AMDs questionable Statement regarding the 7900XTX Hotspot Drama
HTML-код
- Опубликовано: 27 июл 2024
- Support me on Patreon:
/ der8auer
---------------------------------------------------------
Music / Credits:
Outro:
Dylan Sitts feat. HDBeenDope - For The Record (Dylan Sitts Remix)
---------------------------------------------------------
Paid content in this video:
- /
Samples used in this video:
- /
---------------------------------------------------------
Timestamps:
0:00 Intro
1:16 When did the problem become known?
2:17 What problems are possible?
3:12 Defect of the graphics card possible?
4:25 The problem with the Vapor Chamber
5:12 Bad Problem solving?
6:07 Speculation
6:36 Replacement not available?
7:56 The defective 7900 XTX
9:15 We "test" the card
9:51 Why the "test" at all
10:30 We open the card
11:05 The PCB
11:36 Measuring resistors
12:15 Summary/Conclusion
13:55 Outro 
Наука
We need more people like you in this industry that we can trust, thank you for your efforts.
Sad that I have to cycle through tech tubers like used socks. Once higher level monetization comes into play, it's time to move on.
The OP makes a good point. We do need more people who have the engineering know-how, and can callout the BS when it’s warranted.
And cats!
Don't forget about the cat.
@@MNeuwahl Chumean, all we need is AMDUnboxed
Yeah, I really used to like Gamer Nexus a lot, because they used to talk about performance and just that, but now it's 70% sarcasm, 15% sponsors and maybe 15% of actual content.
13:20 Agreed, if AMD are able to track these cards, the right thing to do is to recall the bad ones. Sold from their own site so... no excuses.
Top fella, no click bait and very professional breakdown of the card with no over the top ranting and raving like other channels on youtube. Keep up the excellent work
I think clickbait has got to be scummiest human invention since internet advertising
"very professional breakdown of the card" made me laugh. Nice pun
"very professional breakdown of the card with no over the top ranting and raving like other channels on youtube" 👍👍👍👍👍
totally agree, the others are just influence-rs
And a cat.
Nice to see a well-grounded RUclipsr following up on this. It's not about lashing out at AMD, nor should we take their word on it. DerBauer is perfectly stanced on this with some very good points. Thank you
But the assertion was made that AMD wasn't contacting the distributors. Something that he has no idea if they are or aren't. Thats pretty irresponsible 'reporting' IMO.
As a person who does component level electronics repair for a daily job... I can smell that blowup through the video. An autopsy would be interesting.
As my instrumentation teacher would say, he let all the magic smoke out!
As a person who does companion-level hair grooming for a daily habit... that kitty he has on the table is just so full of static electricity. Has anyone else noticed how that hair just keeps getting puffed up higher and higher throughout the video? That cat's capacitance could fry a threadripper. God dang.
please my dear i want xplanation four you. this no xplosions. no possibility four smell and no xcelent graphics processing unit pcbee componentry making four this issue. mayb four you compilcation but every body my country understand four this. if u wanting four me make teach i ready four xplanation your ears. please dear i no u try your best four your branes. but i proud four u if you make improve. thank you my dear
hello, do you think it is a problem of the single model, or of the whole new architecture? and then it will also extend for lower end models (if it's an architecture problem)
@@alsteooaai i have same problem at rx 6600 xt (110 C hotspot)
Great video and you're 100% right. If they know what batch is the issue they should be contacting the relevant parties to let consumers know. My wife knows nothing about PC hardware and when I was explaining the situation even she said "well why aren't they contacting customers directly instead of waiting for them to reach out to AMD"
I think it's because it's not contained to just one batch and they're not quite sure how big or small an issue it is in terms of units affected
Thanks :)
Exactly, if they knew which cards are affected they would go the route of contacting the people that bought them. Just how car manufacturers do it. If your car has a defect, you get a letter from the manufacturer directly or the authorities telling you to bring the car to the shop.
I doubt they are 100% certain about exactly which batches were affected and to what extent, especially this early. Issuing a recall while still investigating a non safety issue is not a smart move either. Not to say you cannot issue a recall by just saying it. A lot of preparation is needed.
Its more cost effective to do it that Way. Not the nice Way but for them its better.
At least they think. I am with you and it should be recalled. But yeah…
@@flimermithrandir Of course it's better for AMD. But the average consumer has no idea about GPU-Z or how to identify this is even an issue for them. They'll just think the card has poor performance when in reality it's just getting throttled really heavily
They're downplaying the issue (which is huge) as much as possible, RMT any affected units, and quietly phase out the issue. They'll likely do anything to avoid a mass recall.
I think so, too. Companies always lies.
When car have recall they only replaced said piece not the whole car. Likewise GPU.
I honestly hope this explodes in their face like a bomb if they contineu to ignore it and not do right, they clearly messed something up badly something much worse then just the cooler, and everyone thinking its just the cooler lets them shove it under a rug, this issue should still be investigated cooler problem is only part of the problem, remember wrongly soldered caps that caused a lot of heat if soldered wrong and even exploded on Intel boards from Asus, something like this could be happening again as well i bet all that heat can overwhelm vaporchamber as well and cause some weird issues.
Anyway just speculating.
@@Resist_JWO_1984 Well they don't lie so much as they skew the words in a favorable or least negative way possible. It's called media training. He's prob obligated to say it the way he does.
@@ktpt3732 Probably quicker to replace the entire card, then fix and sell the returns at a discount.
I agree with you 100%, Roman.
If AMD wants to take market share from team green, they have to do better in all fronts, including customer support.
The problem is, I guess, that an RMA in a highly anticipated product would hit their stock price harder than simply ask their customer to check and contact them if anything is wrong.
Always remember this, guys, big companies prioritizes their stock holders first and foremost, customers come last, if that... despite been the ones who made them a great company to start with.
Hi from Brasil!
EDIT: 9:23 - whoa!! I can smell it burning from here!
"The problem is, I guess, that an RMA in a highly anticipated product would hit their stock price harder than simply ask their customer to check and contact them if anything is wrong."
--> THIS!!!
Note, however, that EVERY company would do (and many have done in the past) the same thing. It would be good if AMD put out a general notification instructing users on how to check if they have one of these defective products - that would be VERY simple to do !! They could even get the customer emails from every seller and email directly.
ive had 2 5800x go back for RMA, they paid for everything, i told em to send me back an x3d instead and they agreed. They paid for shipping back and shipping to them. Zero complaints here.
@@ChrisM541 Yes, they have ways to contact everyone who bought their cards instead of issuing a public RMA. I don't know why they chose to passively wait for their customer to contact them, it doesn't make sense to be honest.
