Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 17 May 2007 10:59:12 -0700
From:      Nate Lawson <nate@root.org>
To:        Peter Jeremy <peterjeremy@optushome.com.au>
Cc:        acpi@freebsd.org
Subject:   Re: PowerTOP for FreeBSD ?
Message-ID:  <464C97F0.5000709@root.org>
In-Reply-To: <20070516193130.GG1162@turion.vk2pj.dyndns.org>
References:  <20070515050404.GK49628@obelix.dsto.defence.gov.au> <1179249135.1149.21.camel@vonnegut> <20070516032813.GB3773@obelix.dsto.defence.gov.au> <464AAC1E.7060002@root.org> <20070516193130.GG1162@turion.vk2pj.dyndns.org>

next in thread | previous in thread | raw e-mail | index | archive | help
Peter Jeremy wrote:
> On 2007-May-16 00:00:46 -0700, Nate Lawson <nate@root.org> wrote:
>> Normally timers tick every 1 us on FreeBSD.
> 
> Unless something has changed massively, tradtiionally, hardclock
> interrupts every 1msec and statclick interrupts at 128Hz (approx
> 8msec).
> 

Oops, yes it's every 1 ms (1/1000 sec).  I didn't bother mentioning stat
clock since it's derived from hardclock and that's the only one that
makes the CPU exit its idle state.

>>  Most interrupts are
>> serviced with no work to be performed, so it's just wasted computation.
> 
> Using the lapic timer is an excellent example of this.  By default it
> runs at 2000Hz to run hardclock() every 1msec and statclock() every
> 7.5 msec - just under 50% of the interrupts are completely wasted.
> 

Absolutely.

-- 
Nate



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?464C97F0.5000709>