Author: Vincent Diepeveen
Date: 04:30:15 07/27/03
Go up one level in this thread
On July 26, 2003 at 17:22:02, Russell Reagan wrote: >On July 26, 2003 at 16:25:37, O. Veli wrote: > >>Since it is hardware, can >>we expect to be stronger than top software? > >I would expect it to be slower than top software, because cpu improvements >happen so quickly, and FPGA programming (from what I've heard) is not a simple >task. If he spends another two years working on it before releasing it (as >Slater said), just imagine how much faster the cpus will be by then. I completely agree here. Where it is fast in 2003 thanks to parallellism it will be very soon outdated IMHO. Note that he isn't using hashtables either in hardware. Though way better than deep blue, move ordering still is a joke of course. Deep Blue did it more or less at random. >If you're talking about something massively parallel like Deep Blue, that is one >thing, but a single PCI card? I doubt that is going to do any better than break What's faster: a couple of fpga cards without hashtables, or a software program at the 192 processor cluster that they got there? Each processor 3.0Ghz P4 or something. >even with top of the line hardware, so why bother? IBM threw so much hardware at >the problem that desktop cpu improvements wouldn't catch up for a LONG time, but that's bullshit of course. 1 brutus chip is massive faster than the shit IBM had. Also evaluation is a lot better. Further searched db too many plies in hardware. Also without nullmove. So the current chessprograms search within seconds same depth like deep blue. But of course milliseconds for depth of pc hardware at the time. You can't compare 1997 with 2003. I remember the many postings from hyatt at the time saying that nullmove is dubious. Where are they now? Brutus is about a factor 10 faster in hardware than deep blue effectively. Deep Blue had average of 133MLN nodes a second. Note that this is dubiously extrapolated guessed number. Even using *them* it means that in middlegame it had horrible little nodes a second for hardware. We're talking about 30 million or something and in endgames up to 300mln. 30 MLN at 480 cpu's means: 62.5k nps a cpu on average. Note this is EXTRAPOLATED numbers. The real number of nps from deep blue was of course A LOT lower. They just guessed something. But i am sure that Brutus will be a lot faster a cpu EFFECTIVE than deep blue. It is a shame how bad deep blue was in this respect. >a single PCI card doesn't seem to be worth the trouble of programming the thing, >because desktop/server cpus will probably outperform it before too 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.