weights" that they can tune. This does not say >>that (say) 1,000 of those weights are not complex with several different >>features contributing to that term... > >First you report 1000 then 6000 now i see the new number 8000. > >What is it? I don't ever remember saying "1,000". I originally said "10,000". Later it turns out that Hsu and Murray have reported 8,000+ terms at several presentations they have made. So it is their number, and it is well-known. > >>And, if you think about a chess processor running at 24mhz, did you ever think >>"why is it running so slow? *Could* it be that they are adding up several >>hundred gate delays per clock, because they are doing a *lot* of work every >>clock?" Think about it before you start saying what can't be done in 10 clocks. >>I can do *everything* *you* do in *one* clock if I want to design such a chip. >>Because I can build a chip to emulate *any* finite state machine, which your >>program most assuredly is... > >You CANNOT do everything in 1 clock. >That's impossible and laughable that some people believe this. > >You CAN do some simple things in 1 clock. > >You can do masses of simple things in 1 clock, >but you cannot say things about previous evaluated knowledge >in 1 clock. > >That's impossible. such a stupid remark. Can you say "ripple adder"? Do you know what a ripple adder is? It is a device that adds two binary numbers in one cycle. Do you realize that it is *impossible* to add two binary numbers in one cycle, because you have to add 2 bits at a time, and then propagate the carry bit? Of course you can't do that in one cycle... *unless* the cycle is long enough to let the carry bits ripple up to the big end of the word *before* the clock edge falls to latch the result. You don't understand hardware. It's obvious you don't understand hardware, and you ought to stop trying to explain that which you don't understand. Their eval is based on the same sort of gate-delay design as a simple ripple adder... at clock 1, I can ripple add 1024 different integers two at a time, in parallel. at clock 2, I can ripple add the 512 sums, two at a time. At clock cycle 3, I can add the 256 results two at a time... you get the idea... And it is *not* impossible. > >You can mainly do things like piece square tables in 1 clock, that's >true. > >>>So when smoke about knowledge has been removed, we can go to >>>the supposed speed of it. >>> >>>Then it appears that from the few printouts i saw of Deep Blue, >>>that it just searched 11 or 12 ply. >>> >>>That's not much for a 200 million nodes a second program. >> >>my 12 plies is not your 12 plies is not their 12 plies. You should know >>that. How does your depth compare with genius, when genius says 4 plies? >>You get the idea. The only unfortunate thing is you have *never* had to play >>them to see what they can do. I have. *three* times. With a program that >>is so much stronger than Crafty, which is outplaying your program, and I was >>not just barely beaten by them, I was *drubbed* soundly... > >But my 12 ply is roughly their 12 ply, as i do a lot of threat extensions and >do checks in q-search and do pins in evaluation etc. > >Crafty DOESN'T. This means that even with a 15 ply search you don't >see tactically the same others see in 12. maybe not, but it seems to see enough to beat the snot out of certain programs I can name... And 1-0 and 0-1 are the only two measures of a chess game that I know how to understand... > >Some things NEED to be extended, and i'm not referring to 1 reply >check, which in my opinion sucks. > >But this is just my humble opinion, > >So is Fritz doing mating extensions, and Nimzo, and genius, and rebel, >and The King, and... ...do i need to continue (hiarcs i'm not sure but it >sees also way more). > >Perhaps there is a reason why they all do it? > >What i mean to say is: this is not a feature of Deep Blue, this is a WEAKNESS >of crafty. > >Especially at tournament level this kills crafty, although considering the >little depth we (crafty, diep) get in blitz, it might be smart not to do so in >blitz, >as long as we don't get a reasonable depth. it's not "killing me" that I notice... so I don't have a clue what you are rambling about here. Singular extensions are on my list to try again, I have the code from Cray Blitz so it's not a huge amount of work, it just hasn't popped to the top of my list yet. > >>>Now when we consider that it has MASSES of disadvantages a normal >>>chessprogram has at a general purpose processor, not to >>>confuse with single chips, then we see after some calculations that >>>those processors indeed are fast, but their practical speed compared to >>>PC programs is a lot less. >> >> >>define "masses of disadvantages". I see normal hashing, incredibly complex >>evaluation done in basically no time, instant MakeMove/UnMakeMove/Generate >>facilities... I'd love to have those same disadvantages... > >True they CAN do a lot in the same clock, but not EVERYTHING. >And they NEVER make it within 10 clocks, unless their evaluation really >sucks. You know that, i know that. *I* don't know that. But *I* understand digital hardware design. Fortunately, Hsu understands it too, and built that which you say can't be built. > >>> >>>Still more than PC programs of nowadays, so no doubt it'll kick some butt >>>in blitz, but i doubt whether normal users who likes to see some positional >>>insight >>>of a program as well will ever be happy with it. >> >> >>of course not.. a program that could beat Kasparov in even one game could >>hardly provide useful information to us, eh? >> >>> >>>Further i like to point at the fact that the technology used for DB >>>chessprocessors is very cheap. The salary of the PR people who were >>>needed to organize the event needed probably eated up the biggest part >>>of all that money, which IBM claims that the event Deep Blue-Kasparov took. >>> >>>Greetings, >>>Vincent >> >> >>your point would be? the "silicon" in *any* microcomputer is incredibly >>cheap to manufacture. It is the R&D that drives the cost up... not the >>silicon and fabrication. If a machine could design the next generation of >>itself, they would cost under a dollar to produce...
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.