Info AMD confirms Windows 11 slow down its CPUs up to 15%

Page 6 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

deasd

Senior member
Dec 31, 2013
512
724
136



The issue mentioned in the opening paragraph is caused by functional L3 cache latency which has increased by around three times on affected hardware. The programs that will suffer include those sensitive to memory subsystem access times. AMD noted another problem too, explaining that UEFI CPPC2 may not schedule threads on the processor’s fastest core preferentially.


Regarding the latter issue, applications sensitive to the performance of one or a few CPU threads will see a performance hit. The issue will be more noticeable on greater than 8-core processors that operate at over 65W. This issue should also be fixed this month.


hmmmm, whose fault is it this time? It screw up the future hardware review if true? what about Intel side?


AMD.jpg



ndjxev655tr71.png
 

LikeLinus

Lifer
Jul 25, 2001
11,518
670
126
By your own explanations you're pointing that it's NOT AMD's fault.
You yourself said more than one time that the AMD driver (that may be contributing to the problem but may also not be the only culprit) was working as intended on Microsoft's current OS, Windows 10, probably following Microsoft's guidelines. You yourself is saying that AMD's driver is working.
What changed, the driver or the OS?

Yes, the OS.
It's AMD's fault if some new feature that should be for exclusive use by a future rival's product that wasn't even launched yet is causing problems for it's CPUs? Something that AMD have no ways to control, nor say in how that new feature should work?
The most that AMD can do is pay attention to "bugs" on the development builds and ask Microsoft to fix it, and we have evidence that this was done and it was working as it should in recent builds. It's AMD's fault if Microsoft launched W11 on a broken build? That they did this giving zero warnings to the users? Why wasn't the fix available at day 1 if it was already available?
A new driver by AMD is not a fix, it's a workaround.

And isn't concerning that Microsoft rushed a new OS that may have changed rules about how drivers should be made and may cause problems for many others besides AMD?

You are confusing two different things here. There are two separate performance related issues with AMD and Win 11 right now. One is the scheduler (that's on MS, but I'm highly surprised AMD didn't notice this before or after RTM and the actual release) and then there is the UEFI issue that is on AMD for not updating the driver before launch.

No, I'm pointing out the UEFI issue IS AMD's responsibility. It is their driver/software to maintain and did not update it before the release and upon release, issued a KB letting users know they were aware of the issue and working on it.

Let this sink in, why aren't you still using drivers from Windows Vista or past OSs? By nature, things change within the OS during development because code is update, changed and added. This can and does cause functionality issues with old drivers that are based on the code of an older OS. The OS change, but the driver did not change with it! MS isn't responsible for the software/driver.

The UEFI issue has nothing to do with the issue that Microsoft is working on for the scheduler picking the highest performing cores.

No, it's not concerning that Microsoft "rushed" a new OS. This OS has been known about for a long time and AMD had the ability to work on their UEFI issue well before after launch.

Notice on how September 20th, nVidia released finalized version of Windows 11 drivers? Other companies did too, but this is an example.

"he company did release its first Windows 11 ready drivers back in July, but the drivers released today are designed for the final version of Windows 11, which Microsoft plans to release on October 5, 2021."

It's amazing that another company can release finalized drivers several weeks before the actual launch of Win 11. you know, cause Microsoft rushed the OS and everyone was caught by surprise!

Regardless of all of this, the fix is company and this isn't really a huge deal in the grand scheme of things.
 

Abwx

Lifer
Apr 2, 2011
10,847
3,296
136
What evidence? You provided no link with real evidence. All you posted was a Reddit thread with a thousand replies, counting it as the "truth" and you keep dogging the real question. I never said I was a reputable source, That's why I gave you links with comments directly from AMD. You can't even understand that the drive issue is being done via a software update and not a windows update because it's a driver written by AMD. But you state AMD is aware of it and was aware of it before the release? Even saying they told Microsoft and MS ignored it? LOL. Delusional with no proof.

