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

Page 4 - 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
723
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
 

BorisTheBlade82

Senior member
May 1, 2020
660
1,003
106
@jpiniero
In case you did not know: Windows is closed source. That basically means that Microsoft and only Microsoft alone decides what to include into any RTM, Developer exclusive or anything else

On an anectodical note: Back in the 80s Intel proposed to them implementations of Excel functions that were up to 15 times faster than their own. You know what MS said? People buy Excel because of the features.
 

LikeLinus

Lifer
Jul 25, 2001
11,518
670
126
Again: AMD very probably was under the impression that the fix was included in the RTM! That it was not was Microsofts decision. They could have simply included it or otherwise delayed the release. They did not. And for good reason: Win11 is vital for ADL and new hardware sells a new OS.

Do you have an ounce of proof of any statement? You're just making things up to blame MS and not hold AMD accountable at all. As I said - "No one really knows how things transpired."

Whole lot of people saying "What if" and "It seems". MS and AMD are said to be working together (per AMD) to fix the issue and send out an update this month.
 
  • Like
Reactions: Zucker2k

LikeLinus

Lifer
Jul 25, 2001
11,518
670
126
Especially when the developers' build already has the fix before Win11 launched.
Link?


"The other issue has to do with the “preferred core” feature in UEFI systems not scheduling tasks for the fastest core available in a given scenario. This could affect the performance of applications that rely mostly on a single or a few cores, since multi-threaded tasks would use all of the available cores anyway. As such, the performance impact may be more significant in processors with high core and thread counts, specifically those with a 65W TDP or higher. AMD doesn’t explicitly say how big the performance impact may be."

"AMD says it will arrive via a “software update”, not a Windows update, so it may come in the form of a new driver release for Windows 11. This fix, too, is expected sometime this month."

Doesn't sound like it's a part of the Windows Kernel if AMD is releasing a software/driver update. Also, how is MS responsible for UEFI.
 
  • Like
Reactions: Zucker2k

LikeLinus

Lifer
Jul 25, 2001
11,518
670
126
Follow the discussion
How do you fix something that wasn't previously broken? I think you mean something changed in the RTM that caused an issue with performance. AMD and MS are working together to patch the driver or software update and AMD will also have a UEFI update because it's causing performance issues.

As far as the laughable conspiracy theories that MS is doing this for Intel's benefit.

 
Last edited:

leoneazzurro

Senior member
Jul 26, 2016
905
1,429
136
How do you fix something that wasn't previously broken? I think you mean something changed in the RTM that caused an issue with performance. AMD and MS are working together to patch the driver or software update and AMD will also have a UEFI update because it's also causing performance issues.

The problem was known before the launch of Win11 and the part of MS fix was not included, this is obviously for rushing the WIn11 release. Also, when something working in Win10 does not work in Win11, it's clear that someone other than AMD changed the way the scheduler "reads" the processor's capabilities. I am not following the conspiracy theories for favoring Intel - but obviously MS did not a good job here.
 
Feb 4, 2009
34,494
15,729
136
I am going to simplify this entire discussion:

It is Microsofts job to be ensure win 11 runs properly on supported hardware, if they fail people like me will hesitate & have doubts

It is AMD’s job to make sure or patch supported hardware works properly, otherwise buyers become fearful.

They both failed. New OS release shouldn’t noticeably impact performance. Patch should have arrive on or before release date.
 
  • Like
Reactions: podspi and Zucker2k

Joe NYC

Golden Member
Jun 26, 2021
1,893
2,185
106
Did AMD not do throughout testing on Windows 11 before the release and never told the public until after it was launched and others found out?

Microsoft was making wholesale changes, as if this was not a real application, used by billion people, but as if it was prototyping...
 

LikeLinus

Lifer
Jul 25, 2001
11,518
670
126
The problem was known before the launch of Win11 and the part of MS fix was not included, this is obviously for rushing the WIn11 release. Also, when something working in Win10 does not work in Win11, it's clear that someone other than AMD changed the way the scheduler "reads" the processor's capabilities. I am not following the conspiracy theories for favoring Intel - but obviously MS did not a good job here.

