Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: crafty faster on AMD however

Author: Robert Hyatt

Date: 20:39:26 09/27/02

Go up one level in this thread


On September 27, 2002 at 12:11:03, Vincent Diepeveen wrote:

>On September 26, 2002 at 15:11:23, Robert Hyatt wrote:
>
>>On September 26, 2002 at 12:28:16, Aaron Gordon wrote:
>>
>>>On September 26, 2002 at 11:16:06, Robert Hyatt wrote:
>>>>>  93.5 seconds base run time
>>>>
>>>>Yes...  But the Intel duals are blowing the AMD duals out of the
>>>>water, totally..
>>>>
>>>>AMD appears to win the "single cpu war" at the moment.  But on the
>>>>duals (and beyond) they are _way_ behind intel's performance.
>>>
>>>
>>>Where exactly are you seeing the dual AMD's being "_way_" behind Intel's
>>>dual P4 systems? Slates dual XP 1.73(2100+) gets 1.69 million nodes/sec in
>>>Crafty v18.11. A single 2.5GHz AthlonXP gets almost 1.6 million nodes/second.
>>>With the 2800+ being announced on October 1st and the Iwill MPX2 as cheap as it
>>>is someone could make a dual AthlonXP 2800+ box quite easily provided they take
>>>the 5 minutes to unlock the cpus.
>>>
>>>Show me some numbers Hyatt. :)
>>
>>Several have posted dual AMD numbers.  The relevant detail is this:
>>
>>run crafty with one cpu on an AMD box and record the NPS.  Then run it
>>using two cpus and record the NPS.  On an AMD box, the dual cpu speed
>>will be around 1.4X faster than the single cpu speed.  on an Intel box,
>>this ratio is 1.9X.  Which is a _significant_ difference.  IE with Crafty,
>>a dual intel gets 1.9x as much computing power (I am not talking parallel
>>search speedup here, only raw NPS numbers to compare computer horsepower)
>>while a dual AMD gets 1.4X...
>>
>>Eugene posted several such numbers here.  Vincent reported 1.4X on his
>>dual AMD.  At mhz for mhz, AMD is generally faster, using one cpu.  But
>>when you factor in that 1.4X vs 1.9X for duals, the Intel processor catches
>>up in a big hurry when using more than one cpu...
>
>depends upon how well a program is written obviously.

Perhaps the "better" written program runs faster on intel hardware?





>
>>Wasn't my results that were posted, I just noticed them.
>>
>>Others replied to that thread as well and AMD was always behind...



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.