One core , better than 4?

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Rebel, chrisw

bupalo
Posts: 82
Joined: Fri Mar 16, 2012 2:04 pm

One core , better than 4?

Post by bupalo »

Is it possible that the engines behave better with one core with 4 times calculating compare to 4 cores?

That's becuase with one core they see better the crtical moves
jdart
Posts: 4366
Joined: Fri Mar 10, 2006 5:23 am
Location: http://www.arasanchess.org

Re: One core , better than 4?

Post by jdart »

No, in general engines do not do better with one core vs. 4.

But once you start using multiple cores, you will see a lot of variability in search results. Sometimes you get a lot of speedup in finding a key move, sometimes you don't. It is random because the order in which threads are scheduled can affect the solution time.

--Jon
User avatar
tjfroh
Posts: 8087
Joined: Thu Mar 09, 2006 9:51 pm
Location: Near the Intel Plant in the Land of Manana
Full name: Timothy Frohlick

Re: One core , better than 4?

Post by tjfroh »

Well said Jon. I enjoy Arasanchess very much. I was under the impression that four cores would, in general, give a threefold increase in NPS. The more cores would yield less improvement per core. Most important is a well-written program. TJF
jdart wrote:No, in general engines do not do better with one core vs. 4.

But once you start using multiple cores, you will see a lot of variability in search results. Sometimes you get a lot of speedup in finding a key move, sometimes you don't. It is random because the order in which threads are scheduled can affect the solution time.

--Jon
A New Way Comes Upon Earth.
God is an infinitely variable Constant.
Man marks his ground with ideologies.
Galaxies are the dreidels of God.
War is a punishment for implacability.
Peace flows from forgiveness of sins.
User avatar
hgm
Posts: 27794
Joined: Fri Mar 10, 2006 10:06 am
Location: Amsterdam
Full name: H G Muller

Re: One core , better than 4?

Post by hgm »

I think you missed a critical part of his question:

1 core with 4 times calculating
bupalo
Posts: 82
Joined: Fri Mar 16, 2012 2:04 pm

Re: One core , better than 4?

Post by bupalo »

i can post results where stockfish one core performs better than 4 cores
User avatar
Laskos
Posts: 10948
Joined: Wed Jul 26, 2006 10:21 pm
Full name: Kai Laskos

Re: One core , better than 4?

Post by Laskos »

hgm wrote:I think you missed a critical part of his question:

1 core with 4 times calculating
Yes, I think he missed that, and the reply is that 1 core 4x time calculating on average is definitely better than 4 cores 1x time calculating.