Would be rybka 4 stronger on 2 cores, or on 4?
2 Cores active cpu speed 2.53 GHz. 4 Cores active CPU speed 1.73 GHz.
2 Cores active cpu speed 2.53 GHz. 4 Cores active CPU speed 1.73 GHz.
2.53 GHz / 1.73 GHz = 1.46
2 cores => 4 cores is x1.6-1.7 so the 1.73 GHz 4-core is slightly better.
If you can overclock (but this is usually not possible on laptops), that calculation has to be redone.
2 cores => 4 cores is x1.6-1.7 so the 1.73 GHz 4-core is slightly better.
If you can overclock (but this is usually not possible on laptops), that calculation has to be redone.
Im waiting for Vas to respond in
http://rybkaforum.net/cgi-bin/rybkaforum/topic_show.pl?tid=18237
So assuming by doubling the cores rybka 4 only gains 86.6% speed up then we can calculate as follows:
1. 2 processors at 2.53Ghz each will act as 1 processor at [2.53+2.53(.866)]Ghz= 4.72Ghz
2. 4 processors at 1.73Ghz each will act as 1 processor at [1.73+1.73(.866)+1.73(.866)^2]Ghz= 4.52Ghz which is slower
Now if 70% is the actual speedup instead of 86.6% then
1. We get 4.3Ghz
2. We get 3.79Ghz which is a lot slower
So the 2 core engine is faster than the 4 core one.
http://rybkaforum.net/cgi-bin/rybkaforum/topic_show.pl?tid=18237
So assuming by doubling the cores rybka 4 only gains 86.6% speed up then we can calculate as follows:
1. 2 processors at 2.53Ghz each will act as 1 processor at [2.53+2.53(.866)]Ghz= 4.72Ghz
2. 4 processors at 1.73Ghz each will act as 1 processor at [1.73+1.73(.866)+1.73(.866)^2]Ghz= 4.52Ghz which is slower
Now if 70% is the actual speedup instead of 86.6% then
1. We get 4.3Ghz
2. We get 3.79Ghz which is a lot slower
So the 2 core engine is faster than the 4 core one.
I have a 2 core machine and a 4 core machine. With just rybka(not deep rybka) from what I understand it only uses one core but my quad easily increases in chess engine points compared to the dual core when playing on the server. The deep rybka is supposed to utilize the dual and quad core processors.
Each processor for your quad could be faster than each processor for your dual. Look at the average kn/s for the starting position to be sure (not the Ghz listed number which is often misleading).
Or you can do a fritz benchmark with only 1 core enabled for each computer. Also, one must play a tremendous amount of games before coming to such conclusions. For a low amount of games there is still a good chance that a chess entity will outperform another of equal strength.
Or you can do a fritz benchmark with only 1 core enabled for each computer. Also, one must play a tremendous amount of games before coming to such conclusions. For a low amount of games there is still a good chance that a chess entity will outperform another of equal strength.
Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill
