First Windows 11 Patch Tuesday slows down AMD CPUs even more than the release build

Rolled the dice and upgraded to Win 11 on my 5800x/Aorus Xtreme based gaming rig. Haven't seen any loss in performance. Yeah, yeah, I know - anecdotal. But seriously... my Aida64 scores are the same as they were on Win 10. So there has to be more to the story...

VBS was already disabled on my system.
Same for me on R5 3600 + 5600XT. Updated from 10 to 11 dev on July and it had some performance issues but now it just working as it should be. VBS was disabled by default.
 
I know MS and AMD are working together on both issues, so couldn't this be both there faults? Or AMD's? I only say that, because after it was revealed that Zen 1000 series didn't have W11 support, I didn't hear any complaints from AMD.
MS is the who decides which updates get pushed to users.
 
Have you played it recently? CP2077 on v1.31 is still a buggy trainwreck. I know because I've played this version start to finish, naively expecting the game to be more polished at this point. I heard quite a few new bugs were also introduced since v1.3. Nearly one year and several patches later, CP2077 is still buggier than a Bethesda open world RPG on launch day.
Let's not exaggerate there. Bathesda is still the king of bugs. :)
 
Let's get the thin foil hat on for fun: wouldn't it be pretty convenient for intel if Microsoft just didn't have time to fix a performance regression on Windows 11 big enough to get Alder Lake a great first impression on the reviews cycle, then later on when it turns out it's just basically the same than Zen3 without the unfortunate yet perfectly timed issues Microsoft is throwing at AMD well, reviews are out, orders are made and people's mind are not going to move from the "intel indeed is back!" mentality.

Now getting off that conspiracy theory I think it's just a simple case of AMD being a second-class citizen to Microsoft still: They didn't work with AMD for the initial Ryzen launch, the initial threadrippers had even more problems on Windows than Ryzen 1 and now it seems like it's the same story: intel probably gets a dedicated team of Microsoft employees working around the clock to tune in performance while AMD might get some white papers in an email several weeks too late to react in a "here you go, figure this out yourselves" kind of attitude.

This... ^ I am mi;dy curious if you turn off AMD's hyperthreading which is vastly different that Intels if the issue lessens ..

 
Let's not exaggerate there. Bathesda is still the king of bugs. :)

Not anymore. I'm being serious. I have played Oblivion, Fallout 3 and Skyrim on launch, they were buggy of course, but I didn't experience as many glitches and bugs (including progression breaking ones) as I did with CP2077...
 
We already have scheduler issues facing AMD's Zen3 platform on windows10 that are not getting fixed, so these issues facing windows 11 are not going to be far off. I think it will come down to money shifting for special partner support resources (AKA Timed engagements) and since Intel has more money to throw at the issue they will get their stuff worked out before AMD can.

I also do not think your tinfoil hat is wrong. There are things I can say here that would outright violate a couple NDAs that would interest you, but just look at my scheduler comments and look here -
 
Now getting off that conspiracy theory I think it's just a simple case of AMD being a second-class citizen to Microsoft still: They didn't work with AMD for the initial Ryzen launch, the initial threadrippers had even more problems on Windows than Ryzen 1 and now it seems like it's the same story: intel probably gets a dedicated team of Microsoft employees working around the clock to tune in performance while AMD might get some white papers in an email several weeks too late to react in a "here you go, figure this out yourselves" kind of attitude.

Which is funny because is not like Microsoft is some kind of startup company made up of a bunch of college friends based on Johnny's mom basement.

Like, there are two x86 chipmakers around and MS has what? thousands of programmers? I just can't fathom how is it possible to have this kind of issues and then have the issue made worse by their own patches.

Think what you will, but this smells like Wintel from a mile away lol
 
One would assume the MS would iron out these core/serious issues before releasing a new OS. I wonder what they were working on as a priority, visuals?
 
Not anymore. I'm being serious. I have played Oblivion, Fallout 3 and Skyrim on launch, they were buggy of course, but I didn't experience as many glitches and bugs (including progression breaking ones) as I did with CP2077...
I did with Skirim and we have a newer title we can use for comparison, right? Falout 76...

