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

[BlaezaLite]

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!

It was hitting 80C out the box. With a little undervolt and fan curve adjustment, now its at 69C in CoD. I've got some Kryonaut I wanted to use up. It hits 2.5GHz, but it is air cooled.

 

[RJARRRPCGP]

It was hitting 80C out the box. With a little undervolt and fan curve adjustment, now its at 69C in CoD. I've got some Kryonaut I wanted to use up. It hits 2.5GHz, but it is air cooled.

I usually never see 80C, unlike my Ryzen 7 5800X, which is a different story! Rarely higher than somewhere around 65C, maybe the low-to-mid-70s?

80C out-of-the-box, is like the meme of Ryzen 7 5800Xs! (and of course, zero-RPM)

 

[tpa-pr]

PowerColor 6950XT owner here with some anecdotal evidence. Have been running the Dec '22 drivers since day after they released with zero issues. Mind I'm only running 2x 1080p monitors and not overly taxing the card with anything except Elden Ring, but still. Rock solid.

I'm a layman at this stuff but if people are interested I can provide some voltage/temp metrics under load for investigative purposes. Maybe an experienced eye might spot an anomaly. But even from a layman perspective the type of damage done seems a bit out of scope for bad software.

 

[AsRock]

Co-incidentally, I messaged OCUK about re-pasting my card to see if that would void the warranty only 2 days ago. Might be a sign to leave it alone...

Maybe not this time as you say 2 days haha, how ever what about if you want to the manufacture about it ?, as i see it that's bs they stop you from doing that well unless they paying all the shipping at least.

 

[Ando]

Been using the 22.11.2 driver on my 6800 for a couple weeks I think, can’t remember exactly when I decided to update from 22.7. I noticed that the newer drivers were only for the 7900 series, so figured this WHQL driver would be the most recent stable option for some time.
Had no problems, card runs at sub 70 degrees on hotspot with an undervolt at stock clocks. Same as on the prior driver.
There really is barely any information given about this, I think he should have said more in the video. How many of each variant of the die had this issue? How many were aib vs reference? He didn’t even give the specific driver number, just said “december 8th driver”. He might be a stand up, knowledgeable guy, but this video does little except scare people.

 

[AusWolf]

So... what does "December 8th driver update" actually mean? I'm even more confused now.

 

[Bomby569]

This may have already been asked, but does this affect reference cards only or is it AIB's too? I'm on the 22.11.2 driver and have the Sapphire RX6900XT Toxic Air. No issues at all.

one thing drivers definitely can't do is distinguish between them

 

[A Computer Guy]

Been using the 22.11.2 driver on my 6800 for a couple weeks I think, can’t remember exactly when I decided to update from 22.7. I noticed that the newer drivers were only for the 7900 series, so figured this WHQL driver would be the most recent stable option for some time.
Had no problems, card runs at sub 70 degrees on hotspot with an undervolt at stock clocks. Same as on the prior driver.
There really is barely any information given about this, I think he should have said more in the video. How many of each variant of the die had this issue? How many were aib vs reference? He didn’t even give the specific driver number, just said “december 8th driver”. He might be a stand up, knowledgeable guy, but this video does little except scare people.

I just got a 6700XT the other day for my gaming pc. Although the video reported the issue was for 6800/6900 it did get me a little worried cause after I installed my 6700XT it broke my ability to boot after uninstalling the old driver before installing 22.11.2 and my CPU fan now does something it never did before when gaming. Probably totally unrelated but still unexpected.

After reading a bit my top theories for this GPU cracking problem are
- Manufacture defect, bad batch of silicon
- Liquid metal problem, bad batch reacting poorly with GPU die and/or heat sink
- Mining card related problem, severe overheating
- Driver issue causing GPU to be destroyed physically (seems unlikely to me)

 

[Morgoth]

says i have the lates version 22.11.2 date 30-11-2022

 

[Kabouter Plop]

Im highly suspecios of frequency / thermal / voltage limits etc being ignored the first one frequency limit you can easily test firing up doom eternal while on stock and then on custom core tuning but applying the default suggested value and checking max frequency while you fire up doom eternal, and then trying it again but setting it 1 mhz below suggested value.
Seems to only ignore it on default stock
my gpu-z reports boost clocks of 2250 mhz but will boost 2555 mhz when i fire up doom eternal.
And it also does this if i set default core tuning of 2479 mhz but if i set it to 2478 mhz instead 1 mhz lower it only boost to 2468 mhz, this what makes me very suspecious.
Im glad i run watercooled card but given driver issues if been having im very suspecious, AMD ignores me cos they probably know i am right, i know they read other forums as well cos they told me they do.
There probably even reading here.
Nvidia has killed gtx 480 in the past with drivers, mistakes happen, AMD will make them as well but 6+ months of mistakes is ridiculous.

If been 3-4 days stable now tho after i updated ms edge dev build and set it to new webviews2 which bassicly edge vision of chromium for apps, and not all chromium or let alone webview2 apps let you disable hardware acceleration, and outlook appearntly crashes due amd gpu driver for one user so i decided to update my self to dev build, cos this fixed it for this particular user.
I just stopped using whatsapp for video calls as well and no more blackscreens during video calls.
Most my blackscreens are during light loads.
Anyway nvidia killed gpu's cos their fan profiles broke and cards cooked them self with 0 rpm fan, if not read any reports of fans stopped working ever on amd gpu's.

