Author: Lar Mader
Date: 18:59:32 06/26/05
Go up one level in this thread
BTW, here's the info from Hydra's website: http://tournament.hydrachess.com/ahydra.php -= Lar On June 26, 2005 at 21:57:33, Lar Mader wrote: >Hydra is so amazingly strong, not so much because the program is superior to >other programs, but because the hardware is far superior to the standard PCs >that today's commercial programs run on. That being said, the programming is >indeed interesting and somewhat different from PC programs in that it is >specially designed for a fast multiprocessor search that can take advantage of >specially programmed FPGA hardware for search extensions and evaluation. Also, >because some of the chess evaluation is implemented in the FPGA hardware, it is >able to perform evalations that would be too expensive (with regard to >time/speed) for pure software engines. > >I think a lot of people would say that on some idealized equivalent hardware >setup, Shredder 9.0 is a stronger engine than Hydra, from a pure engine >standpoint. But this comparison is pretty loose: it isn't really fair or >accurate. It's very apples and oranges. For many reasons, such as those >mentioned above, you really have to consider the entire machine, the hardware >running the engine, how the hardware resources are designed, and how well the >software is optimized for this hardware, etc. > >Keeping the above in mind, which means that pure numbers don't tell the whole >story, it's interesting to note: the Hydra team claims their program considers >200 million moves a second. Shredder 9, on a very fast dual CPU PC probably >would be doing well to hit 2 million positions per second. Maybe you can get 4 >million/sec on a quad. So you might say that Hydra is roughly 50 times as fast >as Shredder (running on some pretty fancy PC hardware). Again, this doesn't >accurately tell the story of how the strength of the two machines/engines would >compare, but it is impressive. > >I don't think it's any surprise that Hydra is exceptionally strong. > >-= Lar > > >On June 26, 2005 at 21:29:30, K. Burcham wrote: > >> >> >>I assume Chrilly wrote the Code for the Nimzo program. >>If this is true then what were his results from year to year when working with >>chess code against his competitors, Ed, Robert, etc... >>Was Chrilly winning some of the time, most of the time? >>How many times was Nimzo world champion? >>Is Chrilly an IM, GM..? >> >>How come all of a sudden this Chrilly has a program that some say beats >>shredder, all of the worlds top programmers, and now can beat the number 7 super >>GM in the world. You think people in any other sport can all of a sudden beat >>the worlds best, when they have proven in the past they do not have world class >>talent. >> >>Of course if someone has the stats to show me that Chrilly was already world >>class programmer and was winning against world class programmers---then I will >>have learned about Chrilly. I am like many here, I do not know. So I would like >>for some of the chess program history buffs to fill me in on this guy that is >>kicking ass at the top. >> >>Also, now that Chrilly is on top, Frederic would you please tell Chrilly to comb >>his hair---if it will not comb---have him use some hair gel. >> >>kburcham
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.