Can Core 2 Quad really perform better than Core 2 Duo? if so,
how much better with same speed CPUs?
Let's say Core 2 Quad 3.6 GHz and Core 2 Duo 3.6 GHz, Is there any significant difference in performance?
how much better with same speed CPUs?
Let's say Core 2 Quad 3.6 GHz and Core 2 Duo 3.6 GHz, Is there any significant difference in performance?
The below benchmark data is a comparison, as we see Quad is 2x times faster than Dual core
That means,you can get extra 40-80 Elo points (depending on conditions) in favor for Quad
In other words,you will have approx.50-60 Elo extra more, in case of doubling the processor speed
Best,
Sedat
That means,you can get extra 40-80 Elo points (depending on conditions) in favor for Quad
In other words,you will have approx.50-60 Elo extra more, in case of doubling the processor speed
Hardware-Processor Speed Cores kN/s
Intel Core 2 Q9450 @ 3.60 GHz 4 10604
Intel Core 2 Duo E6600 @ 3.60 GHz 2 5001
Best,
Sedat
Sedat,
Thanks for benchmarks.
I am a bit non-technical, How come Core 2 Quad 3.6 GHz is 2X faster than Core 2 Duo 3.6 GHz?
What exactly Quad Core means?
10604 KN/s, what is it? Does same chess engine think 2x better?
Thanks for benchmarks.
I am a bit non-technical, How come Core 2 Quad 3.6 GHz is 2X faster than Core 2 Duo 3.6 GHz?
What exactly Quad Core means?
10604 KN/s, what is it? Does same chess engine think 2x better?
Duo and Quad is the numbers of Two and Four respectively, representing number of physical cores inside one CPU. Each core can be identified as independent CPU (but inside one physical CPU box and socket).
Quad (4) cores can work 2x times faster than Duo (2) given all the independent work they do is unique or it will be a waste evaluating same position twice or more times.
Quad (4) cores can work 2x times faster than Duo (2) given all the independent work they do is unique or it will be a waste evaluating same position twice or more times.
>Thanks for benchmarks.
>10604 KN/s, what is it? Does same chess engine think 2x better?
>How come Core 2 Quad 3.6 GHz is 2X faster than Core 2 Duo 3.6 GHz?
Not at all....
For example,10604 are kN/s speed values,which are generated by Fritz Benchmark tool
kN/s=kilo nodes per second
Of course, 4 core is faster and better for chess (than 2 core)
Simply because,Quad contains 4 cores,Dual contains 2 cores
Lets imagine...
Which Auto machine will be faster,if you drive:
60km per hour ?
or
120km per hour ?
Btw,nowadays is very popular (or best) the machines to drive with 30 km per hour

Name Elo + - games score oppo. draws
Houdini 2.0b Pro x64 6c 3418 17 17 1049 68% 3305 46%
Houdini 2.0b Pro x64 4c 3347 17 17 1030 51% 3344 47%
Deep Rybka 4.1 x64 6c 3358 15 15 1363 59% 3305 54%
Deep Rybka 4.1 x64 4c 3293 13 14 1602 47% 3314 56%
Fire 2.2 xTreme x64 6c 3311 23 22 530 52% 3297 63%
Fire 2.2 xTreme x64 4c 3275 16 16 1165 41% 3328 59%
Note:the above results are based on SCCT Auto232 Rating:
http://www.sedatcanbaz.com/chess/ratings/scct-auto232/
-----------------------------------------------------------------------
kN/s Mate Processor Speed Cores L.P EXE Hardware User
21416 12s Intel Core i7 980X @4.33GHz 6 OFF x64 Sedat Canbaz
10872 47s Intel Core i7 920 @3.00GHz 4 OFF x64 Sedat Canbaz
Note:the above hardwares are used for SCCT Auto232 Rating
For more Fritz/Houdini chess benchmarks can be found below:
http://www.sedatcanbaz.com/chess/fritz-chess-benchmarks/
http://www.sedatcanbaz.com/chess/1228-2/
http://www.sedatcanbaz.com/chess/houdini-2-0-benchmarks/
Regards,
Sedat
But what if program lets two or more different cores/processes to evaluate same line and positions at the same short short time?
10% of time? 30? I heard that Rybka is evading counting same positions twice so her nodes speed don't grow linear after increasing number of cores,
but just to about 1.7x speed each 2x cores
10% of time? 30? I heard that Rybka is evading counting same positions twice so her nodes speed don't grow linear after increasing number of cores,
but just to about 1.7x speed each 2x cores
Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill