Author: Jeremiah Penery
Date: 00:17:26 05/29/02
Go up one level in this thread
On May 28, 2002 at 00:50:01, Mike Byrne wrote:
>modified nonsmp crafty 2:41
>
>Crafty v18.15 (1 cpus)
<SNIP>
>White(1): st 600
>search time set to 600.00.
>White(1): setboard 3q2k1/pp3rpp/2n4r/3pP3/P1pP1p1N/B1P3Pb/5P1P/R2QR1K1 b - - 0 >Black(1): go
> clearing hash tables
> time surplus 0.00 time limit 10:00 (10:00)
> nss depth time score variation (1)
<SNIP>
> 14 9:32 ++ 1. ... Rxh4!!
> time=10:00 cpu=99% mat=0 n=373223764 fh=95% nps=621k
> ext-> chk=24050595 cap=694863 pp=1593152 1rep=3845008 mate=51685
> predicted=0 nodes=373223764 evals=53788687
> endgame tablebase-> probes done=0 successful=0
> SMP-> split=0 stop=0 data=0/32 cpu=9:56 elap=10:00
>
>===========================================================================
>modified smp crafty 1:37
>
>Crafty v18.15 (2 cpus)
<SNIP>
>White(1): setboard 3q2k1/pp3rpp/2n4r/3pP3/P1pP1p1N/B1P3Pb/5P1P/R2QR1K1 b - - 0 >Black(1): st 600
>search time set to 600.00.
>Black(1): go
> clearing hash tables
> time surplus 0.00 time limit 10:00 (10:00)
> nss depth time score variation (1)
>starting thread 1
<SNIP>
> 14 4:19 ++ 1. ... Rxh4!!
> time=10:00 cpu=396% mat=0 n=785776691 fh=97% nps=1309k
> ext-> chk=67658533 cap=922567 pp=2386265 1rep=8557559 mate=60987
> predicted=0 nodes=785776691 evals=62997715
> endgame tablebase-> probes done=0 successful=0
> SMP-> split=1375 stop=124 data=13/32 cpu=39:39 elap=10:00
^^^^^??
The SMP version has a node count (and NPS count) more than twice as high as the
single-CPU version, which seems pretty strange.
But the really strange thing is that the CPU time reported is what you would
expect to see on a 4-CPU system. I'd like to know if others have similar
strange numbers for CPU time usage in an SMP environment, or if something funny
is going on here.
Thanks,
Jeremiah
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.