ThunderslatMS update causes problems with AMD drivers. Clearly it's AMD's fault. Can't make this shit up.

Arent AMD drivers just bad as is? Like a simple update in Windows just breaks it. Or anything really from what I can find.

Clippy doesn't like all this ridicule and shade. Clippy works hard to bring you an edgy experience. We are going to have to help Clippy so he can fulfill his lifelong dream of sleep.

Knight47Arent AMD drivers just bad as is? Like a simple update in Windows just breaks it. Or anything really from what I can find.

If you want to go back two decades then maybe. It's just one of those things that is repeated so many times that it must be "true".

TesselatorI wonder if anyone here is or will ever be, affected by any of these? I know I'm not. No one in my groups are. I guess it's just something people like to post forum messages about.

I also know that depending what information feeds you're on, these sorts of announcements haven't changed in 20 years. And really, for every even slightly complex, actively developed, software - not even just OS's.

These issues don't happen to you, therefore, they don't exist. TPU editors are making stuff up for extra clicks. Sounds legit.

Just like AMD messing up Windows update.

Seriously guys, get a grip.

halcyonIt's as if they are doing this on purpose. Nobody is this incompetent.

I mean... well, umm...... man, look....

Dr. DroThe budget was originally 500, but Tim Hortons coffee is more important

See my location. That's never true. (/s, I'm not a coffee drinker but even I've heard it's decline, enjoy it!)

Knight47Arent AMD drivers just bad as is? Like a simple update in Windows just breaks it. Or anything really from what I can find.

No, they're not. No, they don't. No, they do not.

That's FUD. Please stop spreading it.

in this AI era, microsoft cant really make things work without creating bunch other issue ?

OR they use to much AI to create/fix their code ?

FierceRedSee my location. That's never true. (/s, I'm not a coffee drinker but even I've heard it's decline, enjoy it!)

Hahaha, this was more of an allusion to RTG/AMD's GPU IP division being based out of Markham (old ATI HQ).

LabRat 891It's a surprisingly common sentiment-

https://twitter.com/i/web/status/2098112740552266219

:eek: Things must be real bad if the owner of that account is actually saying "screw AMD" to any capacity. :fear:

TesselatorI wonder if anyone here is or will ever be, affected by any of these? I know I'm not. No one in my groups are. I guess it's just something people like to post forum messages about.

I also know that depending what information feeds you're on, these sorts of announcements haven't changed in 20 years. And really, for every even slightly complex, actively developed, software - not even just OS's.

i was part of the group that was having the driver timeouts last year. it was an update that ended in 3011 that was the cause.

been pretty smooth sailing this year.

but yeah time is a circle bro.

Honestly just more free advertising for Linux.

LabRat 891It's a surprisingly common sentiment-

https://twitter.com/i/web/status/2098112740552266219

So if you want to detect issues but you have to basically consent to giving your data to AMD!!!! This s**t is messed up. 😠

AusWolfAre you suggesting that AMD messed up the latest Windows update?

Well they still havent figured it out how to separate igpu and dgpu drivers causing windows to replace it.

FierceRedI mean... well, umm...... man, look....



See my location. That's never true. (/s, I'm not a coffee drinker but even I've heard it's decline, enjoy it!)



No, they're not. No, they don't. No, they do not.

That's FUD. Please stop spreading it.

But if I go to forums, I see people having issues. Now the solution to fix stuff is to replace everything aside from the gpu itself or the driver.

Like I bought(I dont) a 9070 and started having USB issues, TDR or whatnot. When I put back my old nvidia card it just works. Fix is: my ram is broken, mobo bios outdated, psu issues, registry hack, anything but the gpu, replace/adjust whole PC around the gpu. That's a pretty aggressive leap. Not sure if those are serious or just trolls.

kazuvikingWell they still havent figured it out how to separate igpu and dgpu drivers causing windows to replace it.

