• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

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

This is worth a watch


Who is Kris, he's clearly a pro and earning off repairs and isn't a prolific tuber, he looks legit to me, all odd though still.
I think he is legit I just don' trust the source of the cards being a regular user(s).
 
Hang on a minute.

Wtaf are these people on anyway.

Why are they being taken to a repair shop.

They had warranties.

In normal use that warranty would still have been valid.

This is total rubbish if my GPU popped while at idle or gaming I wouldn't be taking a new in warranty GPU to a repair shop.

A+B doesn't equal fu£#@£ P on this Story.
Common sense is not common on the interwebs, is what this is. I see a pattern :)
yep... i kinda want to call F.U.D. on that story ... it's localized.

but

although, the latest up to date driver for me is 22.11.2 for me which is as i mentioned, 30.11.2022 for me ... tho i suspect AMD did react and might have pulled off subsequent drivers releases ... these report and issue seems fishy ...

nonetheless i would have to point, my RX 6700 XT was my best GPU upgrade i ever had since my XpertVision X1950 GT Super AGP from a GeForce 7950 GT AGP (which where my last AGP cards i ever owned, only kept the XVi card, sadly ... given how rare are the 7950 GT AGP nowadays )

This is worth a watch


Who is Kris, he's clearly a pro and earning off repairs and isn't a prolific tuber, he looks legit to me, all odd though still.
not wrong either, i do enjoy watching Kris video nonetheless ... but still something doesn't add up ... dunno why ...
 
Didn't 6800XT/6900XT/6950XT have some crazy transient spikes going on?

In order to protect the GPU I'm pretty sure there needs to be actual fuses implemented on the PCB, GPU won't just save itself. So with a wonky driver......I need to go look at some 6900XT PCBs again. Just fun speculation

Yes usually you have to go to MPT to run unsafe profiles, but don't driver updates overwrite SPPT anyway?

View attachment 278633

These spikes are nasty! o_O
Something was wrong with Navi 21 and at AMD. You see how these are more or less fixed with the newer RX 7900 series.

1673469520325.png

AMD Radeon RX 6950 XT Reference Design Review - Power Consumption | TechPowerUp

1673469403856.png
 
AFAIK, AMD cards don't have the same level of hardware protection built into them as NVIDIA's do. So it is possible, though unlikely, that a driver bug (or just pushing the power limits a bit too high to eke out some extra perf) is causing some 6000-series cards to go bang. But no point in becoming overly concerned until a more reputable source like GN jumps on this - that said, if I was a 6000-series owner I would downgrade my drivers for the time being.

The fact that we've had 3090s / 3080s being bricked by New World says otherwise.

Mind you bricking a card is one thing, no idea how you get die cracking from a bad driver.

In any case this is highly unusual and should be verified before anyone jumps to conclusions. Input from Debauer or Actual Hardcore Overclocking would be very helpful.

Is there a way to prevent the spiking though configuration? I typically don't bother to OC my GPU just leave the defaults, maybe change the fan curve.

You can reduce the spiking by limiting the power draw. For AMD cards you can do so in the driver. Power spikes start to get bad for last gen AMD and Nvidia cards at above 300w. Both the AMD 7000 series and Nvidia 4000 series fixed the power spike issues.
 
The fact that we've had 3090s / 3080s being bricked by New World says otherwise.

Mind you bricking a card is one thing, no idea how you get die cracking from a bad driver.

In any case this is highly unusual and should be verified before anyone jumps to conclusions. Input from Debauer or Actual Hardcore Overclocking would be very helpful.



You can reduce the spiking by limiting the power draw. For AMD cards you can do so in the driver. Power spikes start to get bad for last gen AMD and Nvidia cards at above 300w. Both the AMD 7000 series and Nvidia 4000 series fixed the power spike issues.
For AMD is that under Power Tuning: Power Limit (%) ?
 
Hardware limits in AMD's GPU firmware exist for years now and until this recent batch of GPUs no one has showed that kind of failure. Mining under terrible thermal conditions with custom BIOS might result in that, not any casual use or even benchmarking.
 
I'd like to know if anyone involved has any position in NVDA, AMD or stands to benefit directly in any other way...

That's what this smells like to me...
 
Just because the driver is the common factor, it doesn't mean there wasn't something else, too (overclocking, mining, FurMark?).

I'll keep an eye out and monitor my 6750 XT, just in case.
 
AMD staying silent right now does not help, all we can do is speculate which can potentially further damage AMD reputation.
If been having driver issues for 6 months my card is watercooled.
My card ignores frequency limits already confirmed firing up doom eternal on stock max clocks 2555 mhz
Firing up on custom default tuning preset says 2479 mhz leaving un changed applying also boost 2555 mhz.
If i lower tuning preset by 1 mhz lets say 2478 mhz and fire up doom eternal it boost to 2468 mhz so only default frequency limit is broken.
Imagine thermal limits also being broken.
AMD making gpu's cook it self to death ?
AMD launching new gpu with hotspot issues ?
Random user reporting display port cable makes ssd not be detected but changing it magically fixes hotspot issue and ssd re appears ?
20th pin issues ?
I have a club 3D display port 1.4 HBR3 cable btw if i am correct this is one of the better cables.
Wish my gpu just died randomly but it can run heaven for 24 hours just fine and superposition and port royal each for 8 hours.