When was this first reported? You are saying it worked in the beta builds, but then also saying it was a known issue before it was released. All of the reports I've seen (even your reddit post) are after the Oct. 5th release. You are saying the "fix" was removed for reasons unknown. Why would they remove something that hurts performance for absolutely no reason? Wouldn't it take more time to remove a working feature rather than just leave it be? Makes no sense. Where are the reports before launch?
logic would dictate that something else was changed that caused the issue was it was unnoticed until released to the pubic. Not some nefarious plan by MS to gimp a working scheduler. :rolleyes:

What about the UEFI issue that AMD says they are working on?

"UEFI CPPC2 (“preferred core”) may not preferentially schedule threads on a processor’s fastest core."
 
  • Like
Reactions: Zucker2k

leoneazzurro

Senior member
Jul 26, 2016
905
1,429
136
When was this first reported? You are saying it worked in the beta builds, but then also saying it was a known issue before it was released. All of the reports I've seen (even your reddit post) are after the Oct. 5th release. You are saying the "fix" was removed for reasons unknown. Why would they remove something that hurts performance for absolutely no reason? Wouldn't it take more time to remove a working feature rather than just leave it be? Makes no sense. Where are the reports before launch?
logic would dictate that something else was changed that caused the issue was it was unnoticed until released to the pubic. Not some nefarious plan by MS to gimp a working scheduler. :rolleyes:

What about the UEFI issue that AMD says they are working on?

"UEFI CPPC2 (“preferred core”) may not preferentially schedule threads on a processor’s fastest core."

The fix was obviously known before because the developer builds where the fix appears are from late August/September. If you are able to do some math and use simple logic, you could also undestand that for a fix to appear there must be a) some reporting -that of course does not have to be public as it was done for an unreleased product- b) some time for programmers to develop it.
 

LikeLinus

Lifer
Jul 25, 2001
11,518
670
126
The fix was obviously known before because the developer builds where the fix appears are from late August/September. If you are able to do some math and use simple logic, you could also undestand that for a fix to appear there must be a) some reporting -that of course does not have to be public as it was done for an unreleased product- b) some time for programmers to develop it.
You keep calling it a fix in previous builds. Then you go on to say "that of course does not have to be public." So you mean there is no proof there was ever a fix or ever even actually an issue previous to the RTM? But for some reason you know for sure, how exactly?

Explain to me how Math can help me understand something you call a "Fix" that you have no proof that was ever broken. Feel free to post links regarding past betas (before launch) that says they fixed an issue with the scheduler specific to AMD CPUs.

Notice you also avoid the entire UEFI topic as well. Can't explain that one, can you? It reduces AMD performance as well. Nothing to do with Windows 11.
 
  • Like
Reactions: Zucker2k

leoneazzurro

Senior member
Jul 26, 2016
905
1,429
136
You keep calling it a fix in previous builds. Then you go on to say "that of course does not have to be public." So you mean there is no proof there was ever a fix or ever even actually an issue previous to the RTM? But for some reason you know for sure, how exactly?

Explain to me how Math can help me understand something you call a "Fix" that you have no proof that was ever broken. Feel free to post links regarding past betas (before launch) that says they fixed an issue with the scheduler specific to AMD CPUs.

Notice you also avoid the entire UEFI topic as well. Can't explain that one, can you? It reduces AMD performance as well. Nothing to do with Windows 11.

If you are uncapable to read, follow a reddit discussion to find the relevant posts, or use simple logic, it is not my fault. J Like, you continue to speak about UEFI, but UEFI works under Win10. So, by using simple logic, who has broken what? Of course it has to do with Win11. And I am tired of your redacted, go trolling somewhere else.




profanity is not allowed in the tech forums.


esquared
Anandtech Forum Director
 
Last edited by a moderator:

Panino Manino

Senior member
Jan 28, 2017
813
1,010
136
My "math" is simpler.

The sole biggest change done in Windows is the explicit support for Alder Lake in the scheduler. That's the biggest and most relevant change in how W11 works, right?

To Microsoft it was more important to release W11 fast in time for Alder Lake than it was important to them to not cripple any AMD CPUs with the same new scheduler that would inevitably produce headlines for weeks on all tech related websites hurting AMD's public image just before a big marketing campaign from Intel to sell the new CPUs.

Very easy to calculate the losers and winners in this situation and once again shows just how strong is Intel's "softpower".


