[TriLUG] Home Power Consumption, was: Re: [Novalug] Comparing Clouds; A trivial test. (fwd)

Maxwell Spangler maxlists at maxwellspangler.com
Sat Nov 2 14:06:16 EDT 2013


On Fri, 2013-11-01 at 18:09 -0400, Aaron Joyner wrote:

> machine.  My power at home runs about 14 cents per kw/hr.  I don't
> know much about the power efficiency of the Phenom II or what power
> supply you're using, but assuming it's around 350w to 400w to run that
> machine, your electricity cost is roughly equivalent, and GCE has no
> up-front capital costs.


I like Aaron's thinking of combining capital equipment costs plus power
costs when comparing running your own systems to the cloud. Add real
estate, insurance, and cooling as well for businesses, etc.

On a related note regarding the details:

Our local libraries in Larimer County, Colorado allow us to check out
power meters to investigate home power use.  I've done this twice and
encourage others to do the same.  The results are interesting.

I'm finding that along with calorie counters, FitBit devices and other
ways of [pseudo]scientifically measuring your world, measuring your
actual power consumption is an interesting and rewarding activity.  You
know turning that light off saves 60, 120, 300w, but how about the
stereo? the spare file server? your typical laptop, etc?

As a result of my testing I found that my four small headless servers
acting as two file servers, one internet gateway and a central
dns/dhcp/pxe server were about equal to a small MT PC in power use.  So
I could load up one server, virtualize everything and consume about the
same amount of power.  I haven't yet, but I'm pretty sure its coming
because now I know power consumption is equal.

NFS File servers; 2each x 55 watt average
Gateway and central server: 2 each at 35 watt average.

Some others:

60w lightbulb: 58.9 watts (reference test)
Kenwood stereo: 44.4 Watts, on.

For computer tests I measured a variety of states such as "ac only" for
plugged in, "charging while idle" for laptops" and "cpu dd test" meaning
run 2-3 copies this for each core in the system to get the CPU running
at maximum speed and consuming [simulated] full power.  (A true test
might exercise all parts of the CPU, exercise integrated video, etc, but
this is just for exploration.)

# dd if=/dev/zero of=/dev/null

HP 2007 era laptop with 2 core Core2 CPU and Nvidia power hungry
graphics card.
AC only: 6.6w
Charging while idle cpu: 44-48w
CPU dd test: 78w
Same test using a dock: 120 for the combination.

My HPE-204F Single socket AMD Phenom II CPU, couple hard drives, low end
AMD Radio PCIe card.  Fairly basic desktop with perhaps a 300W power
supply:

Idle 95-106w
DD test: 180w
Average use: 133-141w

'Antec' a whitebox Asus MB with Phenom I CPU, couple drives, etc:

103w idle
120-135w during a single file copy

HP 2-socket Intel Core2 2-core Xeon XW6400 workstation, 2007 era:

Plugged in: 2.2w
Idle: 215w
15x DD: 289w

These are just some random reports for your curiosity and consideration.

-- 
Maxwell Spangler
========================================================================
Linux System Administration / Virtualization / Development / Computing
Services
Photography / Graphics Design / Writing
Fort Collins, Colorado
http://www.maxwellspangler.com


More information about the TriLUG mailing list