Power dissipation formula for quad core cpu

[Rubycon]

Is there a spreadsheet for this OR a formula for calculating REAL power drawn at the socket?

?GHz @ eVCORE = ? Watts

 

[Idontcare]

Originally posted by: Rubycon
Is there a spreadsheet for this OR a formula for calculating REAL power drawn at the socket?

?GHz @ eVCORE = ? Watts

At the socket? How do folks measure the current flowing thru some 600 power pins simultaneously?

I'm sure the equipment exists, at least at the test and package level but what does a consumer do in their garage to extract this data from their system?

But to answer your question, no I don't know of any such resources.

 

[Rubycon]

Actually I have a system that measures power consumption of the system with precise accuracy however I want to track changes based on changes of input to the CPU.

There must be a formula to calculate power with the known values. (speed and vcc)

 

[Idontcare]

Originally posted by: Rubycon
Actually I have a system that measures power consumption of the system with precise accuracy however I want to track changes based on changes of input to the CPU.

Me too, mine's called Kill-a-watt

(I know you are probably referencing something a lot more sophisticated than such $20 toys, but I couldn't resist)

 

[Rubycon]

Originally posted by: Idontcare

Me too, mine's called Kill-a-watt

(I know you are probably referencing something a lot more sophisticated than such $20 toys, but I couldn't resist)

Dranetz Power 1000.

 

[Idontcare]

Originally posted by: Rubycon
Dranetz Power 1000.

If you aren't lucky enough to find an existing equation is there any reason to not make your own equation?

Do you lack the hardware (the CPU) needed to generate the data?

 

[sonnygdude]

My impression is that could be a pretty thorny calculation. Having to do with efficiency, it seems you would get different power draw/power dissipation at different operating temperatures. Plus the issues of how you calibrate power draw relative to duty cycle (CPU utilization). In my experience whenever you look at power dissipations you need to favor empirically measured dissipations of systems under "worst case" and "typical" operations.

I don't want to pretend I'm an expert - but trying to analytically predict power consumption with a reasonable degree of accuracy seems like a pretty daunting problem.

Then again, maybe someone has a tool based on interpolating/extrapolating empirically measured data?