Measure CPU power

Archived from groups: comp.sys.ibm.pc.hardware.chips (More info?)

"hemanth" <hemanthkc@gmail.com> writes:
>How to measure the power consumed by an application i.e. the cpu power
>used over a period of time?
>In case any 1 knows the answer pls let me know.

If you had alternating current being supplied then you could just
use a clamp-on amp meter, the voltages would be stable so you could
just integrate the current measured over time and you would have
your power.

But the power is not alternating, it is direct current, so I don't
think that will work, I'm not aware of a clamp on DC amp meter.
But there are a couple of other possibilities.

You could measure the voltage drop across a resistance with that
current flowing through it. Suppose you knew the resistance of the
length of wire from the power supply to the CPU was .001 ohms. If
you had a good meter and could measure the voltage, say it was 0.1
volts. Then Ohms law would let you determine that the current was
..1/.001=100 amps, just as an example. But resistance of copper
changes with temperature, for accurate measurements you need to
decide whether that is enough to matter. And measuring tiny voltages
good accuracy inside the radio frequency mess that exists inside
pc cases can be a problem.

A current flowing through a wire produces a magnetic field. You
could wrap a few turns of the wire supplying the power around a
magnetic field sensor, like a linear hall effect device, and get a
voltage out of it that was proportional to the current flowing
through the wire.

You could water cool your processor and measure the flow rate of
the water coolant and the temperature rise of the liquid after it
has passed over the processor. Take that mass of water, the heat
capacity and the temperature rise and you can deduce the power
removed from the CPY by the water coolant.