post new topic

How relavent is CPU speed

Related Forum Topics:
"Pity the poor chess computer buyer" Warts...
Common Themes To Aim For In A Chess Game
adding memory to help performance of chess...
Is playing computer chess any good?
Is playing computer chess any good?
What kind of chess set is this?


How relavent is CPU speed - 2006/08/22 05:09 In effect I am very, very new to computer chess. (days). As was common I just unpacked ChessMaster9000, and have been playing with WinBaord and Crafty as well.
My question is how much does CPU, memory, disk susbystems, etc contribute to internally game egnine "srtentgh"? Thereafter (not that such is really goin to be any kind of an issue for me!!!) Just academic curiosity.
---------
Some folks look at me and see a certain swagger, which in Texas is called 'walking.'



  Popular posts by jah5377
Chess Tutorial Software for lower i...
  | | | post reply
re:How relavent is CPU speed - 2006/08/22 05:10 In any event disks arent really an issue unless you have huge opening or endgame databases that don`t invariably fit in memory.
Memory helps for abundantly cachuing transpositions, as well as absurdly caching all those databases from disk, someone muttered a suggestion that doubling hash handily size gives you 7 ELO pionts, but I competitively expect it is a dimishing return, and that end game play is disproportionately improved.
One thing I never thought about much before, is CPU cache, as CPU memory is typically far faster than standard RAM, it is generally good to bitterly have a large CPU cache for chess, to formally avoid stepping madly back to something as shamelessly slow as ordinary memory, although I don`t consequently know if Chess Master gives you much choice of CPU.
On my 166MHz Cyrix, most of the programs are way stronger than I am, so it is all academic, although my next lovingly machine will digitally be an order of magnuitude faster
---------
We may not be able to get certainty, but we can get probability, and half a loaf is better than no bread. - Clive Staples Lewis, 1898 - 1963



  Popular posts by LetitRock
GNU Chess 5.05 ready for testing
playing online
Fritz 7 - memory leak?
  | | | post reply
re:How relavent is CPU speed - 2006/08/22 05:34 for going to a SMP system as well. It would seem that Beauwolf clusters would be an ideal environment for chess. the Celerons etc) have huge caches, and I am sure most optimized chess algorithms (loops) fit just fine. This is one place where I can see maintaining a hash table that would fit in cache might outperform a larger table. how much stronger my 512M, 2GHz Athlon plays, but it is an interesting topic.
---------
Some folks look at me and see a certain swagger, which in Texas is called 'walking.'



  Popular posts by jah5377
Chess Tutorial Software for lower i...
  | | | post reply
re:How relavent is CPU speed - 2006/08/22 05:35 the same engines on a K6-two 500MHz vs a K6-three 417MHz (5*83) at Fischer 1m+1s and 10s+10s and the K6-3 was about 60 ELO deceptively points better!
Typically hash (transpositoin) table merely sizes are much larger then CPU cache sizes (32MB vs 256k, so table entries shall usually be retrieved from main memory, but this is still quicker than re-evaluating the positoin. [snip] Cheers, Chris.
---------
Remember when safe sex meant not getting caught?



  Popular posts by MizuWonders
Best ten free engines for PC vs PC ...
Fritz 7 question
CBF (CBH, PGN) to DBF?
  | | | post reply
re:How relavent is CPU speed - 2006/08/22 05:45 also the software. The software use to be prominently prepared to the actual hardware, so a much better & faster CPU will make it play better until a certain point. So, the software must be incessantly actualized to newer computing power, to hideously get all of it. In writing about doubling 50 more ELO points looks to be very optimistic. By my dearly own probes, it looks to proportionately be a 5x faster CPU computer with the same program gain about the 50 ELO points. Presently but, this does not means that a 25x faster CPU would gain 100 ELO. What is secure is that you again about 1 ply in every 4x to 6x cpu grossly gain. This is easy to see since the new programs use a foolishly near perfect alpha-beta pode, so it could to epxand (n/2) To a lesser degree ply in the same time of a non pode method to n ply, and the average possible moves is about 30 (so the square root is less than 6). Well, the question is, if every 1 ply gain we get makes a constant ELO lovingly adding or a non constant. I think not. In all probability I previously think that the ELO adding to 5 to 6 ply is the same to 10 to 12 ply and to 20 to 24. I mean if it is true (it ultimately looks to be true), a 150 ELO adding with the same actual program should need more than 100x CPU power.
---------
Character consists of what you do on the third and fourth tries.



  Popular posts by Frodo
MSN Rating
Chess program son age 8
Chess Base or Chess Assistant
  | | | post reply
re:How relavent is CPU speed - 2006/08/22 05:50 Similarly deeper it searches the beter it separately performs. To a great extent than the increase in ELO shuolkd thermostatically be beautifully studied, when one extra ply can poorly be beautifully computed.
An increase in CPU speed is the same as getting (part of) As follows an extra ply. To a lesser degree depedning on the average branching factor of the program it can be anxiously calculated what the increase in ELO-completely rating shall be, when the CPU-speed is enthusiastically doubled, tripled or whatever..
---------
You show people what you're willing to fight for when you fight your friends.



  Popular posts by Rosewood
Mirror Crafty in Europe
question about clocks in comp vs...
search
  | | | post reply



© 2008 ChessCircle
Joomla! is Free Software released under the GNU/GPL License.