Author: Anthony Cozzie
Date: 05:47:22 02/16/04
Go up one level in this thread
On February 15, 2004 at 15:07:25, Joachim Rang wrote: >On February 15, 2004 at 12:30:03, Bob Durrett wrote: > >>On February 15, 2004 at 11:47:04, Joachim Rang wrote: >> >>>On February 15, 2004 at 10:40:47, Peter Skinner wrote: >>> >>>>On February 15, 2004 at 09:38:33, Bob Durrett wrote: >>>> >>>>> >>>>>Chess commentary: >>>>> >>>>>The days of the software programmers are fast coming to an end. >>>>> >>>>>The days of the "hardware programmers" are just beginning. >>>>> >>>>>Bob D. >>>> >>>>Well truth be told, this is the first "good" result for the Brutus/Hydra team. >>>>Nothing can be concluded from just 7 games. It was the best at this tournament, >>>>but what is to say at the next one it only scores 3.5/7.0? >>>> >>>>Now flip the coin on a software program. Let's take Gandalf (which happens to be >>>>one of my favorite programs). It did not score as well as I had hoped, yet the >>>>minute the new version goes on sale, I will be purchasing it. >>>> >>>>7 games is really meaningless. It is a crap shoot that Hydra won, and next year >>>>can easily lose. Luck plays a big part in it. This is _exactly_ why the rating >>>>lists (SSDF, Chess Fun's rating list...) are so important. They show many games >>>>vs the same opponent, but also many games vs a wide array of opponents to attain >>>>a reasonable rating and top program and error margins. >>>> >>>>You must also factor in that Hydra is hardware based and whatever advantage it >>>>had today can be gone as soon as tomorrow. Software can evolve faster than >>>>hardware, thus any advantage can easily be erased with a few key strokes. >>>> >>> >>>sorry that I didn't put it in my first post, but here you err again. >>> >>>The prospects for FPGA-based solutions regarding to hardware-development are >>>_much_ nicer than to CPU-based programs. Together with that the >>>Hydra-FPGA-Solution is not a one-time card but can be updated several times: So >>>you buy the card one time and update the software for the card several times. >> >>Software on the card? That is terrible! When will they ever learn? >> >>That makes Hydra not much better than a mini-PC for chess. : ( >> >>Bob D. >> >most probably not software but you can flash it or something like that I don'T >know the details, but you can change things on the card. > >>> >>> >>>>Personally I have stopped playing the "hardware" game in regards to chess. If it >>>>is fast enough to run the programs I use, then it is good enough for me. I tried >>>>keeping up with the hardware that comes out, but it gets costly, and frankly I >>>>can spend my money elsewhere. >>>> >>>>I read in a survey on a tech website, and I think it was Anandtech that the >>>>average home user computer is still only a 1.5 Ghz P4 computer or below. Only >>>>12% were at a level where they had the most up to date hardware. So that leaves >>>>88% of the computing population that filled out the survey that use "sub-par" >>>>systems in every day computing in regards to chess. >>>> >>>>I would rather pay for excellent software than upgrade a cpu every 6 weeks like >>>>some do on the chess servers. >>>> >>>>The SSDF still uses an Athlon 1200 MHZ computer to do it's testing, and that is >>>>fine by me. It relates more to the hardware I use, and I can reflect that in the >>>>testing I do personally. >>>> >>>>Peter. It is an FPGA -> Field Programmable Gate Array. You can look up how they work online, its not really that complex. I disagree about its potential though - Hydra has hashing problems. anthony
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.