DrMrLordX
Lifer
- Apr 27, 2000
- 21,797
- 11,143
- 136
A320 has more right to support the 3000 series because motherboards are still in production,
Who is still producing A320 motherboards? What models?
@maddie
Here is the data I wanted to produce earlier:
Code:
Voltage 1.375v
4.0 GHz: ~145a peak, ~130a average
3.9 GHz: ~144a peak, ~121a average
3.8 GHz: ~142a peak, ~119a average
3.7 GHz: ~141a peak, ~118a average
Clockspeed 3.7 GHz
1.4v: ~144a peak, ~125a average
1.3v: ~126a peak, ~111a average
1.2v: ~116a peak, ~96a average
1.15v: ~105a peak, crashed
I generated this data using HWiNFO64 for monitoring and 1g runs in y-cruncher (Yukina/"ADX" binary) on my R7 1800x. I reset HWiNFO's timer a split second before the beginning of each run to make sure that my averages were not tainted any more than necessary. Sadly, I was unable to replicate the massive 152a number @ 4.0 GHz I saw the other day. I might get better averages from Prime95 since it does not end (y-cruncher has a factoring phase near the end that is not as stressful as the summing phase).
Anyway, as you can see, dropping voltage while maintaining clockspeed had a significant effect on peak and average current draw. Clockspeed adjustments at static voltage, not so much, but it's there. In theory, my 1800x could draw as little as 70a if I could just get vcore down to 1.0v, which is not possible on this chip (1.1v wouldn't even boot). Will Matisse behave the same way? Well, probably not. Unless you choose static clockspeeds, it will not sit at 3.7 or 3.8 GHz and stay there. It will boost all over the place, adjusting voltage as it goes. It's also a wider core. If I use y-cruncher to bench Matisse, I'll be running an AVX2 binary rather than the 128-bit Yukina designed specifically for Summit Ridge/Pinnacle Ridge.