Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Speedtest with Celeron & Celeron with cache

Author: Mark Young

Date: 03:16:22 10/06/98

Go up one level in this thread


On October 06, 1998 at 05:58:36, Paul Petersson wrote:

>On October 06, 1998 at 02:20:07, Mark Young wrote:
>
>>On October 06, 1998 at 01:44:15, Mark Young wrote:
>>
>>>On October 05, 1998 at 20:42:12, Paul Petersson wrote:
>>>
>>>>Here are some results from testing two Celeron processors. I used the BT16
>>>>position and Rebel 8.
>>>>
>>>>
>>>>REBEL 8                              BT16
>>>>--------------------------------------------
>>>>
>>>>Celeron 266 @266 MHz (66*4)          5:29
>>>>
>>>>Celeron 266 @400 MHz (100*4)         3:39
>>>>
>>>>
>>>>Celeron 300A @300 MHz (66*4.5)       2:56
>>>>
>>>>Celeron 300A @450 MHz (100*4.5)      1:58
>>>>
>>>>Celeron 300A @460 MHz (102*4.5)      1:55
>>>>
>>>>Celeron 300A @504 MHz (112*4.5)      1:45
>>>>
>>>>It gets very hot at 504 MHz, and hangs after a while, but at 460 MHz and lower
>>>>it runs without problems.
>>>>
>>>>
>>>>Paul
>>>
>>>Some of the times you post seem a bit fast for the Intel chip. Did you use 4 mb
>>>of hash, and use the time to complete the full 9 ply search. Your 450mhz time of
>>>1:58 is very close to Ed's time of 1:44 for his 450mhz AMD K6-2 chip. If I use
>>>28mb of hash on my computer. I do a 9 ply search over 1 min faster then I do
>>>using 4mb. One of the reasons 28mb is faster is you do not find the key move
>>>Nxg2 so the program does not hang on that move for a long time, checking out the
>>>move to make sure it is good. Now at 60mb Hash I find the key move Nxg2 and it
>>>take me about 1 min longer to do 9 plys over 28mb hash and about the same as 4mb
>>>hash.
>>
>>Here is what I mean.
>>
>>4mb hash, time to finish 9 ply search 2:33, move found after 9ply search Nxg2
>>
>>28mb hash, time to finish 9 ply search 1:49, move found after 9ply search Nf5
>>
>>60mb hash, time to finish 9 ply search 2:35, move found after 9ply search Nxg2
>>
>>Rebel 8 running under windows. PII 400
>
>I did the test precisely as described at Ed´s homepage. 4 MB hash. At 460 MHz it
>finds Nxg2 in 1:42, and finishes at 1:55 after searching 19211713 nodes.
>
>Paul

I just wanted to check. Because the AMD is suppose to be 25% to 30% faster then
the intel chips at the same clock speed when running rebel. This est. then seems
to be a bit high. At least with the Celeron with cache.




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.