AMD Radeon RX 7900 XTX May Feature Faulty Coolers, Causing Overheating

[AusWolf]

New news.

Thousands of AMD Radeon RX 7900 XTX reference cards could be affected by 110°C GPU hotspot issue - VideoCardz.com

AMD expected to issue a statement. System integrators already returned hundreds of Radeon RX 7900 XTX MBA cards The German investigative journalism continues. Igor Wallossek from Igor’sLAB has a lengthy post explaining the current situation with AMD Radeon RX 7900 XTX graphics cards and their...

videocardz.com

The article is new, the content isn't.

TL,DR from the end of the article:

"AMD has issued a short statement, finally admitting that there is a problem with reference cooling design:
We are working to determine the cause of the unexpected performance limitation of AMD Radeon RX 7900 XTX graphics cards. Based on our observations to date we believe the issue is related to the thermal solution used in AMD’s reference design and is occurring in a limited number of cards sold. We are working to resolve this issue for the affected cards. Customers experiencing this unexpected limitation should contact AMD support."

 

[maxfly]

The article is new, the content isn't.

TL,DR from the end of the article:

"AMD has issued a short statement, finally admitting that there is a problem with reference cooling design:
We are working to determine the cause of the unexpected performance limitation of AMD Radeon RX 7900 XTX graphics cards. Based on our observations to date we believe the issue is related to the thermal solution used in AMD’s reference design and is occurring in a limited number of cards sold. We are working to resolve this issue for the affected cards. Customers experiencing this unexpected limitation should contact AMD support."

Yep the interesting bit is that AMD planned a statement for yesterday but postponed it. So maybe it will come out todayish.

 

[nguyen]

quite a new low for AMD

 

[Why_Me]

quite a new low for AMD

View attachment 277456

Love it .... AMD vs the Germans (other than Mindfactory.de which is an AMD schill).

 

[maxfly]

quite a new low for AMD

View attachment 277456

Mistakes happen, it's how it's dealt with that will determine whether or not this will leave a lasting impression.

 

[AusWolf]

quite a new low for AMD

View attachment 277456

Whether it's a new low or not, we'll see from the statement. This isn't the first time a major company F-ed up. How a F-up is handled is a good indication of a company's attitude towards its customers.

 

[Vayra86]

This is very often in tech articles. If you "get personal" they will explain to you it's you that doesn't understand - it's "clear" that they meant "market share of the products shipped in a quarter".

And it's impossible to get overall market share from the data of shipped cards - what's the "best before" date of graphics cards? Some people retire their cards in one generation, and don't even sell them on, some people buy every two generations, and sell their old card, some people use very old cards - either bought, or are upgrading only once every few generations...

The only more or less useful market share numbers I know of is Steam Hardware Survey, and people have problems with that chart since the introduction! I wasn't polled, so it's inaccurate! It's clearly biased and it's numbers are fake! And so on... So people more or less stopped using it as a data source to create their charts on forums (due to attacks), and Valve doesn't provide very helpful visualizations - they bundle together everything - integrated graphics, laptop cards and discrete desktop cards...

The reality is, to understand these statistics, what you need is combine a crapload of them, apply context of the markets and of real customer journeys and then you've probably got a half-decent idea. The best I gather from these charts of JPR is the ballpark progress. I think what's clear, is that AMD is losing share YoY and Nvidia gains it. What's clear, is that Intel captured a tiny bit of share and pushed out some units. Another thing that's clear, is that especially Nvidia's numbers are heavily skewed because they pumped out lots of chips and then crypto crashed. These chips are in the market and in these numbers, but don't represent actual gaming share.

That's about as far is this goes. Another point is that we don't see consoles here, while these are 'discrete gaming devices' nonetheless, and in many use cases not even all that different from gaming PCs, including peripherals etc. Another point is that Intel commands the IGP market, on which simple games are also played, and we don't know how much relatively to those on which people don't.

Discrete in its pure PC definition is really just such a sliver of the market, it never determined anything. It can lead the way, be the signpost for gaming in the near future, because PC is the most versatile platform for innovation. What I gather from all these stats is mostly trends. Not real numbers.

 

[nguyen]

Mistakes happen, it's how it's dealt with that will determine whether or not this will leave a lasting impression.

Knowingly sell defective products should probably leave a very long lasting impression, who is to say AMD won't do it again in the future?

Kinda funny that there are people who verbally accused Der8uer who uncovered the truth AMD has been hiding for more than 2 weeks.

 

[ZoneDymo]

Knowingly sell defective products should probably leave a very long lasting impression, who is to say AMD won't do it again in the future?

