AMD GPU 6000 series owner should avoid the latest driver from december 22

[Kabouter Plop]

Those are very easy to remove without damaging them, you can obtain replacements easily too.

What about the type thermal paste used ? or you just skip that and pretend its there fault for skipping ?

 

[eidairaman1]

What about the type thermal paste used ? or you just skip that and pretend its there fault for skipping ?

Skipping what?

 

[AusWolf]

Any updates on this?

 

[jesdals]

Not any thing at the site. Gordon at PC World did mention that he had a talk with Kris and he told Gordon, that AMD had send him some cards for testing.

 

[Hofnaerrchen]

Fortunately I run my card underclocked with a 240W power limit.

I still fail to see how, aside from maybe excessive heat, software could cause this kind of damage. I'd expect this from excessive mounting pressure or board flex.

Just watched the video and then read some comments on it.... one caught my eye. It was about used cards from mining rigs. As we all know, there were some major changes to crypto mining lately and I would not be surprised if some or even all of those cards came from a former mining farm.

You do not send cards to repair while you still have warranty to them, so they probably were either quite old or -most likely - a bulk sale of mining cards. He talks about quite some things he is asking his customers but the information if they might have been used in mining and age is missing, so we are back on speculating.

To kill a chip you need to run the voltage out of specs quite heavily without fitting cooling and based on the fact the limit is still the same with the latest driver I doubt that it's a driver issue. Especially when you consider that most cards require BIOS tweaks to allow for higher voltages out off official spec. Quite easy to check with GPU-Z. Maybe someone did sell quite a lot of used cards with new bios - nothing new neither - to make more profit.

 

[RJARRRPCGP]

Did AMD just pull all drivers past 22.8.2 for Radeon RX 6000 series?

 

[Super Firm Tofu]

Did AMD just pull all drivers past 22.8.2 for Radeon RX 6000 series?

View attachment 279946

No.

 

[BlaezaLite]

Did AMD just pull all drivers past 22.8.2 for Radeon RX 6000 series?

View attachment 279946

The 22.11.2 WHQL drivers are still on TPU and this place would of announced something like that.

 

[RJARRRPCGP]

Suddenly, they are no longer offered via normal updates in the Adrenalin control panel.

 

[BlaezaLite]

Suddenly, they are no longer offered via normal updates in the Adrenalin control panel.

Hmm...

 

[A Computer Guy]

Hmm...

Now I'm a little concerned. Currently running 22.11.2 on both rigs. Maybe AMD is just being cautious.

 

[BlaezaLite]

Now I'm a little concerned. Currently running 22.11.2 on both rigs. Maybe AMD is just being cautious.

I only have 1 rig and yep, I'm on 22.11.2... Anything bad happens to my shiny 6900XT and I'll go dotty.

 

[Hofnaerrchen]

Suddenly, they are no longer offered via normal updates in the Adrenalin control panel.

Just checked the official website and 22.11.2 is still available.

A bugged driver would cause system instability but a driver (software) is not capable of causing physical damage like the one on the reported GPUs - cracking dies in particular. Even for a driver to increase voltage you need a BIOS to support this. Why do you think people - who are heavily into OC - do use special BIOS versions or tools to hack their BIOS's? Because the driver is not capable of bypassing the values in the BIOS.

 

[RJARRRPCGP]

Just checked the official website and 22.11.2 is still available.

A bugged driver would cause system instability but a driver (software) is not capable of causing physical damage like the one on the reported GPUs - cracking dies in particular. Even for a driver to increase voltage you need a BIOS to support this. Why do you think people - who are heavily into OC - do use special BIOS versions or tools to hack their BIOS's? Because the driver is not capable of bypassing the values in the BIOS.

I didn't make that up. Yes, it's still at their driver web site, but removed from Adrenalin checking, apparently. Now I click "Check for Updates" and it just says "Up To Date". This didn't happen days ago.

 

[Garlic]

Don't re-paste it if it's a high-tier Sapphire! I have superb temps out-of-the-box, even with it boosting to 2.7 GHz!

eh, I had terrible factory paste for my toxic, 30c hot spot drop after repaste, some guy I know also had similar temps and drops. My 6700xt nitro was fine though

 

[BlaezaLite]

eh, I had terrible factory paste for my toxic, 30c hot spot drop after repaste, some guy I know also had similar temps and drops. My 6700xt nitro was fine though

I'm still under warranty, although it's only 3 months as my card was B-Stock. Might wait until that's finished then pull it apart.

