Als Antwort auf: / In reply to: Re: YABRL: Aristarch 4.37 scores slightly less than Aristarch 4.21 geschrieben von: / posted by: Uri Blass at 09 February 2004 08:53:41:
...I do not think that 1.1Ghz is not normal hardware.In addition to the draw topic from my other posting, I just saw that you use time control 5+2 on a Thunderbird 1.1 GHz. This is comparable to about 2+1 on normal hardware.
Under these conditions, Aristarch might not get above search depth 10, which is necessary for the search algorithm to work efficient. Aristarch is optimized for search depth of at least 9 or 10, better more. I am testing on a Athlon XP 2400 with 4+2.
Not all people hurry to buy the fastest hardware.
The best hardware that I use is AMD 1 ghz and I had to ask for a special operator to run movei on hardware that is almost twice faster.
In reply to Stefan's post:
It depends what is considered "normal" hardware, I personnally do not think that only top-of-the-notch hardware is normal.
With respect to the speed of this Athlon machine: I have also an Athlon XP 2400, which is, however (Crafty benchmark) for chess only 1.7 times faster, which is less than a ply (and the XP 2400 has optimal BIOS settings as well, so no misconfiguration). On the Athlon 1.1MHz Aristarch always reaches depths of 9 and 10 (mostly 10), later in the game then of course deeper.
From various test games I and others have played I also do think that Aristarch is one of those engines that are comparatively stronger at longer time controls, but this observation is already true for 4.21.
With your argumentation line, however, I conclude that 4.37/4.40/4.41 are even more optimised for fast hardware and/or long time controls than 4.21. Is this correct?
Thanks
Robert