Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Which version of Crafty played in Maastricht?

Author: Slater Wold

Date: 11:37:06 08/27/01

Go up one level in this thread


On August 27, 2001 at 14:07:55, Robert Hyatt wrote:

>On August 27, 2001 at 14:03:16, Slater Wold wrote:
>
>>On August 27, 2001 at 13:17:09, Robert Hyatt wrote:
>>
>>>On August 27, 2001 at 12:41:10, Slater Wold wrote:
>>>
>>>>On August 27, 2001 at 10:46:20, Ulf Flörsheimer wrote:
>>>>
>>>>>In the notations it usually has been told of version 18.10x. What does x mean?
>>>>>Was it Crafty 18.10 or 18.10 with a special opening book? Or was it a unreleased
>>>>>version 18.11? If so, when will it be ready for download.
>>>>>
>>>>>I like both style and playing strength of crafty. Its Maastricht-games were
>>>>>really impressive. In my opinion Crafty is one of the most underestimated
>>>>>programs. It definitively prooved in Maastricht once more its legitimate
>>>>>placement among all the professional programs.
>>>>>
>>>>>I hope that crafty's further developement will be as successful as it has been
>>>>>in the past.
>>>>>
>>>>>Good luck for Bob Hyatt and a big "thank you" for the fun he has given to me
>>>>>with crafty!
>>>>>
>>>>>Ulf Floersheimer
>>>>
>>>>Bob said earlier, it's the _SAME_ exact version running on ICC now.
>>>>
>>>>However, it was on a 2x1.4Ghz AMD, which is a _LOT_ faster than his 4x700.
>>>>
>>>>So it packed a little more punch than usual.
>>>>
>>>>
>>>>Slate
>>>
>>>
>>>The 2x AMD is roughly the same as my quad 700.
>>>
>>>quad benchmark is 1.2M nodes per second (gcc/linux)
>>>dual AMD is 1.2M using MSVC/win2000.
>>>
>>>Actually, that makes my quad about 10% faster, since gcc is about 10%
>>>slower than MSVC...
>>
>>Weird.  My AMD 1.2Ghz gets about 800k nps.  I thought for sure a 2x1.4Ghz would
>>do better.  At least 1.5M.  That's the beauty of SMP!
>>
>>
>>Slate
>
>
>The one we had would only do 1.2M...  no idea if there are better ones
>around, of course...
>
>I don't quite follow the math to think that 2x1.4 would be more than 50%
>faster than 2x1.2 however...

Well, my thinking (estimation) was that a 1.2Ghz does 800k.  Therefore a 1.4Ghz
would more than likely do at least 900k.  900 * 2 = 1800 - 20% = 1.44M.

If you figure a 1.2Ghz is getting 800k, than that is 666 nodes per mhz.

666 * 1400 = 932400 * 2 = 1864800 - 20% = 1491840 (or right around 1.5M)


Slate




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.