 

[AleXXX666]

for my 6600, all these latest drivers caused performance decrease, 22.5.1 is the most stable for me. Again, I'm back with team green. Team blue is LMFAO for me, prices are not better at all, while performance is not superior and stability is at "beta-test" stage lol

 

[AsRock]

Any updates on this?

 

[jesdals]

Nice to hear that the drivers was cleared off suspicion

 

[Hofnaerrchen]

LOL just LOL

TL/DR:
Not a driver issue - as expected
I was quite pissed about the first video because it was poorly researched and causing more damage than anything else - like this thread. So I was hesitant to watch the follow up and - he asked the customers some of the questions I did leave on the original video - to no surprise it turns out they did send in the cards because they had no invoice/warranty and had bought the card lately. As it was fucking cold in Germany the time they did buy the cards it is more than likely they did put it in the PC right after receiving and it is a VERY BAD IDEA to put anything under the extreme stress that happens when you have very high temperature deltas - you can safely (at least for your hardware when you do this far away from your pc) test this by pouring hot water in a cold water pitcher. Also it looks like they are from the same seller and probably former mining cards...

 

[LLezerus]

LOL just LOL

TL/DR:
Not a driver issue - as expected
I was quite pissed about the first video because it was poorly researched and causing more damage than anything else - like this thread. So I was hesitant to watch the follow up and - he asked the customers some of the questions I did leave on the original video - to no surprise it turns out they did send in the cards because they had no invoice/warranty and had bought the card lately. As it was fucking cold in Germany the time they did buy the cards it is more than likely they did put it in the PC right after receiving and it is a VERY BAD IDEA to put anything under the extreme stress that happens when you have very high temperature deltas - you can safely (at least for your hardware when you do this far away from your pc) test this by pouring hot water in a cold water pitcher. Also it looks like they are from the same seller and probably former mining cards...

I feel nice and vindicated in regards to my own comments

 

[AsRock]

LOL just LOL

TL/DR:
Not a driver issue - as expected
I was quite pissed about the first video because it was poorly researched and causing more damage than anything else - like this thread. So I was hesitant to watch the follow up and - he asked the customers some of the questions I did leave on the original video - to no surprise it turns out they did send in the cards because they had no invoice/warranty and had bought the card lately. As it was fucking cold in Germany the time they did buy the cards it is more than likely they did put it in the PC right after receiving and it is a VERY BAD IDEA to put anything under the extreme stress that happens when you have very high temperature deltas - you can safely (at least for your hardware when you do this far away from your pc) test this by pouring hot water in a cold water pitcher. Also it looks like they are from the same seller and probably former mining cards...

Yeah on top of that the other card he showed did not look cracked and actually looked separated as you could see the die clearly.

Which made me think enough hot and cold cycles you could get a good view what the die looks like haha.

 

[RJARRRPCGP]

I run my GPU fans at least 45 percent. Not loud and cools well, most of the time.

 

[AsRock]

I run my GPU fans at least 45 percent. Not loud and cools well, most of the time.

Yeah same here i keep mine running due to other parts on the card. so at idle the run no higher than 500rpm.

It got it's hottest when i was seeing if the card would crash with the AMD stress and hit 3001MHz with 1070mV +20 and was using 325w and still only 68 on the hot spot although a little unstable.

Temps seem to be stable so i am happy.

 

[Tropick]

LOL just LOL

TL/DR:
Not a driver issue - as expected
I was quite pissed about the first video because it was poorly researched and causing more damage than anything else - like this thread. So I was hesitant to watch the follow up and - he asked the customers some of the questions I did leave on the original video - to no surprise it turns out they did send in the cards because they had no invoice/warranty and had bought the card lately. As it was fucking cold in Germany the time they did buy the cards it is more than likely they did put it in the PC right after receiving and it is a VERY BAD IDEA to put anything under the extreme stress that happens when you have very high temperature deltas - you can safely (at least for your hardware when you do this far away from your pc) test this by pouring hot water in a cold water pitcher. Also it looks like they are from the same seller and probably former mining cards...

*4 guys walk into a repair shop with cracked GPU dies*

The internet: "AMD is LITERALLY trying to DESTROY YOUR PC with DANGEROUS drivers!!!!!! Return your RDNA2 card NOW!!!!"