@@gustavo_vanni It does make sense from a business (share price) perspective, unfortunately, but only if they are certain that this issue doesn't affect the majority of cards already sold in that 1st drop released, or that 'not too many' affected cards were produced/sold. It's all risk management.
@@bluej511 Free upgrade?
amd is clearly downplaying it so as to avoid an expensive product wide recall
The product is not tha expensive for them. The just use a high margin
@@pedro.alcatra recalling and having to deal with all the defective products is expensive though. First they'd need to contact everyone affected, refund or replace all of them and then deal with all the defective cards somehow...
Like they'd probably be better of never launching the thing in the first place...
That's because they've done an Apple, and dictated what we want from them instead of listening to us. I mean the claims of efficiency over the competition for both their CPU's and GPU's are all bogus and inflated by 10%. AMD stock will slump hardcore from this generations failures
recalling is worse than unlaunching
Why would you product wide when serial numbers lead to batches, especially if using more than one subcontractor on the chamber? Which is a guess.
Lots of respect for your approach!
kina virus released to fk up the elections...millions dead, billion+ with heart issues...evil know no bounds...Godless people will always create hell !!!
Simply one of the best, trusty and objective, hardware reviewers of RUclips. Keep up with the good work!
I have a speculative reason for not recalling the affected serial numbers, right now most main stream media isn't talking about the issue, the issue has mostly been covered by Igor, you, GN, and JayzTwoCents. That's not a super large part of the market, so likely AMD doesn't want the issue to extend into mainstream media and take the PR hit which isn't unreasonable as it's not a safety issue yet. Customers will either not care about the temperatures and performance or will contact AMD about the issue and get offered the RMA. As long as people are getting the RMA and not being told it's normal by support it's an acceptable (not great though) way of handling it.
My speculation is that the subcontractor making the vapor chamber was doing it just like the heat pipe manufacturing in the Gamers Nexus factory tour of coolermaster, whom I'm guessing also made the vapor chamber in question.
They were loading bins of heatpipes into multiple machines filling them with water. Look at the filling step in the video, you can see an identical machine doing the same process farther down the wall than the one he is focusing on.
If one machine was miscalibrated, and they had, say 3 machines doing the job, then 33% of the vapor chambers from that production run will be bad. (25% for 4 machines, ertc.) They just might not be able to determine which individual chambers are part of that 33%.
@@FuburLuck I saw your comment on a different thread, I think this is most likely the case. Though it's likely the cards would come in a certain order if the machines are running at a constant rate the affected cards would be produced in a certain order. Considering filling the vapor chamber is near the end it's likely the part would be numbered with what production line it came from before there's a chance for them to get scrambled. It's likely the serial number would indicate which production line the vapor chamber came from or at least the cooler. So assuming it's a single digit that will between 1 and 4 for each production line and the busted machine is on production line 3 they could go to the corresponding digit of the serial number and recall all the ones with a 3 in that digit.
@@morgan40654 Are you sure they HAVE serial numbers? I've never seen one on any vapor chamber or heatpipe (from other GPUs I've seen in personal teardowns).
@@FuburLuck Doesn't necessarily have to be on the vapor chamber itself, it would likely be included on the main cards serial number or the coolers serial number. They're quite long and they're able to fit a lot of information into them.
thanks for your effort in finding the issue and now in responding to AMD and AMD customers. Hope this issue gets resolved sooner rather than later.
I agree with everything here. AMD seems to be scrambling to find out exactly what went wrong, while trying their best to appear they already know. I suspect that, after enough customers report the fault, AMD may be able to narrow down the affected units and reach out to everyone they can find. In the interim, AMD should send an email to every customer describing how to identify the problem. I'm sure some users aren't following the news as closely as your viewers.
Neah
If they do this if any company does this and proactively investigates a product issue in such a manner, it would leave them extremely vulnerable to getting sued easily and lose more easily.
They might even get forced by certain regulators to issue mass recalls etc.
No company wants that the negative press, the loss of money to do the recalls etc
According to Scott Jerkelman.... I wouldn't believe a word that guy says. Remember, he was the guy that made fun of Nvidia 12vhpwr adapter. Turns out that the adapter had no issues at all other than users not plugging it in all the way. On the other hand, this vapor chamber issue is way more common and way more costly for AMD. What a nightmare. They don't even have cards to swap for customers that have been affected. What a scumbag move by AMD. That's called Karma, Scott. Cheers
@@mattgreenfield8038 I think you're lacking a fundamental understanding of what occurred there: competing companies will poke at each other because as Wendy's Twitter going viral demonstrated, people enjoy some drama and sass. That move was one designed specifically for advertising and monetary gain.
Further, there were absolutely issues with some of the cheaper adapters being made of inferior quality components, and the saga highlighted that before people bought cheapo ones trying to save a dollar. The wattage ratings on some of those wires weren't even sufficient for the current passing through them.
I do recommend using a bit of critical analysis before name-calling, as that typically only highlights an issue with you when the folly in your argument is demonstrated. Cheers.
AMD does not have email of every customer. What they do have and could do is update the AMD graphics driver to be able to detect this failure (and possibly others in the future) and notify the user about the problem and the solution.
@rustler08 you DO know that every Nvidia graphics card comes with an adapter, right? It's already been proven that it was an accidental user error. Either they didn't plug it in all the way or accidentally pulled it out while cable routing. I had the CableMod 12vhpwr cable before I bought my 4090 day 1 anyway. I didn't want wrestle with 4 8 pins coming into my case. 1 CableMod looks much cleaner. Gigabyte Gaming OC 1699$ US MSRP. This beast runs cool as a cucumber. Crushes every card that AMD is capable of making. No faulty vapor chamber. It came with an adapter. No user actually bought their own 12vhpwr adapter. I did see those cheap knockoff ones for sale. But every single card from Nvidia, or an AIB, comes with the same card. The warning was heard loud and clear. Make sure you plug it it all the way and you won't have a problem. And since then, the same adapters are being used and like magic, no more issues.... but this vapor chamber disaster...OMG. Cheers to Scott Jerkelman and the heaviest dose of Karma anyone has ever gotten. "Stay safe this holiday season". Hahahaha Cheers to you Scott
Thank you for calling AMD out on this. From the first statement they made, I can't believe people have been giving them a pass for placing the entire responsibility of determining which cards are affected on their customers. They NEED to tell people a range of affected serial numbers. Honestly the whole situation feels like they're just trying to minimize the number of cards that get returned and it leaves a bad taste in my mouth.
