[mythtv-users] CPU power (was Possible small HD frontend)

Brian Foddy bfoddy at visi.com
Sat Mar 22 17:08:32 UTC 2008


On Saturday 22 March 2008, Brian Wood wrote:
> On Mar 22, 2008, at 9:56 AM, John Drescher wrote:
> > What I do not see in this thread (the 7 responses I have today) is
> > that usually by slowing the clock you can reduce the voltage of the
> > processor by a considerable amount and this lowering of voltage will
> > help reduce power by a considerable amount. Also I know for a fact
> > that AMD systems (Opterons) will use considerably less power measured
> > by a Kila-Whatt when at 1GHz at idle versus 2+ GHz idle.
>
> That contradicts the intuitively unobvious info from another poster.
> As I said I find it hard to believe that slowing the clock does not
> reduce power consumption.
>
> But the point was also made that at a slower clock a given job will
> take longer, and thus the *total* amount of energy may be the same or
> even higher.
>
> In the physical world doing the same work faster usually takes more
> energy than to do it slower, not theoretically perhaps, but real-world
> factors like friction and thermal losses make it so.
>
> beww
>

On my X2, it made a bit difference (reference my just recent post).  I've
also tried it on a dual 3GHz Xeon P4, and tho the cpu slowed and the machine
was stable, the kill-a-watt showed no difference at all, but the machine
was definitely slower.  I hope this weekend to try it on a dual Opteron 
machine I have to see it it makes a difference.

Brian



More information about the mythtv-users mailing list