| 1 |
On Wed, Aug 23, 2017 at 05:29:13PM -0500, R0b0t1 wrote: |
| 2 |
> As an example, I am interested in characterizing the power consumption |
| 3 |
> of rendering a PDF document. I would hopefully only need to run the |
| 4 |
> renderer once. |
| 5 |
|
| 6 |
I remember in college some interesting work in security-related stuff |
| 7 |
that was about characterizing power draw in order to crack passwords. |
| 8 |
Potentially that could be useful, as you could record the waveforms |
| 9 |
going into your power supply while rendering a PDF, but it sure sounds |
| 10 |
complicated. |
| 11 |
|
| 12 |
> Apparently you can talk to these, but my searches can only find code |
| 13 |
> which seems highly experimental. The other replies seem to be for |
| 14 |
> embedded Linux systems running on FPGAs and perhaps Cortex-A parts. |
| 15 |
|
| 16 |
Even if you could talk to these, taking periodic measurements of |
| 17 |
current and voltage, then integrating over time would probably give |
| 18 |
results with large errors. |
| 19 |
|
| 20 |
> If I were using a microcontroller I could get uA or nA draw per MHz |
| 21 |
> and I know my operating voltage and operating time. However, desktop |
| 22 |
> processors are much more complex, and I am not sure if they have been |
| 23 |
> entirely characterized. The most advanced tool I can find is PowerTOP |
| 24 |
> and it does not seem very accurate. |
| 25 |
|
| 26 |
Due to modern processors having things like pipelines and all of the |
| 27 |
caching and memory pre-fetching going on behind the scenes, it's probably |
| 28 |
not easy to measure this with any degree of accuracy. |
| 29 |
|
| 30 |
Have fun, |
| 31 |
|
| 32 |
Alec |
Archives