This entire situation made me exchange my reference 7900XTX for a non-reference version. It is a shame AMD ran into this issue.
,,, du musst mal meine videos sehen ,, die sind bruuuudallll
Damn, you got the 79000XTX
xD
2053, vapor Chamber still not working. 🤣 Thanks and fixed
The solution to this kind of problem is simple: don't buy anything in the first week of release.
companies are getting lazy, launch a new product and let the first customers test it for problems.
And buy second hand on the scalper marked. Not very easy to choose.
@@hansmaulwurf8027 actually very easy wait it out and don't buy from a scalper because scalpers are ones who got the earliest batches and with that will suffer from issues the most
Don't be an early adopter. Ez fix it for 90% of technology.
@@davidepannone6021agreed, let tech tuber/whale do the testing and we can get the best product later.
@@thegreatotaku118 👍 100%
As someone with a reference XTX experiencing the issue, thank you for finding and publishing this issue so quickly.
It was a bad experience overall, and I have no doubt you’re videos are a big part of why I’m getting a refund right now, rather than stuck in RMA hell.
Cheers ❤
Yea, I would definitely take the refund. Get a few extra bucks and just buy a 4090. No strings attached. No excuses. No compromises. Just the best of all time. AMD may not even match it with their next generation RDNA4. That's how good the 4090 is. It wins every benchmark. At every resolution. And what is it 600$ more? Well worth the extra change.
@PlaylistArchive I bought a 4090 because we have never had a graphics card lead by this much performance, over even its own 4080. 4080 and 7900XTX are a dead tie. Hardware Unboxed did a 50 game benchmark between 7900XTX and 4080. They have always been AMD biased. They still show 4080 winning 1440p by 1%. 7900XTX wins 4k by 1%. 4090 is 25 to 30% faster than both of those 2. The 7900XTX is barely 30% faster than a 6950XT. What a weak generation from AMD. Nvidia has never been this far ahead. Facts. Facts. Facts. Other than AMD being cheap, what's good about it? If your gonna spend over 1000$US just get a 4090. The King of Kings. The man among boys. The undisputed champion. The 4090.... best best best 👌 👍
@@mattgreenfield8038 Your suggestion makes no sense. A defective piece of hardware is not only an AMD issue, you can buy an NVIDIA 4090 and still have issues. Also, in order to get a 4090, you'll almost have to double the investment, which is most of the time out of the question, especially taking into account that the 4090 was released previously and, if Shiboe didn't buy it in the first place, he probably won't be able to spend more money now that it was scalped and there are barely NVIDIA cards in stock at MSRP. The best way to deal with this issue is to get a refund and then buy an AIB partner 7900 XTX. The extra money won't be that much and performance will probably be better than the reference AMD model.
@@mattgreenfield8038 Do you realize that you're comparing a 4090, which cost no less than $1850 (the cheapest ones that are listed, though not always in stock), with a $1000 card? That is 85% more money, not just an extra buck. Of course the 4090 will perform better (even if it's 30% better, it's a bad deal). The 4080 is also more expensive. The cheapest ones are currently listed at $1250, which is 25% more money and yet it performs worse in 4K than a 7900 XTX overall. The extra amount of memory will also make the AMD card age much, much better in the next few years, which adds to its value in the mid-to-long term.
@Leo D I would definitely pick 4080 over the 7900XTX. Nvidia has better drivers, better software, less crashing, fewer black screens, and fewer headaches. And who cares about how much more Vram. 16gb is at least good enough for 2 more generations. Why would you even keep a card longer than 2 generations? I don't think you understand how much better the 4090 really is. We're talking nearly 30% better in 4k native. The 4090 is 70 to over 100% faster in 4k native with RT than 7900XTX. Look at Portal RTX. Assuming the 4080 is a tie with the 7900XTX across the board. The 4090 is basically a full generation ahead of both of those cards. We have never seen a card so dominant before. It's not just 30% faster than a 7900XTX. That 30% is GIGANTIC. We're talking about 30% faster than the 2nd best graphics cards ever made. The performance gap has never been this wide. Look at the charts. It's funny how far ahead the 4090 is. It looks like it's from the fkn future, bro. Yea, that's worth the extra money. Nvidia is selling over 160,000 4090s per month. Seems like everyone agrees with me. I don't blame Nvidia. They made the best graphics card of all time. I blame AMD for not being able to compete. Nvidia deserves to be able to have its cake and eat it, too. AMD is just lagging years behind now. It's sad, really. I want AMD to surprise us all and make a 5090 competitor, but it looks horrible for them right now. Jees. I'm gonna have to spend 2000$US or more for a 5090 at msrp. Unless AMD can just show some competence.
Der8auer love your videos, very honest, informative and straight to the point. Thank you for opening my eyes to this issue as no other Tech RUclipsrs I follow have even bothered to do so at all.
I've watched all your videos covering the issues with the AMD cards and I'll be waiting for more.
Gamers Nexus spends months blowing up something. Roman does it on video from the first attempt. 🙂
It's like Steve GN:
- We spent 2 months trying to burn this thing. No luck. No video content out of it on the main channel.
Roman:
- Hold my beer!
That interactive furry frame is awesome.
😸
"Furmark"
Great video!!! And a lot smoke and a lot information you bring up!!
"HEADLINE! 'HOTSPOT VAPOR CHAMBER DRAMA NOW KILLS CARDS!' " - Exact quote from Derbauer
I just hope Intel's 2nd gen GPUs give Nvidia and AMD a run for their money
@n n could say the exact same thing about amd in general
Intel's GPU division is basically dead after their latest reorganization. I wouldn't bet on Intel taking market or mindshare from Nvidia or AMD anytime soon.
Intel never meant to compete at the high end (yet). Maybe in the future but for now they are focused on budget cards.
stop hoping for a megacorp to save you from the other megacorps. They're all megacorps.
🤣🤣🤣 Funny guy over here.
My sapphire 7900xtx had the same problem, luckily with all the attention you guys brought I was able to test it and send it back for a refund. Cheers.
Bro I would've done the same and taken my money somewhere else.
Shame that even the likes of sapphire are being hit with this flaw. Proves it's not a cooling solution problem, no?
@Josh Boucher the sapphire card I had was the same as amd reference card, only the reference cooling is affected I belive, I will buy another 7900xtx once the refund comes through, just not withthe reference cooling, many partners have the reference cooling on their cards, the cards basically look the same so its easy to spot them.
