AMD Radeon RX 480 is currently the talk of the town for all the good reasons, until now. Reviewers have noticed the power consumption of the new card is very awkward as it tends to draw more power from the PCI Express slot than its 6 pin power connector. But replicating this issue isn’t easy as being able to check the power consumed only by the slot is a tedious and expensive job. But some reputed reviewers have gone the extra mile and conducted these tests for us.
Golem.de, A German site, said this:
Our measurements show that the reference design easily overloads PCIE slot. To be specific it’s 5.5 amps at 12 volts, 66 watts total. The Radeon RX 480 draws 78-83 watts out of the slot, which is more than power coming from 6-pin connector. Excessive load from PCIe slot is not limited, because it’s within safe reserves. In the worst case scenario the motherboard would switch off. However with our Asus Z170-Deluxe in the test system, the Radeon RX 480 ran smoothly, and also with our MSI Z97 Gaming 3 there were no abnormalities. Nevertheless, the overload can theoretically lead to reboots or even damage, especially with overclocking.
While the report states that the problem isn’t as alarming as it sounds, and is a minor problem for people with enthusiast grade equipment, the average people with lower end motherboards may face problems while using this card, PC shutdown being the first of the symptoms, continuous abuse of the slot could potentially cause damage, as the cheaper build quality of the board won’t be able to take overloading. And with its overclocking potential said to be at an all time high, amateur overclockers could break their cheap motherboards.
This is what Heise Online had to report after their testing:
In games, the Radeon RX 480 exceeds 150 Watt thermal design power and PCIe specification. On average, we have 156 watts measured at games, in Furmark the card even pulls at reduced clock 169 watts. Extremely problematic is the fact that the Radeon RX 480 draws up to 88 watts from motherboard slot, although specifications only allow 75 watts. This may in some motherboards cause instability. Since AMD card already operates above the standard, far above the limit, it is not advised to perform overclocking experiments with the reference design, which has only one additional 6-pin connector. Here one can only hope that AMD’s partners offer boards with standards-compliant layout and sufficient power supply.
There are 3 possibilities now, AMD would choose to remove the PCI Express logos from these cards, or they’ll either alter the boards to have an 8-pin socket and draw more power from there, or they will neuter the card to ensure it doesn’t draw more power than the PCI Express specification allows.
AMD is well aware of the problem and they addressed this issue with the help of a Reddit AMA (Ask Me Anything) where Raja Koduri said:
“We have extensive testing internally on our PCIE compliance and RX480 passed our testing. However we have received feedback from some of the reviewers on high current observed on PCIE in some cases. We are looking into these scenarios as we speak and reproduce these scenarios internally. Our engineering team is fully engaged.”
Robert Hallock further explained on Reddit:
1) The RX 480 has passed PCIe compliance testing with PCI-SIG. This is not just our internal testing. I think that should be made very clear. Obviously there are a few GPUs exhibiting anomalous behavior, and we’ve been in touch with these reviewers for a few days to better understand their test configurations to see how this could be possible.
2) Update #2 made by the OP is confused. There is a difference between ASIC power, which is what ONLY THE GPU CONSUMES (110W), and total graphics power (TGP), which is what the entire graphics card uses (150W). There has been no change in the spec, so I would ask that incorrect information stop being disseminated as “fact.”
We will have more on this topic soon as we investigate, but it’s worth reminding people that only a very small number of hundreds of RX 480 reviews worldwide encountered this issue. Clearly that makes it aberrant, rather than the rule, and we’re working to get that number down to zero.
Robert Hallock, while I can’t dismiss your reply as false by any means,it’s weird you say that a small number of cards have this problem and yet many reviewers who got the experiment done found the exact same issues.
But it is worth noting that AMD released a driver update that reduced the idle power consumption from 16W to 10-11W, so if software can solve this problem, AMD may have dodged a bullet this time around. I bet the green team will be more cautious on their GTX 1060 launch.
For more news on the GPU wars, stay tuned to Asidcast.