Jump to content

Newly Released RX 480 Appears to Have Power Draw Issues


Recommended Posts

Just days ago, AMD released the long-awaited RX 480 graphics card, containing the new Polaris 10 GPU. One of the features of the new GPU is supposed to be significantly increased efficiency, thanks to the new 14 nm FinFET node it uses and several Radeon technologies, but some reviewers have found the card, at least in the currently available reference design, may have significant power issues.

The card has a TDP of 150 W and a 6-pin power connector, which is meant to provide 75 W, leaving the other 75 W to come from the PCIe slot itself, which is the maximum the slot is supposed to provide. What some reviewers have found, including those at PC Perspective who have done a careful analysis of the card's power draw, is that the RX 480 is pulling above 150 W, at stock (overclocking will naturally draw even more power). On its own, pulling more power than the stated TDP is not necessarily an issue, especially if it draws it from the PSU connector, but the PCPer reviewer found it was also drawing more than 75 W from the PCIe slot. Depending on the tolerances of the motherboard, continuous power draw above the spec could cause damage to the slot. Obviously, this is not a good.

As some comments for the original PCPer article suggested examining the ASUS GTX 960 Strix, which was supposedly caught also drawing more from the PCIe slot in the past, the review pulled out his testing rig again and got to work. He found that this 960 does not in fact exceed that limit, at least not continuously. The reviewer used a second order low pass filter specifically to remove some of the spiking that can occur with phase switches of any DC switching power supply. This makes the data easier to read and the reviewers focus was on continuous power draw, not instantaneous.

It is also worth noting that there has not been any official response from AMD concerning this issue yet. The most I have seen has been an

Share this post


Link to post
Share on other sites

Why is it that manufacturers can't get a card out right without issues with the most basic functions.

If it tells me something, it is that they are all after quick money. If they had taken their time to make it right, they'd make a lot more.

Share this post


Link to post
Share on other sites

AMD simply screwed up this time, power design wise. Im some scenarios card draws almost 200W.

I expected this new 14nm FinFet shit act like or even better than Nvidia 16nm stuff, performance per watt. Not happening.

GloFo process sucks compared to TSMC.

Share this post


Link to post
Share on other sites

AMD has started sending out official statements now (from TechPowerUp):

"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)."

 

Software fix incoming for the apparently incorrect tuning on some cards. More information will be provided on Tuesday, if not the driver update itself.

By the way, I've been seeing reports of people successfully undervolting their RX 480s without having to lower clocks (for some they could actually overclock better because the heat was kept down).

Share this post


Link to post
Share on other sites

They could have fixed this with an 8-Pin instead of using 6Pin. I guess they really wanted people to think it had a low power draw.

 

+1,.. the 8-pin would have been the best way too provide 150W and just use the PCIe lane for headroom/overclocking.

 

Wouldn't want to be the guy who said go with the 6-pin after seeing the power draw during RD, :slap:   

Edited by Braegnok

Share this post


Link to post
Share on other sites

 

They could have fixed this with an 8-Pin instead of using 6Pin. I guess they really wanted people to think it had a low power draw.

 

+1,.. the 8-pin would have been the best way too provide 150V and just use the PCIe lane for headroom/overclocking.

 

Wouldn't want to be the guy who said go with the 6-pin after seeing the power draw during RD, :slap:   

 

Except this level of power draw might not be intentional or known by AMD prior to release. It is possible there is a fluke of situational influences causing the BIOS to request higher volts and more power erroneously or that some cards were flashed with an incorrect BIOS that consistently draws more power than needed.

I'm not saying an 8-pin wouldn't have been a good idea (though I could see the 6-pin being intentionally picked so that the AIB manufacturers can have the additional headroom of other power configurations) but it is also possible this issue did not present itself during the RD and QA phases of development. Until more information is provided on the issue (Tuesday) or the fix goes live and people start searching for the changes, we don't know what the cause is.

Share this post


Link to post
Share on other sites

dont all cards do this on occasion when overclocked??????????

and why does it matter so much.

Overclocking does draw more, yes, but generally that power is drawn from the PCIe cables, not your motherboard. Pulling above rated on the PCIe cables isn't a huge deal...asking your board to do that can quickly cause permanent damage.

Share this post


Link to post
Share on other sites

but this is only effecting older boards that have poor power allocation 

im having no issues what so ever.

The problem is not older mb's, the problem is the card draws over 50 watts more than amd said it would , if they had stated that in all their hype it could fry your mb I doubt they would sell very many.

Share this post


Link to post
Share on other sites

Overclocking new high-end motherboards will cause heat issues just as fast as older mid-range motherboards. The only difference is the high-end boards will handle more heat/volts longer before failing.

 

I was cleaning my workstation last week and noticed it's been getting a bit toasty.   

Share this post


Link to post
Share on other sites

×
×
  • Create New...