@@piemngz I have a couple questions if you don't mind. Where did you buy the card and where did you return it for RMA? Also, were you hitting 110c on stock settings, or with +15%PL? I ask because I got a Sapphire Reference XTX on Newegg a couple days ago and I'm trying to decide what to do. My card doesn't always / usually hit 110c on stock settings, but if I do any power limit increase the junction temp eventually hits 110 and stays there. I just don't know if I'm actually eligible for RMA and if I should contact Newegg or Sapphire. I like the card and the temps are manageable with an undervolt and not going over +5% PL. But I'm definitely losing some performance and clock speeds due to the junction temps being too high.
@@SLAY3R8888 Hi mate, I'm form UK, bought the card from Scan computers, I was still in the 14 days cool off period (you can return anything in those 14 days without giving a reason, I did however contact Scan , they knew about the issues with reference cooling, and would probably RMA it even after that 14 day time, thy even offered to test another card they had, and if it would be good with thermals, they would swap it out for me ..... but I did not want to chance it)
As for my cards behaviour, it would not hit 110 degrees in something like fur mark and some other tests, but then I have tried running a static test in Cyberpunk 2077, everything on ultra + RT ultra, then I left V standing on the street in front of the first apartment, after 10 minutes or so, the card hits 110 degrees junction(no oc and stock power limit), then It started throttling the power down to 320 W , the core clocks went down also to around 2200 , the longer I left it the bigger the throttling became, it wasn't as bad as some peoples core clock of 1600, but I'm not going to pay £1150 for an overheating card, I'm now waiting for my refund, and I will either go for a 7900XTX red devil or ..... 4080 because the prices on both are mostly even, but I will get more RT performance etc.
I would not risk another reference cooled card, as AMD said the fault lies in the quality of the liquid used in the vapor chamber, this means the if the card is reaching 110 degrees (mine was showing 111 in some seconds) the liquid might degrade over time and it might get even worse (not 100% on this though, because we don't really know what was used in the heat pipes), on top of that, its winter here, so ambient temperature is still fairly low, Imagine if its 30+ degrees Celsius in the summer.
From what I saw, the unaffected cards, and the AIB cards never reach 110 degrees, not to mention throttle, so I personally would not want to keep one.
Hope this helps.
Good luck
Always watch your videos. Very informative and extremely in depth. Please never stop!
My take in this issue is that AMD received a "golden sample" from cooler manufacturer. Later when it comes to gpu assembly, cooler manufacturer ships ones with reduced cost/lower quality/cutting corner.
This is very common in automotive industry. Send the perfect sample to decision maker to win the contract, then send the garbage to the front line assembly
As an electronics technician getting ready to perform a test like you did, as I was turning the system on I would have said “Smoke check”.
Then after the result you got I would have said, “Smoke check good!” lol
Der8auer smashing it again with another top video upload 💪🥰
Another great video pal!!! How much we miss of clear sayings and raw truth in tech section. Your videos and explanations are exceptional with no gray zones. We use grayscale to calibrate TVs and monitors not to make an argument about potentially defective products.
I totally agree with you, AMD should call customers that are affected and recall the products. They should also contact manufacturers selling reference designs and do that same thing. I owned a Sapphire reference model which I changed for another one hoping I would get a good one and still the same thing happened. So I'll return the card and just wait for a NON-REFERENCE model, IF I CAN FIND ONE.
Your absolutely right. Every time I've had a recall on a vehicle the manufacturer would always contact us via email and snail mail to let us know so I could set up an appointment to have them fix the issue. Anything less than that shows a lack of care on their part. They really need to revamp their BS department aka marketing team.
To be fair, it is a car and you could die. Vehicle manufacturers don't aggressively contact their customers because they care about them. They are trying to absolve themselves of liability in any potential wrongful death lawsuits. If there is a known defect with the air conditioner or the stereo, or some aspect of the car's design that can't potentially kill you, they will probably never acknowledge it nor offer to repair it.
You guys say this but the reality is vehicles are tracked way more extensively than electronics. The only way an electronics vendor is going to know YOU bought their product, AMD from AMD's website included, is if you go and register your product. All kinds of PC part products have product registration from the vendor for that, and AMD has it for AMD Rewards. They aren't going to know you have whatever serial number until you do that.
TL;DR AMD doesnt record what serial number went were. I'd love to know which PC vendor out there automatically does this.
Quite the opposite actually. One of the recalls was a faulty block heater. I received an email and a letter. Then the dealer called me to set up an appointment to have it replaced. GM.
@@IRQ1Conflict And what was the problem with the block heater? A potential short circuit was causing vehicles to catch on fire, right? So, like I said, a safety hazard.
I also dont think they knew about this beforehand. If they did and released it anyway and this happened there would be a head rolling ceremony.
Yeah, but they SHOULD have known. Specially given that it affects cards that or mounted normally. Not much of a stretch to say that if they really didn't know anything then they also did not have any quality-control to make sure the cards actually can perform what they are specified to do.
There may well be a head rolling ceremony anyway - though tbh there is no such thing as bad publicity /shrug
not amd fanboi but just sayng u have to be crazy out of ur mind to realise cards with defective coolers not a chance they knew it beforehand
@@ABaumstumpf I think we have to wait and see how widespread the issue really is to determine that. These guys have shareholders to answer to and if this is something that could have or should have been avoided and they end up recalling a ton of cards you will probably see people get shit canned over it.
@@ABaumstumpf That is a difficult issue to QC cost effectively.... its literally cheaper for the customer for AMD to assume 99.99% of cards are AOK... and RMA the rest.
Great video, Roman. When Scott says "small batches" I'm assuming it is a random event not necessarily tied to a serial number range. Otherwise, disclosing the impacted serial numbers would be a much better approach in so many ways.
Agreed, You show great wisdom in your discussions and content. You do not jump to conclusions and give all sides the benefit of the doubt. I agree that the more RMAs AMD gets in they can narrow down the effected batches. It is unfortunate that there is always that lag in companies doing their recalls. Yourself, Steve, Jayz, and Linus are all great for getting things out there for testing and working together and bring great info to our Gaming/PC Building community. Iam impressed with the care and wisdom you share. Thank You for being there for us der8auer!
I think the assumption that the supply chain being joined up enough that the AIB partners know the serial number of the vapour chamber and which card it was installed in is very optimistic.
And also depending on the manufacturing process and the way "batches" work, if a single filling machine was defective but the batch has multiple machines in parallel then it could be multiple batches but only maybe 20% of them in a batch, therefore to recall by SN would be more costly and harmful that recall by defective performance
Really appreciate your professional approach. AMD should treat their customers professionally aswell and not rely on the cards failing for returning the faulty ones. That is as you say not very nice but also damages their brand so much more than a recall would do. If there was a recall customers understand that sometimes there are problems in production. If they cannot respect me as a customer when I buy a product for 1300 euros I will not buy from them again.
