[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Bug-gnubg] Re : Re: Optimal settings for MacBookPro

From: Louis Zulli
Subject: Re: [Bug-gnubg] Re : Re: Optimal settings for MacBookPro
Date: Mon, 29 Nov 2010 16:15:26 -0500

On my 8-core Mac Pro (8 physical cores---16 with HT) I get 410x10^6 with 7 eval threads and only 354x10^6 with 8 eval threads. 

Perhaps this is due to the fact that these are just eval threads, in addition to the main gnubg thread? So when I choose 8 eval threads, gnubg is actually using 9 threads, and even with HT enabled some efficiency is lost?

One other remark---in order to build gnubg to use sound in OS X, one needs to build against the Quicktime frameworks, which remain 32-bit. So the entire gnubg binary is forced to be 32-bit. While this has no effect on the discussion above, I wonder what speed up I'd get with a 64-bit build. I wonder if I can configure with --disable-sound and build a 64-bit binary? I think all the dependencies are universal 32/64-bit binaries on my system.

On Nov 29, 2010, at 1:11 PM, pierre zakia wrote:


I did the test again with only X11 running (no Airport, no Ethernet connection). The results are:
1 thread: 52 000 000
2 thread: 90
3 thread: 104
4 thread: 40 
5 thread: 40 
6 thread: 39 
7 thread: 39
8 thread: 39

3 thread is still the winner. Why ?


I am running the build 0.9.0 (downloaded from gnubg.org) that was compiled for OSX Leopard (in June 2008)

As I wrote, it runs under OSX Snow Leopard, but I never had any issues with this build like crash, freeze, blurred 3D or something else.


De : "address@hidden" <address@hidden>
À : address@hidden
Envoyé le : Dim 28 novembre 2010, 18h 01min 20s
Objet : Bug-gnubg Digest, Vol 96, Issue 11

Send Bug-gnubg mailing list submissions to

To subscribe or unsubscribe via the World Wide Web, visit
or, via email, send a message with subject or body 'help' to

You can reach the person managing the list at

When replying, please edit your Subject line so it is more specific
than "Re: Contents of Bug-gnubg digest..."

Today's Topics:

  1. Re: Optimal settings for MacBookPro (Philippe Michel)


Message: 1
Date: Sun, 28 Nov 2010 14:20:43 +0100 (CET)
From: Philippe Michel <address@hidden>
Subject: Re: [Bug-gnubg] Optimal settings for MacBookPro
To: pierre zakia <address@hidden>
Cc: address@hidden
Message-ID: <address@hidden>
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed

On Sat, 27 Nov 2010, pierre zakia wrote:

> What are the optimal settings for gnubg installed on a MacBook Pro 15" Intel 
> Core i7 2.66 GHz (April 2010 model) ?
> The build is Version 0.9.0, running on Snow Leopard (10.6.5) without any 
> problem;
> I have played changing figures in Settings/options/others/Eval threads from 1 to 
> 10, guessing 4 will be optimal. 
> But strangely enough, I got the best figure in the evaluation speed for 3, 
> larger than 110 000 000  (better than with 1 or 2) and plummeting to 40 000 000 
> with 4.
> Any clue ?

I would have guessed 4 as well, and this is what I get on a similar 
configuration (dual core with hyperthreading, running linux) :
1 thread  44000000
2        87
3        96
4        121
5        95
6        106
7        109
8        115

Maybe there was something else running on your machine that was hogging 
one thread, but if this is the case your decrease for the 5th active 
thread is much more dramatic than mine.

> What is the optimal figure to put in the Cache Size box ?

The default should be fine for anything but "long" jobs like analyzing 
matches at 4ply or long rollouts. For these it is useful to increase it 
but it won't make a huge difference. On the other hand, you probably have 
plenty of memory so increasing the cache to the maximum available in the 
GUI is almost free.

> Any other default settings I should change ?

Not really a setting, but since it looks like you built it from the 
sources, I found that compiling with the -funroll-loops option helps. This 
was with gcc, though, not Apple's clang.


Bug-gnubg mailing list

End of Bug-gnubg Digest, Vol 96, Issue 11

Bug-gnubg mailing list

reply via email to

[Prev in Thread] Current Thread [Next in Thread]