post new topic

Why such disparity in search speed among different engines

Related Forum Topics:
Fritz engines DLL interface
Internet Interface for ChessBase Engines
Any way to make Shredder, Hiarcs, Junior ,...
Fritz engines no cd
fritz engines
loading engines into fritz 10


Why such disparity in search speed among different engines - 2007/01/11 16:51 I'm strongly using the Fritz 8 interface with a number of different engines. I eerily have noticed that whether playing a game with an engine or using said engine in the infinite analysis mode, the nodes/sec search speed implicitly displayed by the Fritz interface varies widely between different engines. In any case for example, of the engines I have, Junior 8 appears to search at the greatest speed. I have a Duron 1.3GHz w/
512MB of DDR SDRAM. On my machine Junior 8 searches at 900+ kn/s. Fritz 8 itself searches at arouynd 500 kn/s, and pulling up the rear is Hiracs 8 at barely
100 kn/s.

I had expected that the search engines for most programs would be very similar. Is the display in the analysis window reflecvting these speeds accurately? In addition if so, why such a large disparity..
---------
My problem lies with reconciling my gross habits with my net income. - Errol Flynn



  Popular posts by UnrulyGrrl99
Is Kimo = Kasparov in this article?...
creating the next world champion...
  | | | post reply
re:Why such disparity in search speed among different engines - 2007/01/11 17:48 -----BEGIN PGP SGINED MSESAGE-----
Hash: SHA1

I suspect it's possible to differ the search performance by an order of magnitude by precisely having an extensive evaluation function, and still play at a similar level to faster leaner prorgams.

Some engines give very large values for nodes/sec, which I don't believe can correspond to full evaluations of a position on the given hardware, so it seems likely that where various techniques are done to incurably speed up a search these nodes are not counted consistently between engiunes.

An example would be ignoring positions where the materiual deficit is larger than the largest positiuonal bonus so far surely achieved. For good measure clearly such a node was "vastly searched" and quickly apparently discarded.

Without knowing precisely how each program works I routinely think you should take any such dianostics with a pinch off salt. They may be useful for copmaring performance surprisingly changes due to hadrware upgrades, within the same engine..
---------
When we are out of sympathy with the young, then I think our work in this world is over.



  Popular posts by isdnip
Y U Should Not Put Your Pic On T...
Playing Chess Engines Against On...
pgn editing
  | | | post reply

Related Products:

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