A couple of years ago I got a notice from Apple that my IPhone was manufactured within a month where they had problems with the batteries failing excessively and I could just come to any Apple Store and replace it for free. Which I did. No big deal, I don’t think worse of Apple for that. But it seems both Nvidia and AMD have forgotten how to treat customers. I think a lot of us are on the brink of moving away from gaming due to all these scummy business tactics.
You know its bad when Apple treats their customers better than Nvidia/AMD.
I completely abandoned AAA gaming a few years ago for these reasons + how scummy Ubi/Act/EA were behaving, the whole industry just feels like something best avoided.
While I wasn't a generational upgrader I would buy fairly high end hardware. Now I just have a couple 5700XTs for VMs and a few indy games here and there, have saved thousands and don't miss any of it at all. When it stops being fun, stop playing.
They did replace my 5600x under warranty, but is was murder. It had a cold bug, that only (and always) crashed Chrome (on both Linux and Windows LOL!). So I could do anything except browse the internet on it. Sent it back to the store, they tested it hot running games and went "no fault no fix", so I had to send it all the way back to AMD and it took forever. They seemed to lose it (signed for arrived, but they never found the box lol), so just sent a replacement to me after their time limit on replies expired.
Sad the store I got it from were too silly to do a cold boot/browser test. Cos every other chip in this PC worked (the replacement then my upgrade), so can't see it being user error my end (like bios or whatever).
@@Gandalf721 As always, your statement highly depends on what you're looking for. You're not going to be playing a lot of demanding recent games even with settings turned down if you're looking to do 1440P high refresh rate or 4K. Sure some people don't care about resolution or refresh rate but many do
@@Gandalf721 Was still a mid range card at the time compared to others, I do a lot of CAD/rendering hence VMs get their own card. But I guess I'm just silly...
Point was I'm buying cards that will do for work with no consideration for extra gaming performance anymore.
Well said, those were my exact thoughts at the time as well. IMO, there's 2 options available to them:
1- Recall the batches if it's that simple.
2- Release a simple tool to customers that runs a test then outputs a result confirming who's affected. The user should get easy instructions on how to proceed and/or an option to simply send a result back to AMD as proof, if needed.
No affected user should be without a working card while they wait for a new one. They should be placed in a queue then shipped a replacement with a return label to ship the old card within 60 days or be charged a fair price.
Yes a minority of people will try to abuse this system. That's why you sign contracts with manufacturer to ensure they stand behind a significant QC failure. Forcing users to jump through anymore hoops than this is just sweeping it under the rug and will make even more people reluctant to go Radeon at a time they should be winning back market share.
Thanks for the follow-up video, especially with the magic smoke you experienced attempting to test the issue.
We gotta be sure about what's going on before we just say what something is. No need to make yourself a target for a lawsuit without cause.
And AMD had the audacity to make fun of Nvidia for their power connector.
And rightfully so, now they both are looking like a clown, so no fan can feel superior. a balanced situation.
Yeah I find it very weird that they say its "small amount" of cards, yet the broken cards show everywhere over the world. Its not like there is one shipment to germany/usa that was screwed.
karma there, the gpu market is currently a circus though
@@stevenandino8178 not really. The power connector affected like 0.05% of all customers or something and it failing relied a lot on user error. (It's a stupid delicate design but most users plugged it in just fine)
This AMD thing is just defective cards straight from the factory and seems to affect a lot more people (the power connector thing was really hard to reproduce in realistic conditions yet here youtubers already have multiple defective AMD gpu's bought from users out in the wild....
@@indiesigi7807 yeah NVIDIA RULES, AMD DROOLS! XD
(I really don't care I just like messing with the fanboys, Nvidia is better though:p)
"testing" the card was extremly excite to watch!
9:22 Was the best part of entire video 😳
Good that it was captured by camera. Otherwise no one would belive it!
After failure is it under RMA? Or just extremly expensive paper weight?
Right now it's just a paper weight 😂
@@der8auer-en I have MSI 970 Gaming X with fried VRAM. I dropped a screw when doing maintenance cleaning, which cause a short and some sparks came out. Someone told me it is repairable but costs of repair are way higher than buying used one and of course working fine. So currently is used as a trophy on book shelf in my gaming room. Looks cool between boxes of PC harware. Mostly Asus Strix stuff.
Yours looks much cooler even if it's completly dead
Sorry, 960 Gaming X. 970 Gaming x is working fine in my old PC for very old games before age of IPS panels
Lots of products are sold without any serial number attached to the invoice. It's a common problem - I recently purchased an LG C2 + receiver. None of the them had any serial number on the invoice. It might be the same case here.
I think there is another scenario. They feel like a bigger percentage of customers will not even notice.
The problem we have here in Europe is that between the 7900 xt and xtx there is not 100 dollars difference. In fact it is actually 250 to 350 Euros difference. The price for the xtx on AMD today is 1.132,20 €. If you go on Ebay scalpers are already selling their cards. If you buy one and it is faulty you may not be covered by the warranty as it was sold by a private seller. Be careful people.
It's really simple - don't buy from scalpers!
@@karama300video _[TL;DR - If no one buys from scalpers, they will cease to exist, and things return to normal!]_
Exactly! It really IS that simple, too!
We're not talking life-or-death, mission critical hardware. It's a graphics card for your personal computer. At worst, yes, it impacts some, whose livelihood revolves around needing a capable GPU, but even then we encounter a good lesson to practice: if you're upgrading hardware and plan to sell what you've replaced, DON'T _IMMEDIATELY_ SELL IT! (keep it for at least a month)
That way, if any issues come up, you'll easily be able to keep operating by using the old component(s). But I digress...
What's even _better_ about not buying from scalpers, *_period,_* is the fact that _you no longer are validating their decisions to scalp!_
Again, it's *that* simple. They're taking advantage of you/us. This isn't a limited time only product, or something that's no longer in production, or a which has a multi-year waitlist to buy... so don't justify what they're doing.
You take away their incentive ("easy money") and they'll stop doing it. That's just how it works, and is why this wasn't an issue in the PC industry prior to 2020.
As for MSRP... Well, nVidia has suckled from Greed's Teet, and I'm not sure they'll allow this to return back to normal. It's one thing to simply _"stop buying from scalpers",_ but that's because it's a middle-man...
