AMD's Polaris-equipped Radeon RX 480 is here and while it didn't explode out of the gates, it's shown itself to be a solid performer. It also appears a little too power-thirsty, to the point where concerns have been raised about the possibility of it damaging hardware.
Ryan Shrout over at PC Perspective decided to tackle these issues head-on by testing the power draw from the RX 480 under various scenarios, including stock and overclocked.
On the surface, the results are not in AMD's favour, with the card at times drawing power beyond specifications:
Keep in mind this is NOT with any kind of overclocking applied — everything is running at stock here. The blue line hits 85 watts and the white line (motherboard power) hits nearly 80 watts. PCI Express specifications state that the +12V power output through a motherboard connection shouldn’t exceed 66 watts ... Clearly, the RX 480 is beyond the edge of these limits but not to a degree where we would be concerned.
As expected, the overclocking tests were a little more disturbing, but that's entering "at-your-own-risk" territory anyway:
When we zoom in we find that the motherboard is actually providing more than 95 watts of power over the +12V line and maintains the 5 watts from the +3.3V line, proving that we are indeed getting more than 100 watts through a PCIe connection that is only rated at 75 watts. The 6-pin PCIe power cable is almost crossing that 100 watt barrier too.
Despite the numbers, Shrout's conclusion is cautious and level-headed — he states he's yet to run into stability problems with the card and hasn't "heard of any other reviewers" running into issues either.
Those hoping for a cliffhanger will be disappointed: AMD responded to Shrout about these concerns:
As you know, we continuously tune our GPUs in order to maximize their performance within their given power envelopes and the speed of the memory interface, which in this case is an unprecedented 8Gbps for GDDR5. Recently, we identified select scenarios where the tuning of some RX 480 boards was not optimal. Fortunately, we can adjust the GPU's tuning via software in order to resolve this issue. We are already testing a driver that implements a fix, and we will provide an update to the community on our progress on Tuesday (July 5, 2016).
So, where does that leave us? I don't think we'll be seeing mass recalls anytime soon, but it will be interesting if this "not optimal" tuning has long-term ramifications, especially for those tardy about driver updates.