power management

[Dan Williams]

On Fri, 2007-03-30 at 01:40 -0400, Albert Cahalan wrote:
> On 3/30/07, Richard A. Smith <richard at laptop.org> wrote:
> > Albert Cahalan wrote:
> >
> > > Well, how many steps do you have? Divide the battery life by that
> > > to get the maximum rate. If that's not more than 1 per minute,
> > > then you might as well report every change.
> >
> > Nobody has presented a use case for a continuous system so what I'm
> 
> Here are two.
> 
> Use case: a battery display

Counterpoint here; cellphones only display 4 - 6 bars for battery.
While most cellphones longer on standby than the XO is supposed to last,
most people recharge their phones every day or so which is what's
expected out of the XO as well.  I'm not sure there's a really good case
for a battery _display_ that has fine-grained steps.  For other stuff,
perhaps.

Dan

> Use case: power use measurement for software developers
> 
> I'm sure others will discover more uses.
> 
> As long as the wakeups wouldn't be frequent enough to
> have a real impact on battery life, providing the info is good.
> Ignoring unneeded info is easier than hacking up new
> EC firmware to get the info.
> 
> I've been thinking that one could hack gcc to optimize
> for power usage or make valgrind estimate power usage.
> Good power measurement on unmodified hardware is
> an enabling feature for such projects.
> _______________________________________________
> Devel mailing list
> Devel at laptop.org
> http://mailman.laptop.org/mailman/listinfo/devel