We consumers become complete idiots when it's the manufacture's that price-gouge the fuck out of us, and all reason of logic goes out the window...
_"I must own this no matter what, or I won't be cool/good/capable!"_ (aka "sheeple") - or - _"If they charge that much, it MUST be good!! Shutup and take my money!"_ (aka "fanboys")
But, again, chances are we don't *need* the product, and can make do with what we have for a bit longer.
_The exact same thing_ will happen as it does with scalpers, where a message gets delivered: consumers won't stand for unjustified prices. The problem will sort itself out.
(this is called _"speaking with your wallet",_ since companies often don't give a shit about customer's opinions, but they *do* care deeply about money... take that away and they'll sing a new tune you like!)
_[also, yes, AMD is playing the me-too game with pricing, which isn't acceptable behavior, but on the flip side... at least they're not at nV's 🦇💩🤪 pricing......... yet 😒]_
Isn't that _with_ VAT? We have sales tax in the US, and while it varies by location, it nonetheless means actual prices are higher than the stated MSRPs. Where I am, the 7900 XTX would be about $1,087.67 for the reference design.
@@sonicboy678 Everyone forgets the US doesn't put tax in the sales price.
@@sonicboy678 Yes , that is with Vat ( MwSt) in Germany which is 19% at this present time.
Love the stuff you do here, not many out there can.
Thanks for another great video Roman, but where did you get that shirt?
Roman, You, Steve and Jay along with a few others really do such a great job of holding companies no matter their size accountable, fairly that helps all of us consumers in the end. Great work.
That interview... He was trying to do "damage control". The funny thing is that he was the one mocking the competition for the first drama
Karma comes faster and faster in the end days
Once again thank you for making these informative videos. I don't even own a 7000 series card but I still find this stuff very interesting.
I sold my 6900xt last year to buy one wish i didnt ill wait til q4 next year for a 8900xtxh
If car manufacturers legally have to do the same thing for defective cars with recalls, then tech companies should be held accountable too. Luckily here in Australia we have really awesome consumer laws that protect us from this kind of nonsense. So if any Australian buys this card from any distributor they can legally RMA it no questions asked by law for a full refund. Even if it is outside the manufactures warranty, it may then deemed by the Australian consumer commission that it should be covered up to a certain time period depending on the price of the product and what it's used for.
another great video thanks for the update on these and i agree if amd had a handle on this situation and isn't calling their customers to send affected batches back its not ok
AMD is really shooting themselves not in the foot but in the head, this is not a batch problem but a design problem because 6900 reference did the same when you turned it around but now someone finally made enough of a stir for AMD to notice.
Yup. They've been saying 110c is normal since 5000 series
Really guess i was lucky with mine then altho i only tested it for 1 month on reference cooler and i just ran the fan curve to 100% for a month before going watercooling, hotspot hit about 80c back then i believe 50c 60c i think on edge temp
It's possible AMD isn't 100% sure which cards are affected by serial number. The cards were put together by sub contractors, so things like repairs to the vapor chamber filler machine might not be documented well. So they might not know what cards to recall. It could even be a micro leak in the copper and thus they don't want to recall all the cards of the effected dates because they aren't all going to go bad. What I will say is if you have one of these AMD cards. Test it in the normal position in a couple of months if you're not using it in that pos.. (the fan down and the card sidways like a desktop tower. ) Why a couple months? Well, if it's micro leak it might be okay today. (Make sure the warranty doesn't expire first) and two they might figure out what serial numbers are effected in that time. (Saving you from testing the card.)
Yep, my first thought just after hearing about the high temps and throttling down a bit, -what if nvidia asked the heatsink manufacture (might be "coolermaster") to make a boooboo as a revenge for the AMD trying to pick on nvidia for the new power connector during the rdna3 launch. :P
A jokingly thought but hey, who knows :D
Thank you for the good you do with these videos for people who are computer enjoyers!
G'day Roman, Shiek & Makita,
Happy New Year
9:22 😲😭
Totally agree it should be AMD reaching out to customers to get this resolved, not waiting for the customer.
Great video very fair and measured response to AMDs statement while still highlighting issues. great work.
On a side note that is one of the best clips of a GPU failure I have seen, lucky/unlucky to capture that.
I must say, just because you've identified the issue doesn't mean you know which batches were impacted. Perhaps the issue was a faulty water metering system during manufacturing that impacted cards randomly. Maybe they are hoping it is just a few dozen cards or maybe a few hundred. I suspect it is likely they don't know which serials are thus impacted. Still not good... I wonder if it is possible to have people start posting their serials to see if there are any similarities to the lines.
In that case they should not say "one batch" 😁
@@der8auer-en to be fair, a small batch doesn't entail one batch... but it can certainly be taken that way. I found your comment on the PCWorld video and if it is correct that there are already 100s in Germany alone, that really doesn't bode well: I wouldn't' call that a small batch... unless it is isolated to Germany (which certainly doesn't seem to be the case).
@@der8auer-en If there are 10 benches in the factory that fill the water chamber and only 1 was defective, they might not be able to tie the exact serials to the faulty bench even though they know which batch and serials MAY be impacted.
Of course, if this is the case they should just tell us!
Excellent use of logic, great commentary. Well done!
Superb coverage. Thank you, Roman.
I wish that I could give this video two thumbs up. Technical expertise AND ethics!
If AMD really knows the batches affected, a targeted recall is the only real solution for the situation. Otherwise, like you said they really don't know, and they didn't learn their lesson and are not being transparent about something yet again.
There is a far simpler explanation:
1. AMD knows the technical cause - insufficient water in some vapour chambers
2. AMD traced it to a source - a defect in one of the filling machines on the assembly line, for example
3. AMD *can* tell the range of cards that *could* be affected - like all cards that have been filled on this particular filling machine between start of production and date X, possibly with serial numbers, *but* this does not tell them much because that includes a *huge* number of cards, for example it could be 10% of the *whole* production amount so far if there are just 10 filling machines on the assembly line
4. AMD *can not* tell for sure if a particular card from set 3 is defect or not *and* vast majority of them are not expected to be defective - imagine that the defective machine *only* undefilled the cards 5% of the time, at random
5. AMD never said that they could determine if a card was defective or not *from serial number* - they could have ways of testing a card they have *in their hands,* for example by doing a quick X-ray or ultrasonic check of the cold card in a particular position so that water line inside the vapor chamber could be reliably detected and from its position one could see if there is enough water inside or not
With that in mind, it makes no sense to recall a huge number of cards when only a few of them will be defective. On top of that cards going via multiple vendors, including cash sales in shops would make such recall really hard to complete.
