@thebluebumblebee - maybe with
@Dinnercore s picture there, we can say that it might be cheaper to buy the two Ryzen 1700/2700 etc, but to run it definitely the 1700/2700 will save you a few quid on electric!!
Wow that is a lot of power but it is overclocked am I right
@Dinnercore ?? Have you tried a stock speed crunching test for it??
It is overclocked, but not much. I pulled it down already from 3.9GHz @ 1.325V down to 3.725 GHz @ ~1.27V in that picture. Currently I have it one step less with 3.675GHz @ 1.25V, which only dropped about 15W. I will try to get further down, still did not encounter instability yet, it may be ok to go for 1.225V.
EDIT: I never intended this system to be power efficient, it is after all an 'enthusiast' build with dual loop watercooling. BUT I learned something already from fooling around.
One thing I see is, what on earth does the power draw in HWMonitor read? It never makes sense compared to what I measure.
Another big point to mention is, that rated TDP != power draw. It is more like 2x TDP = ~power consumption. The CPU itself may take 180W, or more like loose 180W as heat, but before that those 180W are drawn from VRM which has quite some loss to it, and then this number (lets guess it is 80% efficient) of 216W is demanded from the PSU which has another loss of 5-10% depending on the rating. In the end the 180W CPU will draw more like 235W from the wall.
In my case it must be far worse efficiency wise, at idle I measure 240W and CPU 100% load it´s 640W now. That is 400W just by putting load on the CPU, and at this point I´m kind of starting to question my tools of measurement. Sure there are 16 cores that want to eat and feed an extra thread. But HWMonitor reads from the mainboard in idle 60W on the CPU and under load it´s currently 188W. That is 128W increase under load which translates to 400W (!) more power draw from the wall? Excuse me I got a titanium PSU. At these loads and 240V input it should reach 94-95% efficiency. Feeding 400W and actually delivering just 128W = 32% efficiency.
Either I miss some key knowledge about power figures around PCs or there is a big error somewhere in all of this. Just for fun I ran Heaven benchmark with my Vega (not even on LC settings, just stock air voltage and clocks) and got 910W power draw total. With CPU sitting at 20-30% usage.
Dropped down to 1.225 Vcore btw:
load
idle
The third key note I found out was, I really don´t need any core speed on this machine.
The bottom two results are stock settings, with boosting up to 4.X GHz but heavy powerlimit. The 3.68GHz is my current result on 1.225 Vcore and 10. = 1.25 Vcore and 9. = 1.28 Vcore.
I will now continue to drop voltage until I crash, then I will go down with clocks and see how it behaves around 3 GHz.
Very many thanks to 
Sadly it's a bit far to send to you 
) and my personal experience. My 1700 running stock clocks is pulling 130 watts with a GTX 460. Guru3D's results for the 2700 appears to be 150 watts, which I see as totally believable. They show a ~110 watt jump for the 2950X and an additional ~100 watt jump for the 2990WX at stock clocks. I also don't think it's fair to compare the non X 2700 with the TR's. Would be interesting to see AMD come out with some energy efficient TR's.