On an anectodical note: Back in the 80s Intel proposed to them implementations of Excel functions that were up to 15 times faster than their own. You know what MS said? People buy Excel because of the features.

Do you have a source for this story?
 

LikeLinus

Lifer
Jul 25, 2001
11,518
670
126
If you are uncapable to read, follow a reddit discussion to find the relevant posts, or use simple logic, it is not my fault. J Like, you continue to speak about UEFI, but UEFI works under Win10. So, by using simple logic, who has broken what? Of course it has to do with Win11. And I am tired of your bullshit, go trolling somewhere else.
Reddit is a forum. Same as this. It is not a news source and there there are no reports that Win 11 had this originally working and then broke it on purpose. Which is exactly what you are claiming. Simple logic doesn't pertain because there is no proof from a reputable source. If you sit around and believe everything you read in that thread, you will also notice that someone in that thread said this also wasn't just an issue with AMD. They claim they are having the same issue with an Intel CPU.

The UEFI issue isn't a fault of Win 11, It's the part of the motherboard's firmware and the AMD driver.


"AMD strongly recommends that your motherboard firmware (“UEFI”) be configured full UEFI Mode to ensure optimal performance, compatibility, and stability with the Windows® 10 operating system."


"Second, 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."

This is an issue with AMD's own driver with a new OS. This isn't a fault of MS. This happens with any new OS release where old drivers that haven't been updated during the OS development (via testing) may not function correctly due to changes in said new OS.

But please, enlighten us on how wrong ExtremeTech is and how they can't do "math and simple logic".

No one is trolling. I simply said AMD and MS are both to blame. You also can't seem to provide a link to anything that backs up your claims from a reputable source. Odd, isn't it?
 
Last edited:

leoneazzurro

Senior member
Jul 26, 2016
905
1,429
136
You are twisting again everything.
A) Reddit is a forum where users are signalling that WIn11 developers build with date way earlier than the launch date of Win11 have no issue with the L3 cache problem- and specifically with the cache latency. This is called evidence but you are ignoring it. The intel users reports were about the security features causing performance loss in games, which is another well know problem of WIn11 that was discussed elsewhere. In any case, the fact is that developer versions coming earlier than the release date of Win11 included a fix that did not come with the release version and MS decided to release it anyway. And for having a fix, there must have been a report, too.
B) UEFI works with Win10 but when using Win11 it does not work anymore. This is the only fact. It may be AMD's fault but it can be also MS fault, you only know that the driver is not working properly with the new OS. But the point is,, this too has been known earlier, because AMD already knew on the launch date what did not work and why (as they promised a fix in a very short time). Do you think that they did not tell it to MS, and why? When they went public about the issue on the launch day?
No conspiracy theories, only thing I see is that MS wanted to rush the release date at all costs.
 

jpiniero

Lifer
Oct 1, 2010
14,509
5,159
136
A) Reddit is a forum where users are signalling that WIn11 developers build with date way earlier than the launch date of Win11

The RTM is probally from August 31st. That's the date they announced the release date of last Tuesday. Need time to elapse to get the RTM onto OEM machines.
 

LikeLinus

Lifer
Jul 25, 2001
11,518
670
126
You are twisting again everything.
A) Reddit is a forum where users are signalling that WIn11 developers build with date way earlier than the launch date of Win11 have no issue with the L3 cache problem- and specifically with the cache latency. This is called evidence but you are ignoring it. The intel users reports were about the security features causing performance loss in games, which is another well know problem of WIn11 that was discussed elsewhere. In any case, the fact is that developer versions coming earlier than the release date of Win11 included a fix that did not come with the release version and MS decided to release it anyway. And for having a fix, there must have been a report, too.
B) UEFI works with Win10 but when using Win11 it does not work anymore. This is the only fact. It may be AMD's fault but it can be also MS fault, you only know that the driver is not working properly with the new OS. But the point is,, this too has been known earlier, because AMD already knew on the launch date what did not work and why (as they promised a fix in a very short time). Do you think that they did not tell it to MS, and why? When they went public about the issue on the launch day?
No conspiracy theories, only thing I see is that MS wanted to rush the release date at all costs.
"Do you think that they did not tell it to MS, and why? When they went public about the issue on the launch day?"