Feel free to ignore me. :beermug:

Any software vendor should check that his product work correctly with already supported CPUs, so it is on MS duty to make sure that their new OS work at least as well as the previous one with AMD CPUS, rest is useless argumentation to shift the blame on the victim.

Indeed W10 scheduler did work efficently, it would have been enough to keep it in use in W11 once AMD CPUs are detected.

Anyway you look at it that s very unprofessional from MS part.
 

Borealis7

Platinum Member
Oct 19, 2006
2,914
205
106
oh i like this game! (first 4 are my own opinions)

Win95: BSOD
Win NT 4: 👍
Win98: 👎
Win98 SE: 👍👍
2kSP4: 👍👍👍
Millennium: 👎👎👎
XP (RTM): 👎
XP (SP2): 👍
XP (SP3): 👍👍
Vista (RTM):👎
Vista (SP2): 👍
7: 👍
8: 👎
8.1: 👍
10: 👍
11: 👎👎
 
  • Like
Reactions: Tlh97 and coercitiv

LikeLinus

Lifer
Jul 25, 2001
11,518
670
126
Any software vendor should check that his product work correctly with already supported CPUs, so it is on MS duty to make sure that their new OS work at least as well as the previous one with AMD CPUS, rest is useless argumentation to shift the blame on the victim.

Indeed W10 scheduler did work efficently, it would have been enough to keep it in use in W11 once AMD CPUs are detected.

Anyway you look at it that s very unprofessional from MS part.
As I've stated several times, it's on both AMD and Microsoft. You're just cherry picking what you want to hear because it fits your agenda. It's also very unprofessional for AMD to not update their driver/software previous to the launch of the new MS. They are the developers of said driver. Note, I even provided a link how nVidia and other vendors had driver updates well before the launch. But the UEFI issue is on AMD, not Microsoft. The Scheduler is on MS.
 

Abwx

Lifer
Apr 2, 2011
10,847
3,296
136
As I've stated several times, it's on both AMD and Microsoft. You're just cherry picking what you want to hear because it fits your agenda. It's also very unprofessional for AMD to not update their driver/software previous to the launch of the new MS. They are the developers of said driver. Note, I even provided a link how nVidia and other vendors had driver updates well before the launch. But the UEFI issue is on AMD, not Microsoft. The Scheduler is on MS.

So much for advancing an agenda.

UEFI is not designed hazardously but according to Windows requirements.

What happened here is that MS modded profundly the scheduler to correctly work with an heterogenous cores design, namely ADL, they just didnt do extensive tests with other CPUs, as such a scheduler misbeahavour should have been detected very early during conception if they had payed one guy or two to do a few benchmarks.
 

jpiniero

Lifer
Oct 1, 2010
14,509
5,159
136
What happened here is that MS modded profundly the scheduler to correctly work with an heterogenous cores design, namely ADL, they just didnt do extensive tests with other CPUs, as such a scheduler misbeahavour should have been detected very early during conception if they had payed one guy or two to do a few benchmarks.

If it was really a scheduler thing, pre Alder Lake Intel CPUs would have been affected too. I haven't seen anything suggesting that.
 

Zucker2k

Golden Member
Feb 15, 2006
1,810
1,159
136
As I've stated several times, it's on both AMD and Microsoft. You're just cherry picking what you want to hear because it fits your agenda. It's also very unprofessional for AMD to not update their driver/software previous to the launch of the new MS. They are the developers of said driver. Note, I even provided a link how nVidia and other vendors had driver updates well before the launch. But the UEFI issue is on AMD, not Microsoft. The Scheduler is on MS.
Even the scheduler is on AMD as well. W11 is a different OS under the hood, and I don't think MS would release it without extensive performance assessments with it two x86 partners. Like I said before, the fact that the fix is supposedly in the channel means that the necessary code from AMD arrived late. Things like scheduling and such are buried in microcode (AGESA?), only AMD has access to this proprietary data and the necessary changes that need to be made to ensure smooth communication between OS and CPU.
 