Causing Windows? It's not causing Linux or any other OS? Hmm, weird. Totally AMD's fault like everything wrong with the world is. :rolleyes:

AusWolfCausing Windows? It's not causing Linux or any other OS? Hmm, weird. Totally AMD's fault like everything wrong with the world is. :rolleyes:

Different driver system, different code completely and Linux with RADV and AMDGPU doesn't have to worry usually about running two different instances between three different branches because AMD doesn't update older architectures which are used by half or more than half of their current integrated graphic products.

AleksandarKthe focus on security issues has resulted in Microsoft introducing many new bugs into the operating system

In my opinion, this is how it should have been phrased:

The focus on AI code generation has resulted in AI introducing many new bugs into the operating system.

Thanks for your attention.

ymdhisSo Windows 10 no longer getting patches is, in fact, a good thing.

Sometimes your hunch is right, but often you’re completely off the mark:
KB5129236 Windows 10 22H2 Out-of-Band Update

Regards,
Fred.

AusWolfCausing Windows? It's not causing Linux or any other OS? Hmm, weird.

Why even bring up linux when the drivers are completely different?

LabRat 891It's a surprisingly common sentiment-

https://twitter.com/i/web/status/2098112740552266219

I believe AMD removed this from the FAQ because it doesn't work anymore, but I the old KB article still exists. Is it a long-term solution for anyone here?

www.amd.com/en/resources/support-articles/faqs/gpu-kb205.html

Knight47But if I go to forums, I see people having issues.

This is true of everything, everywhere, ever. Let's say I grant you that you're right and it's true - it isn't even close to new.

Just like everywhere there's a car in a garage getting work done, doesn't mean cars don't go through long periods of time without issues whatsoever.

Damn, I made a car analogy too early. I have lost this conversation and must commit seppuku. :) /s

Knight47Now the solution to fix stuff is to replace everything aside from the gpu itself or the driver.

This is very standard troubleshooting procedure and you know it.

For decades (1987~2017) we went through large slot and architecture changes every other generation from both Intel and AMD, which means in order to troubleshoot if one device was malfunctioning, you had to either have a 2nd computer of the same era to Trust But Verify the user's assumption, or to swap out a piece of it to accomplish the same in a targeted way.

The number of people who had GPU issues until they changed their monitor which tangentially changed their DVI->HDMI or HDMI->DP is greater than 0.

Knight47Like I bought(I dont) a 9070 and started having USB issues, TDR or whatnot. When I put back my old nvidia card it just works.

Just like people - Hardware Unboxed noticed this as well - will build a new PC with Nvidia 5080+ series and have monitor detection problems on startup and on Windows boot.

Then they switch to an AMD 9070 XT or whatever and it would auto-detect the monitors the first time, remember the refresh rate independent of Windows, and work every single time afterwards. Same mobo, same cables. Different stability results.

The sword cuts both way.

Knight47Fix is: my ram is broken, mobo bios outdated, psu issues, registry hack, anything but the gpu, replace/adjust whole PC around the gpu.

Ever since the GTX 780 (though I would argue since 3dfx and the Glide API) the GPU easily became the most complicated piece of hardware people could own, as well as the cause of most of it's issues. That's 50.1% or more.

Which means in order for someone who doesn't have money for a 2nd PC to fix it, they had to focus on changing something they already owned to assist.

And lets not denigrate the registry hack or the BIOs update. Tweaking the registry hive is a time honored tradition for getting a PC deshitified, and the sheer number of people who never update their BIOs since purchase is alarming.

Knight47That's a pretty aggressive leap. Not sure if those are serious or just trolls.

Yet a necessary one due to the above. They're serious, as I remember those times.

And those times are still happening, unless I can borrow your spare 5090? Please?


No snark intended in these replies, only open conversation.

kazuvikingWhy even bring up linux when the drivers are completely different?

The point is, I find it really dumb to say that it's AMD that messed up the latest Windows update. Microsoft seems to be pretty active at that on their own.

AusWolfThe point is, I find it really dumb to say that it's AMD that messed up the latest Windows update. Microsoft seems to be pretty active at that on their own.

