AMD has released an official statement regarding the user reports of Radeon RX 7900 XTX graphics cards. In a nutshell, the chipmaker urges affected consumers to contact support.
Although the Radeon RX 7900 XTX is one of the best graphics cards, MBA (Made By AMD) models appear to suffer from overheating. The reference design exhibited hotspots peaking up to 110 degrees Celsius; user feedback claims that the RDNA 3 graphics card shows this behavior even with the cooling fans at full throttle.
"We are aware that a limited number of users are experiencing unexpected thermal throttling on AMD Radeon RX 7900 XTX graphics cards (reference models made by AMD). Users experiencing unexpected thermal throttling of an AMD Radeon RX 7900 XTX should contact AMD Support," stated AMD in a statement to Tom's Hardware.
According to AMD, 110 degrees Celsius is the normal junction temperature for the Radeon RX 7900 XTX. However, the current suspicion is uneven contact pressure from the cold plate could be causing the overheating problem on reference Radeon RX 7900 XTX graphics cards. Some user accounts claim that repasting or swapping the TIM for a thermal pad is a viable solution. Nonetheless, AMD's team is currently looking into the issue, whereas PowerColor started to collect user feedback to help in the investigation.
AMD hasn't provided a firm date on when we can expect guidance on the problem. In the meantime, the chipmaker has advised affected users to contact the company's support team. Many consumers have reportedly returned their Radeon RX 7900 XTX graphics cards to AMD or the retailer.
Stay on the Cutting Edge
Join the experts who read Tom's Hardware for the inside track on enthusiast PC tech news — and have for over 25 years. We'll send breaking news and in-depth reviews of CPUs, GPUs, AI, maker hardware and more straight to your inbox.
Zhiye Liu is a Freelance News Writer at Tom’s Hardware US. Although he loves everything that’s hardware, he has a soft spot for CPUs, GPUs, and RAM.
Melting connectors, 100 degree hotspots, deceptive naming schemes...Reply
These manufacturers can't wait to take ppl's money by the thousand but don't even have the decency to make sure their products are up to minimum standards.
While it is well within my skill set I shouldn't have to take a brand new graphics card apart and re-paste it .... Likely they are using the same manufacturing technique for the chiplet design as they did for the older monolithic design and there needs to be a modification to the assembly process to make up for the differences in the dies when assembling the heatsink.Reply
This is why I don't recommend buying the new AMD GPUs because it's going to take another generation or two to work out the bugs with the Chiplet design/Infinity Cache just like I don't recommend buying a 1st gen Intel GPU .... Let Captain Kirk go boldly where no man has gone before .....
Those who live on the Bleeding Edge often die a Death by a Thousand Cuts
Speculation typically leads to inaccurate conclusions. Unfortunately until the issue is properly investigated and the cause determined the theories can be a disservice.Reply
Would be interesting to know which company manufactures the reference cards.Reply
Today standards if don't melt out the box can't be good. My meltdown card lolReply
Hm... This is a positive look, I'd say?Reply
It does appear they're trying to do good to early adopters, so good on AMD for that. I hope they do find the problem and post a clear guidance on what to do about it soon.
Arctic Freezer just went through a replacement situation with some of their AIOs and mine was affected. I'm skilled enough to be able to do the required maintenance on it, so it was painless and helped me avoid returns and such. I hope it doesn't come down to people having to re-paste their own cards, which would be less than ideal, but at least AMD offering quick returns and/or a quick "we'll fix it for you at no cost" service once they find the issue.
Yet another case of "support" = "no support at all"!Reply
Following the link provided, you end up at a page which offers articles (NOT support), with a cokkie-gated button at the bottom saying "Need more help? Contact Support". Clicking on that takes you to a page with the following message:
This page is temporarily unavailable as our systems are currently undergoing maintenance.
This maintenance period is expected to end on January 5th, 2023, when this page will be available once again to take your inquiries."
Having raised 2 other issues with both AMD and Sapphire regarding my 7900XT (vertical compression of entire screen on 2nd monitor and completely borked Realtek sound output, both of which remain unresolved for 2 weeks), I won't be holding my breath.
Why are you writing your articles like all the reference RTX 7900 XTX have this issue? It's not all the cards, it's just some of them. I read many comments saying that their temperature is fine and I got mine yesterday and it's also good: GPU temp tops at 70 and the hot spot rarely goes above 82-83 (so same thing than my RTX 2080, actually). It's a Sapphire but the reference model and I was so worried in the last few days because sites like yours talk like every reference model has this problem but it's not true. It's very sad for people who got a bad one and someone really messed up in a factory somewhere but most cards are fine.Reply
By the way, the card is incredibly powerful as compared to my 2080 and runs almost everything at 144 fps ultra settings and 1440p (still struggles with RTX in some games like Cyberpunk like all the reviews said). I just hope AMD will address some issues in their next driver like the crashes with DX12 in The Witcher 3 but other than that it's great.
It's refreshing to see a giant manufacturer actually responding to negative news and providing some sort of help/support this soon rather than waiting for some lawsuits to be filed and even then only responding in generic vague statements about they will fight some lawsuit of some false claims!Reply
Rx6000 series did the same thing, Hotspot 110c core 70c.Reply
It boosted up to 110c or at its MHz cap.
It was the design, not a flaw - I don't understand why people freak out about it.
Just buy a waterblock if you are concerned.