Re: An embedded-Linux power-management framework

On Sat, 2002-08-03 at 03:29, Bishop Brock wrote:

On the first point, we're concerned with system-wide power management,
which includes the
memory subsystem.  Our processors are so power-efficient today that we burn
more power

IBM's may be, but this isnt an IBM PPC specific operating system
project. If the servers are for example AMD dual Athlons or 1U Pentium
IV rack boxes then it -really- begins to matter. To put the scale of
this into perspective on x86 - I can run another entire system in the
power differential between idle and full running dual athlon.

In some respects the Athlon is closer to what you claim since there is a
chunk of power saving that occurs only when both cpus are stopped.

I'm not arguing against your claim. I want to be sure we can support
both.behaviours.

power management policy lead to slack, i.e., idle time.  Our proposal
allows the system to
capture the slack by aggressively idling the processor for short periods,
e.g., between video
frames, without any effect on the application, and with significant energy
reduction.

Some processes show short interactive behaviours. You can quantify and
measure that, indeed the scheduler already has an extremely good idea
what is and is not an interactive process that will execute short bursts
then sleep a lot. Whether that is enough I don't know.

Alan



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf