Saturday, July 2nd 2016
Official Statement from AMD on the PCI-Express Overcurrent Issue
AMD sent us this statement in response to growing concern among our readers that the Radeon RX 480 graphics card violates PCI-Express power specification, by overdrawing power from its single 6-pin PCIe power connector and the PCI-Express slot. Combined, the total power budged of the card should be 150W, however, it was found to draw well over that power limit.
AMD has had out-of-spec power designs in the past with the Radeon R9 295X2, for example, but that card is targeted at buyers with reasonably good PSUs. The RX 480's target audience could face troubles powering the card. Below is AMD's statement on the matter. The company stated that it's working on a driver update that could cap the power at 150W. It will be interesting to see how that power-limit affects performance.
AMD has had out-of-spec power designs in the past with the Radeon R9 295X2, for example, but that card is targeted at buyers with reasonably good PSUs. The RX 480's target audience could face troubles powering the card. Below is AMD's statement on the matter. The company stated that it's working on a driver update that could cap the power at 150W. It will be interesting to see how that power-limit affects performance.
"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 8 Gbps 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)."
358 Comments on Official Statement from AMD on the PCI-Express Overcurrent Issue
hint: there is something in the user control panel which says "system specs" : perhaps you should check it another time, contains useful info such as ...... the system specs ! ;)
As Cadaveca pointed out, Fermi and the HD2900XT had issues. Though I'd like to know if they were right on the edge, and multiple cards pushed it over, or how much they actually were pulling from the PCI-E slot. I know I melted my 24-pin with a pair of Fermi cards. But nVidia obviously learned several lessons with Fermi and their last 3 generations haven't pulled a lot of power through the PCI-E bus.
P.S. i guess it's forbidden for someone to own an old motherboard because this ruins the "defensive line" for the AMD fanboys.:shadedshu:
EDIT:i have to loose my mind since it's my system. Of course i would care!!
2. Also, check my post #43 . I've been clear about my feelings for AMD from my early posts.
Out of curiosity did you loose a board to an AMD/NV GPU? Have you met anyone who has?
check your post #193. You were the one that told me NOT to buy this GPU because it might destroy my mobo!!!
P.S. No i wouldn't buy anything from AMD after the FuryX period (*but someone else with a similar system could). I simply don't like their policy.
www.overclock.net/t/1604979/a-temporary-fix-for-the-excess-pci-e-slot-power-draw-for-the-reference-rx-480-cards/10#post_25320606
But there are cases where people don't even imagine they are pushing their hardware, and that's why I keep repeating the example of the GTX 950 with no power connector. Putting a two slot beast with 8pin connectors and full controls for the voltages on a PCIe slot, can make someone much more nervous than when putting on a PCIe slot a tiny(compared to the beast) innocent GTX 950 that doesn't even have an extra PCIe connector, doesn't give you probably voltage controls to play and is advertised as a power efficient model. How can something like that be a possible danger for your PCIe slot?
When I saw W1zz's review of the card I realize that, if I haven't understood something wrong, that card could be having - after overclocking - the same power draw problems as the RX 480, because it can turn only to the PCIe bus for power. And that 20% extra performance that W1zz gets after OC, can't come out of thin air. And the card is already at 74W at defaults. W1zz's testing with that card on power draw, wasn't including the overclocking scenario, because his job is to test the card at it's defaults, with the OC page being just the icing on the review's cake. But with all this mess with RX 480, thanks to AMD's stupidity, I believe it's a nice opportunity for professionals to show to all those blind overclockers, that things aren't as simple here as "AMD messed up with RX 480". AMD shoot it's own feet, but probably there are others out there with their gun pointing at their feet, not knowing about it.
1) The first option is to throw extra load to the 6pin PCIe connector, if you trust your PSU.
2) The second is to lower power consumption if you want to stay in specs.
The first option will be used as an argument that their fix didn't had any performance loss on the card's performance. That's why it will be the default. Hardware sites test cards with top equipment, so I believe they will ask tech sites to use the first option if they want to retest the card. If a user uses sub standard PSU that doesn't trusts, that's not AMD's fault anyway. I guess that's going to be the logic behind option one which is going to be used as a fix that doesn't affect the card's performance and also doesn't make AMD look like they acknowledge there is a problem. With this option they are not fixing any problem because there isn't one. They are just calming users who are nervous about the whole story with the PCIe bus power draw.
The second option, is what users who are really concerned about the power draw, will use in the end. To be fair, this is the real fix that will bring the card in specs. Probably we will see lower frequency from the GPU, maybe 1200MHz instead of 1266MHz and 1.1V instead of 1.15V GPU voltage. Some sites will choose and advice users to use this option, instead of the first one, and will retest the cards. AMD hopes that, whatever optimizations and performance increases they manage to archive this week in their drivers, that 3%, will be enough to make the card look like not having lost any performance at all.
Then the custom cards will come and everything will go back to normal.
The effect is not huge, but it is significant enough to alleviate the problem, especially when combined with undervolting and/or underclocking the card.
I think we can let the issue rest now, knowing it's fully manageable. But we should definitely continue to investigate every aspect of performance - including detailed insight into power draw - of all future VGAs under review.
But in the end, kudos to AMD for listening to the grumbling and doing something about it. This shows that they are committed to addressing user concerns, and truly care what people think about their products. Doing this driver change costs them money. Yet I want AMD, and yet, all companies, to adhere to specifications of supporting parts 100%. Spec says 75W, you don't overstep it one bit, and now they give the end user the options!