LikeLinus

Lifer
Jul 25, 2001
11,518
670
126
So much for advancing an agenda.

UEFI is not designed hazardously but according to Windows requirements.

What happened here is that MS modded profundly the scheduler to correctly work with an heterogenous cores design, namely ADL, they just didnt do extensive tests with other CPUs, as such a scheduler misbeahavour should have been detected very early during conception if they had payed one guy or two to do a few benchmarks.

It's not designed to Windows requirements.

The Unified Extensible Firmware Interface Forum or UEFI Forum is an alliance between several leading technology companies to modernize the booting process. The board of directors includes representatives from twelve "Promoter" companies: AMD, American Megatrends, ARM, Apple, Dell, Hewlett Packard Enterprise, HP Inc., Insyde Software, Intel, Lenovo, Microsoft, and Phoenix Technologies.

AMD is updating their UEFI driver to work correctly with Win 11. This should have done pre-RTM as other companies have done.

Again, for the millionth time:

"AMD’s UEFI CPPC2 driver may not preferentially schedule threads on a processor’s fastest core.” CPPC2 stands for Collaborative Power and Performance,” and its part of how the CPU communicates with the operating system. We know AMD uses CPPC2 to communicate with the OS regarding which CPU cores are preferred for boosting in both Windows 10 and Windows 11. It seems the version of this software that runs under Windows 11 isn’t working quite right.

It's AMD's software at fault because it wasn't updated to work correctly with Windows 11. ALL vendors have to check their drivers for a new OS launch. Blaming MS for AMD not fixing this or announcing it after the launch is silly.
 
Last edited:

Kedas

Senior member
Dec 6, 2018
355
339
136
A conversation like this:

AMD: What is going on in Win11?
MS: in WIn11 green is red and red is green. :)
AMD: uhhm wait what? o_O
MS: it's in the updated documents page 13 table 2 row 16 (copyright intel) :cool:
AMD: uhhm ok, didn't notice that, I guess we are going to fix our drivers... :rolleyes:
 
Last edited:

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
ALL vendors have to check their drivers for a new OS launch.

That's fantasy-land level propaganda. Every time MS updates their OS, there's either an update to the driver model or there isn't. In the case of Win11, there's no announced update to the driver model. You're lucky to get hardware vendors to update drivers in the event of a change in driver model. But if there isn't one?

How many other drivers are broken under Win11? It's possible that there are others, but I haven't had any issues on my end and I certainly haven't heard of any elsewhere.
 
  • Like
Reactions: Tlh97 and Thibsie

LikeLinus

Lifer
Jul 25, 2001
11,518
670
126
That's fantasy-land level propaganda. Every time MS updates their OS, there's either an update to the driver model or there isn't. In the case of Win11, there's no announced update to the driver model. You're lucky to get hardware vendors to update drivers in the event of a change in driver model. But if there isn't one?

How many other drivers are broken under Win11? It's possible that there are others, but I haven't had any issues on my end and I certainly haven't heard of any elsewhere.
I should have said "All vendors should...". I realize it doesn't always happen, but the major vendors generally test and update drivers, if even needed.

I showed the link to the Intel Network driver issue. nVidia developed new finalized Win11 drivers. AMD themselves updated other drivers.


Win 11 has only been live less than a week and it's also limited on all of the hardware. You'll start hearing about more issues as time moves on. But that doesn't mean a company shouldn't test something as important as UEFI drivers. This isn't the normal "oh my driver doesn't work and so I can't use my printer!". This is a software driver that AMD has stated is causing performance loss issues. How that got through all the testing, is beyond me.

This isn't a regular Windows update. This is a new OS launch. So it's a bit different.
 

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
AMD themselves updated other drivers.

Do we have any evidence that AMD didn't update their CPPC2 driver? And/or didn't do any testing whatsoever? Other than the fact that (currently) it doesn't work correctly?

Also while AMD may have updated their chipset drivers, I didn't update mine in particular and either a). I'm using drivers provided to me by MS or b). the old ones I do have installed work just fine.