Kinda funny that there are people who verbally accused Der8uer who uncovered the truth AMD has been hiding for more than 2 weeks.

what people?
and...if they "have been hiding" for more then 2 weeks its not related to Der8auer as his initial vid was 8 days ago and the conclusion that it was the vaporchamber was only 3 days ago sooo yeah.

And as was mentioned earlier, are we just going to gloss over Nvidia selling defective 3090's that roasted themselves to death playing an MMO? oh wait, they did not know what would happen.

As also was mentioned earlier, its about how the company handles it, will it screw up completely like Gigabyte with their psu's leading me to never buy Gigabyte again?
Or will they act responsible like Fractal with that USB hub problem or Arctic with their coolers and paste and do the right thing removing it from shelves/recalling products and/or offering replacement parts?

Hell in that same realm Nvidia acted properly as well it seems to the burned cards as a result of the new pin connector on the 40 series, we will have to see.

At least this issue with AMD isnt that bad, the product does not break, at most it will throttle and thus underperform but its still usable, certainly not what you paid for sure, but its not as bad as some of the problems I mentioned above.

 

[Why_Me]

what people?

Go back X amount of pages on this thread and you will find TPU posters vs Der8uer ... yes he was posting on here.

 

[TheoneandonlyMrK]

Knowingly sell defective products should probably leave a very long lasting impression, who is to say AMD won't do it again in the future?

Kinda funny that there are people who verbally accused Der8uer who uncovered the truth AMD has been hiding for more than 2 weeks.

Accused him of what, personally I accused him of sensationalism click hunting.
Just like I acuse you of being a troll.

Personally I still think I am right.

Your conspirational tone is possibly excessive, knowing you have a possible problem isn't the be all without knowing the scale of the problem.

Same as 12pin cables burning, or games melting 3090's, no one acts on the first realization, that would also be daft.

The Nvidia fans can rejoice, they're that bit closer to paying double again next time.

 

[Vayra86]

Go back X amount of pages on this thread and you will find TPU posters vs Der8uer ... yes he was posting on here.

We saw the exact same thing 'against' IgorsLab for uncovering that 12VHPWR cable adapter junk.

Its not strange people express doubts based on their own idea of what's normal or not. Its very healthy people are hyper critical of YT/online content, as well.

 

[Why_Me]

We saw the exact same thing 'against' IgorsLab for uncovering that 12VHPWR cable adapter junk.

Its not strange people express doubts based on their own idea of what's normal or not. Its very healthy people are hyper critical of YT/online content, as well.

Let's be honest .. Der8uer and Igor are two of the best ... but they met their match on TPU. No other board imo would have posters with big enough balls to take on those two.

 

[Vayra86]

Let's be honest .. Der8uer and Igor are two of the best ... but they met their match on TPU. No other board imo would have posters with big enough balls to take on those two.

That's why I'm here and not on LTT or any other piece of crap, honestly. The focus here is heavily on no-nonsense, and I love it. Nobody and nothing is sacred. I love that too. That's how you get real info.

 

[Why_Me]

That's why I'm here and not on LTT or any other piece of crap.

This ^^ ... I love this board in a 'Lord of the Flies' kind of way.

 

[nguyen]

Man MSI sure smelled what was going on and just kept quiet&watch other vendor gets burned

 

[AusWolf]

Knowingly sell defective products should probably leave a very long lasting impression, who is to say AMD won't do it again in the future?

Kinda funny that there are people who verbally accused Der8uer who uncovered the truth AMD has been hiding for more than 2 weeks.

Knowingly? Hiding the truth? Why on earth would they do that?

 

[Dirt Chip]

I am confused by the reasoning debauer uses to make his conclusion. like some have said already he bought broken cards and then went on to find them broken and then declares that all cards are there for broken? I get that there are many claims from other people that they also have had this issue, but until proven otherwise thats all they are "claims". Looking back at the 4090 connector issue there were also a ton of people claiming to be having that issue and that turned out to be BS. there will always be people jumping on these type of situations trying to make it appear as if its worse than it is. doesnt matter what brand it is, AMD, intel, Nvidia, there will always be people adding to the claims to make it seem worse. this testing that he did doesnt actually prove how widespread this issue is, just that in these cards that were having that issue he was able to narrow it down to something specific, which is great. but to then say it's an issue with all 7900xt(x)s cards because these broken ones are in fact broken doesn't add up to me. it shows that with the cards that are having issues it's probably related to the vapor chamber. do we know if it's a problem with all vapor chambers? no we don't. we have no idea whats causing it and further investigation is required to find that. maybe there's a specific condition that causes this to happen? maybe it's a defect in the manufacturing of specific batches of cards? all his testing shows is that there could be an issue with the vapor chamber on the cards that are having issues. for someone that has such a huge following this feels totally reckless to be making such claims at this point about needing to do recalls. this would be like if someone said Nvidia needs to recall all 4090s because of the melting cables, without regard to the number of cards actually having issues.

again if it turns out that they need to do a recall then thats what they need to do, but i dont see anything in debauers video that shows that is the case. again this would be like if someone sent him four 4090s that had melted connections and after testing and seeing that they would in fact melt, he then thinks nvidia should do a recall. This just doesn't feel scientific at all how he has come to the conclusion he did.

