AMD To Issue Software Fix To Address RX 480 Power Consumption Problems

News
By Paul Alcorn
published

In our recent AMD Radeon RX 480 8GB Review we discovered that the GPU exceeds its target 150W TDP and has a less-than-ideal power distribution between the PCIe and 6-pin power connector. This results in excessive over-current through the PCIe slot (we measured up to 90W during the stress test). The results of the tests became a big topic on Reddit, and other hardware sites also issued follow-up testing of the power issue.

We notified AMD of the issue several days before we released the review. This morning, AMD (which appears to be burning the midnight oil on the fix) finally issued a statement in regards to our findings.

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).

AMD is apparently moving quickly on the issue and will provide a fix. The company will follow up with more news on Tuesday.

Paul Alcorn is a Contributing Editor for Tom's Hardware, covering Storage. Follow him on Twitter and Google+.

Follow us @tomshardware, on Facebook and on Google+.

Paul Alcorn
Paul Alcorn
Managing Editor: News and Emerging Tech

Paul Alcorn is the Managing Editor: News and Emerging Tech for Tom's Hardware US. He also writes news and reviews on CPUs, storage, and enterprise hardware.

See more GPUs News
92 Comments Comment from the forums
  • nukemaster
    I hope Tom's will do a re-test. I am interested to see how how much of a performance impact this may have.
    Reply
  • InvalidError
    Yup, really curious to see how much performance the reference RX480 will need to sacrifice to get down safely within PCIe slot power specifications.
    Reply
  • Robert_230
    Yes, Nukemaster this needs to be done, if AMD is gimping the cards AFTER the reviews, then this is super fishy for sure, makes you wonder if AMD knew this was an issue and tried to get stellar reviews for performance before gimping the cards?
    Reply
  • InvalidError
    18213900 said:
    makes you wonder if AMD knew this was an issue and tried to get stellar reviews for performance before gimping the cards?
    AMD CANNOT not have known about the issue. If it genuinely did not know, it would be an admission of gross incompetence or neglect regarding PCIe qualification testing.

    Admitting that it pushed the limits of the PCIe specs too far to get better review numbers would be less damaging to its reputation but the down-tuning it will hurt pricing and raise more questions about why the reference card did not ship with an 8-pin connector in the first place.
    Reply
  • Robert_230
    https://community.amd.com/thread/202410 appearantly some people got their mobos fried from this, not sure if it's true but there are pics.
    Reply
  • Robert_230
    18213931 said:
    18213900 said:
    makes you wonder if AMD knew this was an issue and tried to get stellar reviews for performance before gimping the cards?
    AMD CANNOT not have known about the issue. If it genuinely did not know, it would be an admission of gross incompetence or neglect regarding PCIe qualification testing.

    Admitting that it pushed the limits of the PCIe specs too far to get better review numbers would be less damaging to its reputation but the down-tuning it will hurt pricing and raise more questions about why the reference card did not ship with an 8-pin connector in the first place.

    I hear you, and you are right if they admit it, that's the end of AMD. If they went with an 8pin connector then the 150 tdp rating they gave this card would be out of the question. The card, as it is, is less effecient than a 1070 anyways, they will have to pick either performance or low power usage, can't have both it seems like.
    Reply
  • Omegaclawe
    Interestingly enough, the card uses less power AND performs better when undervolted... goes about 5% faster, runs a few degrees cooler, and uses less power when the maximum voltage is dropped down to around 1.1V... it wouldn't surprise me if the "fix" was to simply make that the default.
    Reply
  • hannibal
    Well, it is more like current rebalancing. A Little bit less From pci and more From 6 PIN. But underwolting seems to be sensible Also.
    Reply
  • USAFRet
    18213931 said:
    18213900 said:
    makes you wonder if AMD knew this was an issue and tried to get stellar reviews for performance before gimping the cards?
    AMD CANNOT not have known about the issue. If it genuinely did not know, it would be an admission of gross incompetence or neglect regarding PCIe qualification testing.

    Admitting that it pushed the limits of the PCIe specs too far to get better review numbers would be less damaging to its reputation but the down-tuning it will hurt pricing and raise more questions about why the reference card did not ship with an 8-pin connector in the first place.

    Either they knew and released it anyway, or they did not know.
    Either way, gross incompetence.
    Reply
  • InvalidError
    18213985 said:
    Interestingly enough, the card uses less power AND performs better when undervolted... goes about 5% faster, runs a few degrees cooler, and uses less power when the maximum voltage is dropped down to around 1.1V... it wouldn't surprise me if the "fix" was to simply make that the default.
    AMD cannot make that "the fix" unless it is confident that at least 95% of GPUs will work properly for the foreseeable future at the reduced voltage.

    With AMD trying to make improved power efficiency a big deal in its RX480 pre-launch marketing campaign, I would be a little shocked if the RX480 launched with voltages cranked up higher than was actually necessary. Who's running engineering at AMD?!?

    I hope AMD does not drop the ball as badly on Zen's launch.
    Reply