I do know that I can use old AMD dGPU drivers without a hitch under Win11. And I have, too.
 

Zucker2k

Golden Member
Feb 15, 2006
1,810
1,159
136
Do we have any evidence that AMD didn't update their CPPC2 driver? And/or didn't do any testing whatsoever? Other than the fact that (currently) it doesn't work correctly?

Also while AMD may have updated their chipset drivers, I didn't update mine in particular and either a). I'm using drivers provided to me by MS or b). the old ones I do have installed work just fine.

I do know that I can use old AMD dGPU drivers without a hitch under Win11. And I have, too.
And when your system crashes you'll blame MS for releasing an OS prematurely. Hehe
 

LikeLinus

Lifer
Jul 25, 2001
11,518
670
126
Do we have any evidence that AMD didn't update their CPPC2 driver? And/or didn't do any testing whatsoever? Other than the fact that (currently) it doesn't work correctly?

Also while AMD may have updated their chipset drivers, I didn't update mine in particular and either a). I'm using drivers provided to me by MS or b). the old ones I do have installed work just fine.

I do know that I can use old AMD dGPU drivers without a hitch under Win11. And I have, too.
You know what. You're right!

1. AMD did update their CPPC2 drivers, unfortunately, they did a horrible job doing so and needed to release a KB telling customers their systems performance will be impacted until they decide to take another attempt at fixing said issue.

2. AMD thoroughly tested the system and was unable to correct the issue with the CPPC2 drivers and it just doesn't currently work, but they know they can fix it in the next 2 weeks!

3. They decided not to tell the consumer until other people reported the issue after the launch on Oct. 5th. It's like a happy surprise!

Good job, AMD! :rolleyes: Sigh, it's like trying to point out the obvious to the oblivious.

AMD did nothing wrong. Happy now? LOL.
 

Justinus

Diamond Member
Oct 10, 2005
3,166
1,508
136
You know what. You're right!

1. AMD did update their CPPC2 drivers, unfortunately, they did a horrible job doing so and needed to release a KB telling customers their systems performance will be impacted until they decide to take another attempt at fixing said issue.

2. AMD thoroughly tested the system and was unable to correct the issue with the CPPC2 drivers and it just doesn't currently work, but they know they can fix it in the next 2 weeks!

3. They decided not to tell the consumer until other people reported the issue after the launch on Oct. 5th. It's like a happy surprise!

Good job, AMD! :rolleyes: Sigh, it's like trying to point out the obvious to the oblivious.

AMD did nothing wrong. Happy now? LOL.

I have no idea why someone would white knight so hard for microsoft and intel.
 

LikeLinus

Lifer
Jul 25, 2001
11,518
670
126
I have no idea why someone would white knight so hard for microsoft and intel.
Uh, this has nothing to do with Intel. This also isn't really about MS. It's pointing out that both parties are responsible. Why is it so hard for people to understand and try to make every excuse to not accept the truth? They are just white knighting for AMD. There are two sides to ever coin.
 

Justinus

Diamond Member
Oct 10, 2005
3,166
1,508
136
Uh, this has nothing to do with Intel. This also isn't really about MS. It's pointing out that both parties are responsible. Why is it so hard for people to understand and try to make every excuse to not accept the truth? They are just white knighting for AMD. There are two sides to ever coin.

You're worse than some of the trolls in P&N. Bai.
 

dacostafilipe

Senior member
Oct 10, 2013
771
244
116
IMO, it looks like AMD found a bug in the new OS and can't fix it via driver alone.

Microsoft will certainly release a Windows update and AMD will follow with a driver to handle the performance issue.

As to why is was release like this, it's certainly a regression that happened just before code freeze and because it's not a critical issue Microsoft choose to fix it post-release, specially when the build was already approved by multiple other vendors. Something that happens all the time in the software business.

It's not like this would fry your CPU, right? ;)
 

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
And when your system crashes you'll blame MS for releasing an OS prematurely. Hehe

. . . no? Only crashes I've had are from buggy applications that were also crashing under Win10.