 

[ChristTheGreat]

you sure? do a search for nvidia cards melting after a driver update..............probably about 8-10 years ago??

Talking about this?

"The problem seems to be related to the fan controller, causing the fans to slow down, and even stop. This happening while the card is in use is a very bad thing indeed, and can cause poor performance (as the GPU tries to cool itself down by reducing power) and possibly even overheat your GPU to the point where the card stops working."

WARNING! NVIDIA 196.75 drivers can kill your graphics card

I'm getting a steady stream of reports in from NVIDIA users that the latest 196.75 drivers can cause severe cooling problems, even possibly resulting in the death of the hardware.

www.zdnet.com

 

[mechtech]

Talking about this?

"The problem seems to be related to the fan controller, causing the fans to slow down, and even stop. This happening while the card is in use is a very bad thing indeed, and can cause poor performance (as the GPU tries to cool itself down by reducing power) and possibly even overheat your GPU to the point where the card stops working."

WARNING! NVIDIA 196.75 drivers can kill your graphics card

I'm getting a steady stream of reports in from NVIDIA users that the latest 196.75 drivers can cause severe cooling problems, even possibly resulting in the death of the hardware.

www.zdnet.com

I think thats it, ya wow 2010, I'm getting old

 

[eidairaman1]

Talking about this?

"The problem seems to be related to the fan controller, causing the fans to slow down, and even stop. This happening while the card is in use is a very bad thing indeed, and can cause poor performance (as the GPU tries to cool itself down by reducing power) and possibly even overheat your GPU to the point where the card stops working."

WARNING! NVIDIA 196.75 drivers can kill your graphics card

I'm getting a steady stream of reports in from NVIDIA users that the latest 196.75 drivers can cause severe cooling problems, even possibly resulting in the death of the hardware.

www.zdnet.com

Detonators

 

[TheoneandonlyMrK]

See at least this tries to assist in some way instead of just ha look at this company, well ish.

 

[LLezerus]

GPU repair service reports a sudden spike of broken Radeon RX 6900/6800 GPUs - VideoCardz.com

AMD Radeon RX 6900/6800 GPUs are mysteriously dying and no one knows why German graphics card repair service notices a sudden spike of broken Radeon GPUs. Radeon RX 6000 card with broken GPU, Source: KrisFix KrisFix who is running a professional graphics card repair service, claims that out of...

videocardz.com

Update: KrisFix posted the following photos of the card shown in the video. The card’s GPU has visibly cracked die. Such defects are not related to the driver.

It's miners screwing coolers back on too hard. 'Tis a complete nothingburger.

 

[Bomby569]

It's miners screwing coolers back on too hard. 'Tis a complete nothingburger.

unless they screw it with a flame thrower i don't think it has anything to do with that

 

[LLezerus]

unless they screw it with a flame thrower i don't think it has anything to do with that

Modified mining BIOS with "unsafe" voltage + overtightened thermal solution will get you there 100%

 

[AusWolf]

Modified mining BIOS with "unsafe" voltage + overtightened thermal solution will get you there 100%

Isn't the frame around the GPU die supposed to protect it from that situation?

 

[LLezerus]

Isn't the frame around the GPU die supposed to protect it from that situation?

The copper plate protrudes above this and the GPU die protrudes above the frame. The frame is mainly for getting the correct alignment when attaching the thermal solution.

 

[Vya Domus]

The frame is mainly for getting the correct alignment when attaching the thermal solution.

It also does usually prevent uneven pressure because you can in fact crack the dies if that's done improperly. This was a real problem on HBM GPUs.

 

[Kabouter Plop]

Do we know if the cards krisfix received still had the void warranty sticker on screw cos if its still on there then thats very bad news for AMD

 

[evernessince]

my CPU fan now does something it never did before when gaming. Probably totally unrelated but still unexpected.

If by something it never did before you mean audibly ramp up, that's normal. You have a new GPU and it is using more CPU resources.

 

[A Computer Guy]

If by something it never did before you mean audibly ramp up, that's normal. You have a new GPU and it is using more CPU resources.

Ramping up in config menu? During the game it ramps back down again back to normal almost silent. It just struck me as odd but 6700XT is the beefiest card I have ever owned.
Funny thing is my CPU fan has never really needed to ramp up very high during any gaming and stays pretty quiet even after unzipping something for an hour.

 

[evernessince]

Ramping up in config menu? During the game it ramps back down again back to normal almost silent. It just struck me as odd but 6700XT is the beefiest card I have ever owned.
Funny thing is my CPU fan has never really needed to ramp up very high during any gaming and stays pretty quiet even after unzipping something for an hour.

Might be the game doesn't cap FPS in the menu. In any case you can run HWInfo to find out pretty quickly if it causes your temps to spike in the menu.

 

[eidairaman1]

Do we know if the cards krisfix received still had the void warranty sticker on screw cos if its still on there then thats very bad news for AMD

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