I really had hoped they would have lowered the voltage about 10%, to get that 21%, as that would really be awesome.
If they had lowered the voltage, the card would already be at its best.
But with standard voltage there ist still the possibility of tweaking it even further...
By the way, i am really pushing this card today. The GPU is running dirt2 replays for hours now to heat up my Reserator. After that I ran Furmark for hours without overheating anything on the Mainboard (that's where the power comes from). I really think the claims of Powercolor are right.
The vr heatsinks and the mem heatsinks are at 65Â°C (measured with a decent ir thermometer).
Furmark measures 57Â°C for the GPU watercooled with ZM-GWB3 and the water Temps are at 40Â°C which seems high but is perfectly alright for the Reserator (ambient is 24Â°C).
So everything is cool.
Edit: I tried to measure the cards power consumption as good as i can(with a Conrad Energy Check 3000 at the wall socket).
From the Test with my minix board i know that idle power consumption is around 15W wich is pretty much the same as a good standard card.
In my rig with the MSI 770T-C45 and a
Brisbane dualcore @5x200 @0,8VCore fixed with crystalcpuid
the idle consumption is ca. 83W.
When Furmark kicks in (wich creates ca. 95-97% use), the consumption rises to 134W. That means a rise of 51W. If we add 15W standby consumption, we get 66W at load without taking into account that the cpu also uses a little bit more power at load (1 core is used 60-85%)
I think this definitely shows that Powercolor did their homework.
This card is a nice piece of efficient engineering.
Efficient engineering of power circuits has proven to save up to 20W at load with recent motherboards so why shouldn't that be possible with graphics boards where some vendors don't even bother to put the right circuitry on their products to get the reference consuption values.