Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 16 Mar 2004 09:36:17 +0300
From:      "Artem Koutchine" <matrix@itlegion.ru>
To:        "Erich Dollansky" <oceanare@pacific.net.sg>
Cc:        freebsd-hardware@freebsd.org
Subject:   Re: Multiprocessor system VS one processor system
Message-ID:  <026901c40b22$2c557c50$0c00a8c0@artem>
References:  <CD1C1AC62C4DA944914E36E7B3DD904313DB17@sestosrv004p.ad.octapharma.se> <002f01c4081c$bd720810$0c00a8c0@artem> <40568DC4.7020209@pacific.net.sg>

next in thread | previous in thread | raw e-mail | index | archive | help
> > of processes with short run periods you will get overal
> > perfomance somewhere near 5000Mhz. However, when i 
> > subsituted two 500Mhz CPUs with 1500Mhz one (even with
> > less 2nd level cache) on a heavy loaded web server i notice
> > that sites started to load faster. So, it seems as 
> > one 3X Mhz  CPU is faster that two X MHz CPUs, at least
> > for web server with sql base and many perl scripts.
> > 
> This is true as long there is no load until the CPU cache comes 
> into the game. It also depends on how the CPUs are connected to 
> the main memory.


Hm..
Let me make up  a case.
 Two boxes:
1) Dual CPU  X Mhz with Y KB of cache
2) Sinnge CPU with 2X Mhz and 2Y KB of cache

Which one is faster under FreeBSD? I think the
second one, because SMP overhead is gone.

Artem



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?026901c40b22$2c557c50$0c00a8c0>