Author: Jorge Pichard
Date: 10:28:23 07/12/02
Go up one level in this thread
On July 12, 2002 at 12:51:46, Joachim Rang wrote: >On July 12, 2002 at 12:39:39, pavel wrote: > >>On July 12, 2002 at 12:33:30, Jorge wrote: >> >>>On July 12, 2002 at 12:27:39, Jorge Pichard wrote: >>> >>>>I evaluated the opponents that Shredder and Junior 7 faced and it seems as >>>>Brutus had the best performance based on the level of opponents that is faced. >>>> >>>>Pichard. >>> >>>Brutus is Nimzo8- right? If it is, Nimzo8 is the strongest of all the Nimzos at >>>the moment. >>> >>>jorge >> >>We don't know if it is Nimzo. >>Most likely It is Nimzo ported on special-purpose hardware. >>I doubt He would write another commercial level program from scratch; one is >>hard enough ;). >> >>cheers >>pavs > > >No Brutus is completely written from the scratch. It has nothing to do with >Nimzo, neither on hardware (Brutus uses some special chip called FPGA), nor >software. It is only made by the same programmer, who probably uses his >experiences while programming Nimzo for programming Brutus. But the code is >completely different because of the different hardware. > >There is a nice interview with Chrilly Dollinger, but only in German: > >http://www.computerschach.com/sprechstunde/archiv/chrilly1d.htm I only understood the last part in english, probably somebody can translate the entire interview. Gian-Carlo Pascutto: Hello, Nimzo seems to be a program that is more focussed on a fast and deep search rather than an extensive evaluation. Chrilly: You are right. When I designed Nimzo 98, I felt that more speed than in Nimzo 3 is necessary. But I think in the meantime the pendulum has swung back. Gian-Carlo: With the Brutus project, will you use the ability to do nearly everything ‘for free’ on the chip to make the evaluation much more extensive? Or will you keep the chip simple, so it is both cheaper and can be made faster - even deeper search? Chrilly: The aim is to make a quite sophisticated evaluation. In hardware there is almost no conflict between speed and knowledge. So less knowledge is not significantly faster. There is the conflict between money and knowledge. I have now got a bigger chip (Virtex-1000). This chip is 2.5 times bigger than the chip I used in Paderborn. This chip is definitely big enough
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.