What they *could* do is add a function to the driver to detect the issue at runtime and notify the user about the RMA possibility right there and then.
Thanks Roman for this great video, fully agree on this topic. From a electrical engineering point of view this VRM looks very well designed and should last the power requirements for this card.
I'm curious if this is a case of AMD buying too cheap mosfets or other components for their designs because it fried instantly when powered up, that is not a typical mosfet behavior from my perspective. Keep up the good work!
The real question is when did AMD find out about the defective batch of vapor chambers. It's possible they found out earlier but figured it was OK for most use cases. It's possible they don't track vapor chambers batch to card serial numbers (but unlikely). Their tepid response leaves much to be desired from a consumer standpoint. Theor only saving grace us that their competitors are so much worse at damage control.
They alreay had this issue with the 5700xt models back in 2019. GN did a video about it.
@@mesicek7 and that makes them good at damage control--putting them under the rug for generations. Nvidia can't even do that lol
@@AkitaMix Is that supposed to make them better than Nvidia?
3:50 "It could be that the card was defective anyway..." Isn't that worse though? It could point to the cards having more defects than just the vapor chamber. It's probably better for AMD if cards are dying just due to the defective vapor chamber, because they've at least identified the problem and it's not that hard to resolve, but if there's other issues that could result to cards going up in smoke, like you just experienced, that's a much bigger issue.
Just FYI there are aways 1-3% defective products when it comes to electronics.
@@DavAnderson Yes, but there isn't always a 1-3% chance of a product becoming a fire hazard. If something simply doesn't work you RMA it and no harm no foul, but if it blows up it could cause all kinds of damage to the other components in your PC or even start a fire. So even a small percentage of failure is unacceptable for products that sell in the thousands. Obviously we shouldn't rush to judgment, I do hope this is an isolated incident and not indicative of a widespread failure in AMD's QA for these cards.
@@chrys9256 I can't see any fire in this video
@xXDESTINYMBXx 1st off the card was producing smoke after it failed, but let us remember that after disenssembled, it was revealed to have been blown mosfet*** the cause of that is different then the vaper chamber. Let's remember that. I am not defending amd but the one-off situation regarding one or two cards regarding them dying off isn't great, but let's not hype it up until we get more information. The 110 c issue is horrible and needs to be resolved as soon as possible though.
Excellent video. Your integrity is showing and I think we all can see it. Keep up the great work!
My guess, it's a design flaw and they changed production of the cooler... The affected batches would therefore be all made before CES, but no chance they would want to say 🤷♂️
VRM efficiency is related to temperature, so while there is no proven causation, there are examples of correlation. My biggest problem with the issue is that customer support should not expect average customer to understand that GPU temperature does not only affect the GPU. People who have defective cards with thermal runaway should shutdown their computer until the problem is addressed and not rely on the product throttling. Customer support should not go the opposite way and promise users it will run fine.
Agreeing with you Roman. What AMD is doing, is the absolute bare minimum. There is no proactive customer service what so ever. The strategy seems to be to downplay the quantity/issue, and maybe hope, that not every customer even realizes the problem.
My guess is that currently their information indicates that the batch(es) with affected parts would require a massive recall (due to the number of total units in each batch, not the number of affected units), and they weighed that would be worse PR then "hey contact us if you have any issues". Other guess would have been that they want to do it once, they need people to contact them to know which batches were affected, but they could still recall the known ones and wait for information on whether or not they'll have to do another round.
What I'd be interested in is if they contacted the retailers for a recall of the yet to be sold units.
Great work Roman. You should be at a million subs with your reporting.
Holy shit. When you powered that card on, the last thing I expected was lots of magic smoke. I'd honestly be surprised if the vapor chamber had anything to do with that kind of failure, as most cards these days have some kind of fail-safe throttling to protect them, even under extreme conditions. It's a shame watching AMD completely bork the resolution to this problem for their customers, which will only further push people into overpriced Nvidia purgatory. Both brands offerings are expensive, but people would rather pay more and have the card actually work.
Thx Roman for yet again a interesting and clear analysis. It shows how little consumers really understand their pc.
100% alligned! In term of quality service management as there is risks of hardware damages on a new card just exit out of the box, AMD have to manage this crisis by identifying the part number cards potentially related with the issue and at minima, commucate to Customers to ask them to contact AMD support to organize the replacement. Having all those cards with potentials issues will impact the sales and not only on new builds... it is a question of trust with the image of the company... It is something that any company may not manage too lightlyand like it seems like it is here...
Roman: definitely did NOT make it successfully through the "smoke" test! 😄
man when the card almost burst into flames I was so surprised.
how to fry a card in 2 second AHAHAH 9:18
I'm currently "on hold" (AMD's words) for an RMA of my 7900 XTX as they don't have stock (Ireland/EU). I did have to email them asking for an update on the status of the RMA to get that nebulous info! No date or timeframe was provided
Thanks for this info
@@der8auer-en Thank you for putting these videos out, as soon as I mentioned your previous video to them on the first call it shortened the back and forth to them acknowledging they needed to replace it in the first place, "without officially commenting " on your video that is *chuckles*
They probably have stock, but it's not everywhere where people are with this issue.
@@xXDESTINYMBXx True, I'm just saying what they told me, selfishly I don't really care what the excuse is more that I am still waiting on a replacement. Here is the direct quote from AMD Customer Care when I asked what was happening with the RMA process:
"We have placed you on hold to receive a replacement and will notify you as soon as new stock becomes available. We are currently awaiting inventory, which is being expedited to our warehouses."
@@SascosCorner I hope you can get a replacement soon
You are 100% right i hope there is more people like u in tech industry and also people in like tech service and support cuz the way you have tested this graphic card and the way u talked about customers support should be like and how AMD should step up it in which way its the only good and right thing god bless you.
Thank you for all your hard work. AMD has to do better. A recall of all units affected should be done.
It's likely only some of the cards _within_ a batch are missing the necessary liquid. If they order a recall on every card it means RMA costs for potentially thousands of properly manufactured cards. It's not a safety issue and as such they won't have a problem letting the user figure it out themselves. The fact they specifically mentioned it was not a safety issue is what makes me think this is the case. They wanted to clearly convey they did not need a safety-mandated recall program so they could allow users to deal with it themselves.
You're assuming AMD has records of which coolers (w/ serial) was put on which PCB (w/ another serial), then boxed (w/ yet another inventory number) and send to a distribution network (w/ a different inventory system), which sold them (w/ yet another order number) to customers.
Now do that for every single part in the GPU.