Am i to offtopic ? if so sorry mods but i just want to rant right now im suffering driver issues even today, things are more stable on 22.11.2 already mailed this YouTuber as well if been on 22.5.1 least for 4 months.
Im not trusting AMD one bit, anything that comes at them is their own fault for lack of communication, and leaving everything to users them self to fix.

Also Nvidia has had drivers kill gpu's before in past as well i believe it with with gtx 480 and 1 pulled driver with gtx 1080.
 
Didn't 6800XT/6900XT/6950XT have some crazy transient spikes going on?

In order to protect the GPU I'm pretty sure there needs to be actual fuses implemented on the PCB, GPU won't just save itself. So with a wonky driver......I need to go look at some 6900XT PCBs again. Just fun speculation

Yes usually you have to go to MPT to run unsafe profiles, but don't driver updates overwrite SPPT anyway?

View attachment 278633
Yes they do ( might not be, doing more test due to new findings.)

 
Last edited:
Phew, I have the older ones. (22.11.2), though he mentioned only 6800/6800 XT/6900 XT and not 6700 XT.
 
Phew, I have the older ones. (22.11.2), though he mentioned only 6800/6800 XT/6900 XT and not 6700 XT.
Isn't 22.11.2 the newest one for 6000 series?
 
Isn't 22.11.2 the newest one for 6000 series?
Yes 22.12.1 wont install but if read it installed for some other users which is strange.
 
Isn't 22.11.2 the newest one for 6000 series?
There is 22.12.2 but only for 7900 Series....'tho it's seems that this could be potential issue as some people claim that was possible to install this driver on 6000 series of cards.....here video bellow explains this

 
Yes 22.12.1 wont install but if read it installed for some other users which is strange.
Maybe that's why their cards failed? If the driver is not designed for X card, but misrecognises it to be Y card and installs anyway, that could lead to a million problems.

Edit: Chris (or whatever his name is) said the common factor is the newest driver. He didn't say which one - 22.11.2 or 22.12.2?

There is 22.12.2 but only for 7900 Series....'tho it's seems that this could be potential issue as some people claim that was possible to install this driver on 6000 series of cards.....here video bellow explains this

I'll watch it after work, thanks. :)
 
Maybe that's why their cards failed? If the driver is not designed for X card, but misrecognises it to be Y card and installs anyway, that could lead to a million problems.

Possibly but that would be still AMD fault cos they should have failsafes in place to prevent that, i honestly do not think its from installing those drivers there have been driver issues for 6 months there plenty of users that will report the same, maybe not here and with AMD small market share most of them are probably posting on reddit.
 
Possibly but that would be still AMD fault cos they should have failsafes in place to prevent that, i honestly do not think its from installing those drivers there have been driver issues for 6 months there plenty of users that will report the same, maybe not here and with AMD small market share most of them are probably posting on reddit.
I don't know. I've had my 6750 XT for about 2-3 months now, and no issues so far. Though the video also said it's about 6800 and 6900 cards only, so I don't know. I'm on 22.11.2 by the way.
 
I don't know. I've had my 6750 XT for about 2-3 months now, and no issues so far. Though the video also said it's about 6800 and 6900 cards only, so I don't know. I'm on 22.11.2 by the way.

If seen issues being reported even on older cards heck some of the newer 6000 series cards i believe cant install 22.5.1 so they have driver issues and cannot rollback either, i forgot which cards i believe one of the new 6x50 cards, anyway now you can atleast disable MPO and avoid most problems, just games wont run stable with it.
 
Possibly but that would be still AMD fault cos they should have failsafes in place to prevent that, i honestly do not think its from installing those drivers there have been driver issues for 6 months there plenty of users that will report the same, maybe not here and with AMD small market share most of them are probably posting on reddit.
How could you blame AMD if some does that by getting around the failsafes? By default you cannot install the Driver for the 7000 series and if you use the 6000 series driver with a 7000 be prepared for black screens and other issues.
 
There is 22.12.2 but only for 7900 Series....'tho it's seems that this could be potential issue as some people claim that was possible to install this driver on 6000 series of cards.....here video bellow explains this

I've just watched this (I'm at work, but never mind).

He explains what he thinks quite logically, though it's still just his opinion. Without Chris confirming which driver version he meant by "December 8 update", we can't know for sure
 
I use RX 6900 XT and don't even see a (actual) December driver that comes up as an option. In fact AMD Radeon Software suggests the latest my GPU could use is 22.11.2 which was released (or maybe technically "built") on Dec 8th......makes total sense......released in December with a November name scheme using 11 as the 2nd digit. I suppose it was a beta to WHQL version or else I wouldn't know why they would name stuff like that.

This also aligns with how things are laid out on their website

I don't update unless I need to or have a legit want like a bug is fixed. I'm still on 22.11.1 for this reason

1673476061697.png


1673476992134.png


but if you wanna be "that guy" and download drivers not meant for your GPU, or better yet, specific to another GPU. but with all that said.....I thought drivers do a hardware check at start of installation to even see if you have hardware installed that supports them? if nothing found, install terminates.....I don't get the hub bub around any of this.
1673477039564.png
 
Back
Top