It's almost as if NO ONE in this thread has actually read AMD's OFFICIAL statement... Drivers will address the excessive current draw from PCIe. Considering I'm fiddling with my GTX 980 that has similar fully configurable power control logic, I know it can be also controlled through driver (which is just an extension of BIOS). Meaning what AMD said isn't just load of BS and that it is a valid solution.
As for 8pin being a solution, just a small hint since everyone is screaming EVERYONE SHALL FOLLOW PCIe SPECS!!!!11111oneoneone. 8pin is actually not officially supported by PCI Express certifying body. Meaning, if card has 8pin power connector it's kinda violating PCIe specifications (well, not violating, just not following it). Dual 6pin, no problem. single 8pin or 6pin+8pin ain't officially supported by PCIe specifications.
Also, going single 6pin ensures maximum compatibility with wide range of PSU's. If you go with single 8pin, it's already questionable if target PSU's even have it. Because I think cheaper ones still only have just 6pin...
Actually... 8-pin PCI-E
does exist in the PCI specification, but it was never implemented as specified. Namely the additional two pins (as compared with 6-pin) were supposed to be used for the voltage sense and regulation signal return path. However, now both pins carry the GND/COM wires, which don't really help all that much with making power delivery more stable.
The
real issue here is that the card seems to draw excessive amounts of power from the PCI-E
slot, which is at most fed by two +12V wires on the 24-pin ATX connector, and those are meant for all +12V needs of the motherboard and all connected devices, sans the CPU. Add to that the fact that power traces leading to the PCI-E slot aren't normally very beefy; and the fact that there are only 5 flimsy less-than-paper-thin pins on the card accepting the delivered power and you have a situation where you really want to limit PCI-E slot power delivery as much as possible.
It would actually be fine if the card drew 120W from the 6-pin and up to 50W from the slot, the unnecessary drama would be far less pronounced.
The 6-pin may only be declared as 75W-capable; in reality it can handle the full 150W quoted for the 8-pin in the majority of cases... The only time it can't is if it's on a shitty PSU with wires thinner than 18AWG and cheaply made pins (like on a $15 Diablotek).
Which brings us to...
There is a lot wrong with this statement. No one running a PSU with just a single 6-pin should be running this card. When you have the likes of the bottom of the barrel $30 eVGA 430w that has a 8-pin, if your PSU doesn't have an 8-pin at this point, it's shit. Go buy a new one, it's only $30!
Plus, if the PSU only has a single 6-pin it is probably very close to the edge of actually being able to provide enough power to actually use that 6-pin, so going over spec on power consumption is probably a very bad thing. You're talking about generic shit units that might be rated for 500w, but probably can't do 250w reliably. Do you really think people with those types of power supplies should be using a card that consumes 170w?
Do you really think AMD wanted people with those types of units to use the card? I don't. I think the real reason behind the single 6-pin was marketing. They wanted to hype the card, to say "look at how power efficient it is, it only uses a single 6-pin!" But it backfired on them.
Even the aforementioned Diablotek could handle powering one of these, paired with a latest-gen Skylake CPU, a couple sticks of RAM and some storage. It would all easily fit into a 250W envelope (absolute peak power draw, realistically less than that), which even the worst of the worst PSUs can manage, at least for a while.
That being said, everyone should have the common sense not to skimp on the PSU. No need to go crazy, a nice $30-or-so PSU from a reputable manufacturer should do fine, as
@newtekie1 pointed out.
