On Mon, Feb 02, 2009 at 11:17:29PM -0500, Andy Gospodarek wrote:
And your implied claim that it's a majority is nothing more than
anecdotal as well. In fact, I would even say the verification of your
wattage-savings is a bit of a guess based on the description I've seen
of your testing.
Oh, I'm certainly not going to claim that a majority of our users care
about power consumption. I've got absolutely no idea what the relative
weightings are, just anecdotes in both direction. That said, if you've
got any better ways to measure the power difference, I'd be happy to
provide more figures. The 2W one pretty closely matches the expected
figure (40ms of smoothing plus a frame for the change to take effect
times two transitions per second, gives a figure of a little more than
10% of the time in an upclocked state, and 18W or so between the
upclocked and downclocked figures on the hardware in question) so I'm
fairly happy with the result.
--
Matthew Garrett | mjg59(a)srcf.ucam.org