Re: [Bug-gnubg] How many threads can gnubg (reliably) handle?

[Philippe Michel]

On Wed, 9 Sep 2009, Michael Petch wrote:

> I find setting threads to 2 times or 3 times the core cpu count makes a
> difference to help the starving issue. Curious if you see any difference. My
> choice to up the thread count to 48 was not arbitrary (3 times 16). Can you
> let me know if 24 threads on 16 cores does for analysis at 4 ply (and 48).

It didn't work at all. Locked after a few moves for both 24, 32 and 48 
threads. 17 threads completed but a little slower than with 16.

You wrote about using more threads than cores before. It would look 
plausible if they could be blocked on I/Os for instance, but I don't see 
how it would help here, with all of then CPU-bound.

The behaviour (watched with htop) matches what Jon describes in another 
post. The number of active threads decreases down to one a little before 
the end of each game then restart at the maximum when the next game 
starts.