From my (albeit limited) professional opinion, part traceability is hard and above all EXPENSIVE. I'm sure the risk of batch batches was factored in and they concluded that waiting for RMAs (or issuing a mass recall by date range, in case of safety concerns) were the most cost-effective options.
@n n They might have some idea of when a batch of component was used (which is often enough in the industry), but full traceability? From a sub-constractor's serial on a sub-component to an international, private consumer? No, unlikely for something that isn't safety-critical. That's the kind of traceability you use on seatbelts and airbags, not GPUs vapor chambers.
"Don't jump to conclusions" is a very good advice and not just for PC hardware.
Extremely professional. Excellent content.
You know I was fairly impressed with the statement and still am that they largely 'said the right thing', but I'm glad you made this video to highlight they may not be 'doing the right thing' at least not full marks. Perhaps they aren't certain with the batches, that does mean full recall I suppose.
They could probably want to collect as many cards from returns first before giving out serial numbers. If a batch is missed, it'll be another hell of even worse PR.
From a PR standpoint they definitely said the right things. But not from the standpoint of people who paid 1300€
There is saying the right thing, and then there is backing it up. I have one of these GPUs and want AMD to succeed here... But their default response to date doesn't give me warm fuzzies. If my gpu we're to fail at this point I too would probably ask for refund and jump ship to NVidia
If the card died due to vapor chamber - then placing that cooling solution on another card you would see how this particular cooler perfroms, as you have other reference design cards.
Man I just love how Der8aur give constructive criticism and does it while maintaining being friendly and respectful and keeping composure.
Lots of respect for this, even me being the AMD fanboy I am, love this example of calling out inaccuracies but being respectful the whole time. Super cool and super hard to do.
Cheers!
If they did not know anything about this before release. Then they did not do adequate testing since there are way too many.
Why are we failing to consider why 110C is such a problem in the first place? Is it because soldier can flow at 110C? Is it because the functionality of some on-board devices can be unpredictable after hitting 110C? etc. Well almost all of the above is possible after hitting 110C so yes your failure could every well be because the card hit 110C but near impossible to prove unless there is some visible soldier short that was not caused by the failure directly. But this is why the card is not supposed to hit 110C during normal operation.
Nvidia: Burn the socket
AMD: Burn the card
Intel: burn THE WHOLE PC
I think his statement that they'd replace it right away may have been sincere but he may not have expected the amount of RMAs in that period of time. At least they offered the refund alternative.
"At least they offered the refund alternative."
It should have prompted a recall though. What they're doing now is trying to minimize the impact on themselves by hoping a large segment of the consumer base with defective cards never realise it.
Don’t ever take a statement from a company head sincerely…He might not even be aware of all the details yet, but more focused on the actual event for their CPU.
Lol? When this first happened amd said 110c is normal and won’t give Rma. Amd is a joke
I had the exact same thoughts when I watched that video. Keep up the great work!
the thing is with the them knowing which batch is affected, they don't know exactly which ones in that batch, because not every one of the cards in that batch would necessarily be affected. how many customers are in that batch that have a perfectly good card and now your telling a customer they have to send the card back before they even get the new one. I for one would be pissed that i have a perfectly fine card and have to ship it back and wait for a new one to show up.
this is like power steering hose's on f250's when i was a technician at a dealer ship, they knew what series of trucks had the bad hoses but we had to go through every truck in that series that we purchased and check the number on the hose's to put it on hold for repair or open to sell the truck because they didn't know exactly which trucks in that series were affected. we as the customer to ford (the dealership) had to check.
i don't think what he said is unreasonable to call AMD and say hey this card is getting hot can i get another one.
I tried to order a 6700 xt and it didnt work well from first install. So i returned it got a plain 3070... Asus stable and works. Specifically installing drivers would not complete.
To be precise AMD said they know "which cards are good", but that doesn't mean that they know know which cards are defective. Very likely they know which batch(es) have affected cards in them, but I'm pretty sure that not every card in those batch(es) is defective, so a recall might be quite expensive and more wide ranging than needed. (Although of course if I was an affected customer it would be nice)
Great video, you saved me over $1000 as I was about to buy this card next week.
I would do this experiment, mill off a small amount from the top of the vapour chamber, in a vertical setup, just large enough to insert a needle with water in it, when it hits that 110º inject some water and watch if anything happens. If it drops, we'll then know 100% the problem. It would appear between AMD and the Nvidia release competition of their cards, standards have dropped to an all-time low, for this reason I have bought the ARC graphics card instead.
Great video with some really good points.
FYI - they aren't reaching out to customers on purpose because they don't have enough cards to replace all of them at once, and then they'd have to offer customers a full refund when they would inevitably get upset about the long replacement lead times. So they are doing it this way so they can get a slow stream of defective cards in and refurb them to send onto the next group of customers that open tickets with their support for this issue. This will allow them to phase out the replacement out and not spend money on manufacturing extra cards to support a recall. Also, I suspect they don't even have the materials and manufacturing/testing throughput to make that happen even if they wanted to.
Thank you De8auer, excellent analysis! 🙏 agree with your position, AMD could have handled this a lot better vs putting the responsibility on the customer who is out of pocket 1,300 euros to test a problematic product that AMD is already aware of. Wish there was a 💖button for your video, keep up the excellent work!
AMD no se puede manejar ni mejores y como puro CPU de intel es que controlar mucho asi y ese GPU de AMD esta descuidando sin sentidos
Knowing how to make sure a unit is good/bad (ex.: by weighing them) isn't the same as being able to narrow it down to a specific batch. Maybe it is 5% of vapor chambers across multiple batches that are affected and there is no specific serial number range for the defective ones.
Stop shilling for AMD, bro.
That would make it an inherent weakness with either the prosess or the design and would be much worse, since it could happen to any card launched going forward.
Scott did right PR in saying “yeah, there is a problem” but he also had to do damage control, hence the “small batch”. Plus, we don’t know for sure how many they built worldwide to know for a fact how “small” its batch is.
As you have proven, some cards do work in one position while others do not work properly at all. If yours work in a position you are never going to change, then it is a defect with a workaround, so maybe you yourself might opt not to return your card. Or maybe you got this card to put on custom loop, so it will have plenty cooling.
The bottom line is they are willing to replace the card by a working one, supposedly no questions asked. If this keywords “no questions asked happens, I would not see this as a major disaster.
Not going after customers is also expected, but again shows the batch might not be as small as thought. They would if this was a mandatory recall, then they must inform customers and reach out to as much of them.
this actually explains a lot considering the mix of results I had seen in reviews