You also, like others, didnt fully understand his video: the sample of known 4 bad cards only used to isolate the cause of the problem- not to show how widespread it is.
The reason to mention recall is based on 48 confirmed cases by him and remember that he has a very limited reach compared to all 7900XTX owner worldwide so more are sure to exist, and the understanding it is a vapor chamber (VC) problem that no user can fix by himself (it is not a driver, cable, adapter issue). If the VC is the problem and it`s made by automated process than whole banches are possible effects and that`s thouthends if not more. Hance it is seems like a widespred problem so recall, and not RMA, is the right way to hendel it properly from all sides (AMD, retailers, consumer).
As time gows by, we see more and more cases of people showing the same hot spot problem so his assumption of a must happend recall is very much valid and plusiable.

The high widespreadness of the problem is a quite established by now, all is left is to see how AMD handle it.
After all, it`s not a core\electrical design issue so just replace the collar and it`s fix.
Non-reference models can keep on sale with no problem.

AMD can still close the whole incident quietly and efficiently.

 

[kurosagi01]

Well at least for now its only isolated to XTX and not XT, to be honest the GPU market in the 2 years as a whole has been a big joke.
The Eth mining craze, RTX 3000 series launch to now, the 6000 series XT have been a mixed bag, the 4000 series is a joke with price despite being good perf and the silly connector. The 7000 series are also showing problems.

 

[AusWolf]

You also, like others, didnt fully understand his video: the sample of known 4 bad cards only used to isolate the cause of the problem- not to show how widespread it is.
The reason to mention recall is based on 48 confirmed cases by him and remember that he has a very limited reach compared to all 7900XTX owner worldwide so more are sure to exist, and the understanding it is a vapor chamber (VC) problem that no user can fix by himself (it is not a driver, cable, adapter issue). If the VC is the problem and it`s made by automated process than whole banches are possible effects and that`s thouthends if not more. Hance it is seems like a widespred problem so recall, and not RMA, is the right way to hendel it properly from all sides (AMD, retailers, consumer).
As time gows by, we see more and more cases of people showing the same hot spot problem so his assumption of a must happend recall is very much valid and plusiable.

The high widespreadness of the problem is a quite established by now, all is left is to see how AMD handle it.
After all, it`s not a core\electrical design issue so just replace the collar and it`s fix.
Non-reference models can keep on sale with no problem.

AMD can still close the whole incident quietly and efficiently.

What der8auer did wrong is he made an investigatory video with his own personal opinion placed as the conclusion. No one with the tiniest inclination for scientific thinking should do that. Ever.

Other than that, I agree. Let's see when AMD issues their statement and what it says.

 

[ARF]

This is the only time from what I remember that a Radeon has a defective cooler. Something went wrong and if they find what they changed compared to the last time, the issue will become apparent.

 

[lightning70]

The 7900 XTX Reference model's Cooling problem seems to be getting bigger.

 

[pavle]

Most intriguing.

 

[Veseleil]

Although I'm probably not the type of person to jump on a BIG problem for AMD train, I must say I really don't like their latest stuff. AM5 IHS design compromise, and this RDNA3 reference design (let's keep the 450W card at 2.5-3 slots) bullshit. Reference cards from both of the war rooms were always an unobtanium for me anyway.

 

[Dirt Chip]

What der8auer did wrong is he made an investigatory video with his own personal opinion placed as the conclusion. No one with the tiniest inclination for scientific thinking should do that. Ever.

Other than that, I agree. Let's see when AMD issues their statement and what it says.

My view is that the video is very much well balanced, form the headline to the effort being made to acquire 4 test sample to the wide range of hard mod testing to the problem isolastion to the conclusion and the assumption- all according to the definitiv findings.
There is no problem with personal opinion when it is clear that they are opinions.
At scientific article you have room for assumptions (you can call it personal opinion if you wish) in the end- the end of the discussion. There you can hypothesis according to the hard findings you shown.
It is very much acceptable as long you are within reason with that.
This video assumptions\personal opinion are well within the acceptable range, imo.