Author: Robert Hyatt
Date: 12:16:16 11/12/98
Go up one level in this thread
On November 12, 1998 at 15:03:08, Bruce Moreland wrote: > >On November 11, 1998 at 20:16:27, Bruce Moreland wrote: > >>Having said that neither of you has gotten the name right (it was neither "Deep >>Blue" nor "Deep Blue Prototype", it was "Deep Thought II"), I don't think it >>matters what they called it. >> >>The version of Ferret that played in that tournament ran on a machine that's >>like 5x slower than the one it runs on now, it didn't know what a bad bishop >>was, it used null move R=1 instead of R=2, it didn't really think on the >>opponent's time, and it was out of book after 1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. >>Ba4 Nf6 5. O-O Nxe4. > >It has been argued privately that the difference between the beast that played >Fritz and the best that played Kasparov is quite a bit larger than this, and I >have no reason to doubt this argument. > >bruce I believe you are correct. Fritz was played using the last "deep thought" chips... I believe Hsu said he was searching about 2-3M nodes per second or so. The last kasparov match was played using the second iteration of the deep blue chip (deep blue 1 was the first fab done at IBM for match 1) while DB 2 was the second fab done and this is the version with the much larger eval and new features... Speed differential alone was a factor of 200, but the eval was probably another 10X over that... If you look at DT-2 and DB-2, there is enough difference to not want to say much in comparison. IE the first Cray Blitz steamed along at 1,000 nodes per second in 1980. By 1983 we were doing 20K. By 1986 200K and in 1989 600-800K. And the program got better along the way, just like Ferret has. Someone talking about a game we lost in 1980 would be talking about a far different animal than what we were when we won the world championship in 1983 or 1986... And I'd hope such comparisons would at least include a disclaimer that "yes, we beat them in 1980, but the CB in 1980 was *nothing* like the CB in 1986." That would be the responsible way to write such an article...
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.