You know what. You're right!

1. AMD did update their CPPC2 drivers, unfortunately, they did a horrible job doing so and needed to release a KB telling customers their systems performance will be impacted until they decide to take another attempt at fixing said issue.

2. AMD thoroughly tested the system and was unable to correct the issue with the CPPC2 drivers and it just doesn't currently work, but they know they can fix it in the next 2 weeks!

3. They decided not to tell the consumer until other people reported the issue after the launch on Oct. 5th. It's like a happy surprise!

Good job, AMD! :rolleyes: Sigh, it's like trying to point out the obvious to the oblivious.

AMD did nothing wrong. Happy now? LOL.

Or maybe the point of my post was

You don't know.

You have no idea what caused the old driver to not work, or to prompt AMD to release the kb, or what the state of the fix is or how long they've been working on it. I do know that MS advertised no change in driver model and that many old hardware drivers still work in Win11. So something MS did specifically broke the CPPC2 driver, at least to the extent that AMD felt the need to mention it.

According to my current test run in CBR20 ST, the system is doing a pretty good job of sticking to Core 3 which is the "good" core on CCD #1 of my 3900X. It switches between Core #1 and Core #3, and Core #1 loses 10-30 MHz (~4470 vs 4480-4500 MHz). Granted I have no idea what version of the CPPC2 driver I'm running. Could be there's an internal test build of the driver that ships with the dev channel builds.

(fyi my ST score under 22471.1000 is 518, which is pretty normal)

I also tested SuperPi Mod 1.5 XS (old skewl) and got good results. Again it mostly ping-ponged between Core #1 and Core #3. I got 9.098s in 1M which is pretty normal for a stock 3900X. The reference score I got was 9.129s.

Seems like much ado about nothing, or maybe AMD already has it fixed?
 

Grooveriding

Diamond Member
Dec 25, 2008
9,107
1,260
126
It's fixed in the newest version of Windows 11 on the dev channel. So, whenever that comes to release preview, beta and the main channel, it will be available to everyone.

The scheduler fix will have to come in a new chipset driver from what I've read. Sometime later this month.

Sad they released this without it fixed.
 

LikeLinus

Lifer
Jul 25, 2001
11,518
670
126
Or maybe the point of my post was

You don't know.

Seems like much ado about nothing, or maybe AMD already has it fixed?
Maybe if you actually followed the conversation and read everything wrote and not just one of the post, you would have noticed I've said multiple times ""No one really knows how things transpired."

But people want to keep only blaming MS and even AMD has acknowledge their driver needs to be updated. That's all, nothing more, nothing less. The blind loyalty is just weird around here. It is a lot like P&N in that way.

I've also said the same thing about the fix. This will be fixed in no time and I bet no one would have noticed if it weren't for benchmark tools. For most software, it's a 3-5% loss for the scheduler issue. That's nothing. I'm not sure what the loss is for UEFI. Don't think I've seen anyone tell that.
 
Last edited:

dr1337

Senior member
May 25, 2020
309
503
106
But people want to keep only blaming MS
I use windows because of legacy support. If microsoft is going to start changing software so much that it breaks hardware support and ruins the stabilty/consistency I've come to expect then why should I continue to use windows over forcing linux to run the games I play? In my eyes windows 8 and windows 10 are already steps away from the core reasons I actually use the OS. If the developer of the OS is gonna start breaking parts of the code so much that it causes random hardware problems, it is very concerning to me what the future of windows stability will look like. What other drivers and hardware functions will also be crippled, especially in more obscure instances? we don't know until people find out and report bugs. Microsoft either didn't test w11 on AMD hardware, didn't notice the bugs, or just didn't care. This really doesn't sell me on upgrading from w10 tbh.

I do believe this issue is more in M$'s court than AMDs. I think the implicit writing on the wall especially with the seeming intel favoritism doesn't bode well when their excuse is supposedly incompetence.
 
  • Like
Reactions: Tlh97 and Ranulf