So, why you're also saying that AMD was aware of the scheduling issue because the issue was released on the same KB article.

You can't have it both ways. They were either aware of the issues before the launch or after. If before, they are just as complicit as MS in the handling of the situation.

What they should have done, for the UEFI issue, is proper testing on a Beta OS and already finalized a new driver for launch day. Not let it ride out and then say "Oh yeah, there is a problem with our driver, we'll fix it in the next few weeks!".

MS didn't just surprise drop Win 11. AMD had time to evaluate any performance issues and such and to warn users, BEFORE HAND, or to flat out fix the driver issues.

Either way, facts and evidence stand, both companies messed up. FYI - The conspiracy theory joke wasn't about you. It was about previous ones implying they did this for the upcoming Intel CPUs, lol.

FYI - I'm not "ignoring evidence". Feel free to point me to any facts and data in that Reddit thread that is actual evidence. I'm not scrolling through a thousand replies to find a needle in a haystack from some forum member's response that you claim is evidence that this was a "FIX" (your words) in a previous Beta and worked fine and then MS removed the fix. I gave you links and plenty of real evidence from AMD and reputable sites. You grabbing a link to Reddit and calling it evidence is laughable.
 
Last edited:

leoneazzurro

Senior member
Jul 26, 2016
905
1,429
136
I am saying that they knew about both issues beforehand, they told MS and MS completely ignored it because they had a launch date to meet. You cannot know if AMD did not tell MS beforehand, as evidence says that MS knew also this beforehand as the fix for one issue were present much before the release date. About the "broken driver": things are not so easy. We have a long hystory with MS and "broken drivers" at launch.
 
  • Like
Reactions: Tlh97 and Thibsie

leoneazzurro

Senior member
Jul 26, 2016
905
1,429
136
The RTM is probally from August 31st. That's the date they announced the release date of last Tuesday. Need time to elapse to get the RTM onto OEM machines.

Yes, I know, however one of these builds is from 24/8 and it takes time for releasing developer builds, too...

FYI - I'm not "ignoring evidence". Feel free to point me to any facts and data in that Reddit thread that is actual evidence. I'm not scrolling through a thousand replies to find a needle in a haystack from some forum member's response that you claim is evidence that this was a "FIX" (your words) in a previous Beta and worked fine and then MS removed the fix. I gave you links and plenty of real evidence from AMD and reputable sites. You grabbing a link to Reddit and calling it evidence is laughable.

Lol, what is laughable is that no one of your links gives you proof about you are blabbering, they all only state the very same thing, that is "there are issues with Ryzen and Win11, a fix is coming". No attribution of responsability to either company. Nothing more.
 

LikeLinus

Lifer
Jul 25, 2001
11,518
670
126
I am saying that they knew about both issues beforehand, they told MS and MS completely ignored it because they had a launch date to meet. You cannot know if AMD did not tell MS beforehand, as evidence says that MS knew also this beforehand as the fix for one issue were present much before the release date. About the "broken driver": things are not so easy. We have a long hystory with MS and "broken drivers" at launch.
MS isn't responsible for AMD's own driver. I'm not sure how you are unable to understand that. I gave you a link that says"

Second, 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.

AMD is warning that the performance hit may be more detectable in chips with 8 or more cores and a TDP above 65W, which basically covers the Ryzen 9 3900X/5900X and 3950X/5950X.


AMD makes the driver. Yet you are saying that it's not AMD's responsibility. You're funny, buddy.

Not to mention, you are talking out of both sides of your mouth!

"I am saying that they knew about both issues beforehand, they told MS and MS completely ignored it "

"You cannot know if AMD did not tell MS beforehand, as evidence says that MS knew also this beforehand.."


Wow, just wow. Please provide evidence that MS knew before the RTM launch that there was a scheduler issue with AMD CPUs. LINK it from a reputable source. No excuses this time, put your money where your mouth is.

Again, IT'S NOT A FIX IF IT WASN'T PREVIOUSLY BROKEN. You keep using the word FIX. So PLEASE provide proof that there was a scheduler issue with AMD CPUs in Betas before the Betas you say they "fixed".

Enjoy your night. This is getting ridiculous and you want to talk about logic. LOL