It really depends on the context.
My bro has a 6900 XT, he was on 26.8.1 and has been playing Granblue Fantasy: Relink again since late August/early September and had applied all the Windows updates until September 11th, and it would always blackscreen and cut audio 15-30 minutes in. The only logged error which came out of a few dozen crashes was a single occurence of memory access violation. We sorted out overheating and RAM related issues, as even undervolting, underclocking, temperature limits and base JEDEC speed crashed the PC on that one version of AMD drivers.
I told him to revert to 25.12.1 last week, and it hasn't crashed once since that moment. Arguably, 26.8.1 and possibly 26.9.1 broke something severe enough to reliably crash the PC when the GPU was under load. Eventually, I should check if he did update Windows again in the meantime, but so far it still hasn't crashed to this day.
To me, this driver rollback working tells more about AMD drivers than Windows itself.

RinSenpaiIt really depends on the context.
My bro has a 6900 XT, he was on 26.8.1 and has been playing Granblue Fantasy: Relink again since late August/early September and had applied all the Windows updates until September 11th, and it would always blackscreen and cut audio 15-30 minutes in. The only logged error which came out of a few dozen crashes was a single occurence of memory access violation. We sorted out overheating and RAM related issues, as even undervolting, underclocking, temperature limits and base JEDEC speed crashed the PC on that one version of AMD drivers.
I told him to revert to 25.12.1 last week, and it hasn't crashed once since that moment. Arguably, 26.8.1 and possibly 26.9.1 broke something severe enough to reliably crash the PC when the GPU was under load. Eventually, I should check if he did update Windows again in the meantime, but so far it still hasn't crashed to this day.
To me, this driver rollback working tells more about AMD drivers than Windows itself.

That's a driver issue. I'm not talking about that. When something gets messed up after a Windows update, that's a Windows update issue.

AusWolfThe point is, I find it really dumb to say that it's AMD that messed up the latest Windows update. Microsoft seems to be pretty active at that on their own.

Windows always replaced the amd drivers until very recenty because it thought it was an igpu driver.

kazuvikingWindows always replaced the amd drivers until very recenty because it thought it was an igpu driver.

Yes, because Windows update is dumb. It also replaced my Intel iGPU driver with an older one every time I updated it.

kazuvikingWindows always replaced the amd drivers until very recenty because it thought it was an igpu driver.

To be fair, that's in part AMD's fault because of the mess their driver packages for a "same version" is, you have 3 different packages and only one is a combined one that users normally won't be downloading as AMD usually recommends the individual one because the combined one is around double the size by having to bundle both drivers branches files.

If for example one were to have a recent desktop Zen 4 or Zen 5 with the 2 CU RDNA 2 iGPU with a RDNA 3 or RDNA 4 dGPU and keep the integrated graphics enabled for any reason, the driver AMD recommends to download for RDNA 3 and RDNA 4 is the -b package that only has the latest branch for these with nothing for previous architectures, so one will end with the integrated graphics having no installed drivers.

If you were to also want to add Polaris and Vega to the count, then you would end with no 3 but instead 4 different usual "main" Adrenalin packages that are different and not necessarily compatible between them anymore.

RulumiTo be fair, that's in part AMD's fault because of the mess their driver packages for a "same version" is, you have 3 different packages and only one is a combined one that users normally won't be downloading as AMD usually recommends the individual one because the combined one is around double the size by having to bundle both drivers branches files.

If for example one were to have a recent desktop Zen 4 or Zen 5 with the 2 CU RDNA 2 iGPU with a RDNA 3 or RDNA 4 dGPU and keep the integrated graphics enabled for any reason, the driver AMD recommends to download for RDNA 3 and RDNA 4 is the -b package that only has the latest branch for these with nothing for previous architectures, so one will end with the integrated graphics having no installed drivers.

If you were to also want to add Polaris and Vega to the count, then you would end with no 3 but instead 4 different usual "main" Adrenalin packages that are different and not necessarily compatible between them anymore.

I'm not saying that it's not a problem, but that's not why Windows Update is bad. It's bad because it works with its own driver library. If it finds a discrepancy, it'll install its own version over it.

Like I said, even updating my Intel iGPU driver caused Windows Update to automatically roll back an older version from its own repository.
The only way to keep my drivers up to date was to disable driver updates in Windows Update's registry settings, as ironic as it sounds.