search

sosy-lab / cpu-energy-meter

A tool for measuring energy consumption of Intel CPUs
BSD 3-Clause "New" or "Revised" License
321 stars 29 forks source link

add support to watts as output units #24

Closed rkmax closed 3 years ago

rkmax commented 4 years ago

I use https://www.rapidtables.com/convert/electric/Joule_to_Watt.html as reference to convert the joules to watts

PhilippWendler commented 4 years ago

Thanks for this contribution!

We cannot simply label this as changing the unit, because it actually shows something fundamentally different: power instead of energy. And CPU Energy Meter has "energy" in its name, so we should make sure that we do not confuse users about this.

Furthermore, the output you propose is the average power over the whole measurement time. So we should also make sure that users to not confuse this with peak or current power, in which they also might be interested in, but which CPU Energy Meter cannot provide.

What is your use case for the way, btw.? CPU Energy Meter is currently mostly considered as a low-level measurement tool that provides raw data, and when I postprocess the data (using other tools) I would usually compute average power there if I am interested in it.

Having said that, if it its communicated clearly to the user and there is a use case for it, I certainly would not be opposed to adding this.

rkmax commented 4 years ago

Hi, thank you for taking the time to explain to me. my use case it's a way to see the consumption of the processor, in the same way, I can see the consumption of my graphic card and as you mentioning probably it's better to build something on top of this to show what I want.

and abusing of you time I have few questions for you

thanks in advance for you time

PhilippWendler commented 4 years ago

Sure, no problem. If it is labeled "average power", for example, this should be pretty clear. We could show both, and if the use case is primarily for looking at it, maybe only in the human-readable output? There it could even be in a separate column after each value. The calculation seems correct in any way.

PhilippWendler commented 3 years ago

Are you still interested in continuing with this?

PhilippWendler commented 3 years ago

Please reopen if there is still interest.