- Feb 22, 2001
- 3,044
- 544
- 136
I used a Kill-o-Watt P3 to measure wattage used at the wall.
I set my multiplier to 38/35/Auto -- and never touched the CPU voltage in the BIOS ( Auto )
I'm showing the numbers in multiplier order ( 3.8ghz/3.5ghz/Auto ).
The numbers are me eyeballing the Kill-o-Watt - Taking lows, what I think was average, and highs into consideration.
Desktop Idle
68/61/59 - average after 2 minutes waiting for Windows to settle
Dirt Rally Demo Mode
212/206/203 - average
246/238/217 - peak
Cinebench 2015 CPU
215/208/147 - average
217/213/163 - peak
The Cinebench surprised me because it was the biggest drop in power - probably 50% of time was 138w.
I did this purely as an exercise to see how inefficient would I be running the 38x multiplier "full time", and, in my mind, the penalty is very minor.
I set my multiplier to 38/35/Auto -- and never touched the CPU voltage in the BIOS ( Auto )
I'm showing the numbers in multiplier order ( 3.8ghz/3.5ghz/Auto ).
The numbers are me eyeballing the Kill-o-Watt - Taking lows, what I think was average, and highs into consideration.
Desktop Idle
68/61/59 - average after 2 minutes waiting for Windows to settle
Dirt Rally Demo Mode
212/206/203 - average
246/238/217 - peak
Cinebench 2015 CPU
215/208/147 - average
217/213/163 - peak
The Cinebench surprised me because it was the biggest drop in power - probably 50% of time was 138w.
I did this purely as an exercise to see how inefficient would I be running the 38x multiplier "full time", and, in my mind, the penalty is very minor.