I think this Internet Historian video kinda sums things up rather well:
 
They lost their QA team about the same time they replaced all their competent UX developers with Junior Concept Artist work experience interns...
They didn't "lose" them, they replaced them with ignorant, incompetent Insiders.
 
I did with Skirim and we have a newer title we can use for comparison, right? Falout 76...

I think this Internet Historian video kinda sums things up rather well:

I often even forget Fallout 76 exists.

I have never played it, but I've seen people saying at its worst, it's just as bad as CP2077...
 
> To do so, go to Settings -> System -> Recovery. You should see a "Go Back" button, which can be used to start the downgrade process.

Isn't that an upgrade at this point?
 
I wonder how much cash Intel gave M$ for this.

I'm wondering how much cash AMD gave Microsoft for this. It's a cool move to get CPU's obsoleted and pointing the finger of ultimate-blame at Intel / MS.

Good move AMD. Good move.
 
I see Microsoft is once again finding a way to help the other half of Wintel again. Every time AMD outdoes Intel,k Microsoft finds a screw them up started on socket 7 with 350 and 500 processors and continues on to this day.
 
I see Microsoft is once again finding a way to help the other half of Wintel again. Every time AMD outdoes Intel,k Microsoft finds a screw them up started on socket 7 with 350 and 500 processors and continues on to this day.

Zomb~ AMD soldered a 16C/32T Ryzen Threadripper together, THE biggest processor of the decade, WINDOWS clobbers AMD with Intel's Alder Lake CPU!

Fanzboys sure are making the 21th Century Wintel collab look good. /s
 
Let's get the thin foil hat on for fun: wouldn't it be pretty convenient for intel if Microsoft just didn't have time to fix a performance regression on Windows 11 big enough to get Alder Lake a great first impression on the reviews cycle, then later on when it turns out it's just basically the same than Zen3 without the unfortunate yet perfectly timed issues Microsoft is throwing at AMD well, reviews are out, orders are made and people's mind are not going to move from the "intel indeed is back!" mentality.

Now getting off that conspiracy theory I think it's just a simple case of AMD being a second-class citizen to Microsoft still: They didn't work with AMD for the initial Ryzen launch, the initial threadrippers had even more problems on Windows than Ryzen 1 and now it seems like it's the same story: intel probably gets a dedicated team of Microsoft employees working around the clock to tune in performance while AMD might get some white papers in an email several weeks too late to react in a "here you go, figure this out yourselves" kind of attitude.


Why would perfectly normal business cheating be "thin foil hat"??

- Wouldn't it be great to make a computer virus, and then be the first one to produce antivirus for it and earn millions? McAfee joined the chat. Even though he's probably dead (this time for real).

- Could it happen that car manufacturers agree on the prices, so that all the cars from "B" segment cost the same, all the cars from "C" segment cost the same, etc. European car producers joined the chat. And were penalized for monopolistic agreement back in 2005.

- Imagine if someone would add a special mode to their drivers which is activated when they recognize a benchmark running. So they can "optimize" their graphics cards to run faster in the benchmarks, but with a lower quality (even though higher quality is set in the benchmark). Nvidia joined the chat.

- What if someone creates a special mode for gasoline/diesel engines so they produce less CO2 but only during testing. VW and all other engine manufacturers joined the chat (but only VW got penalized).

Etc, etc. Cheating is pretty normal in business. Shouldn't be, but in our imperfect world it is.
 
One would assume the MS would iron out these core/serious issues before releasing a new OS. I wonder what they were working on as a priority, visuals?

You obviously aren't an experienced Windows user.

Remember Windows ME? Total failure. As if they didn't have anyone testing the product before release.

Or Windows Vista. What a fail.

Or Windows 8. They should have fired at least 100 people for that fail. How hard would it be to send the app to 1000 test users and collect their reports? But nope, they just released that crap. Fortunately, they fixed some of the problems in the patch called Windows 8.1.

So........ testing is not their stronger side. Seems they only have one guy to test Windows, but if a release happens while he's on vacation...... Windows Vista (or similar) gets released as soon as it successfully compiles.
 
Let's get the thin foil hat on for fun: wouldn't it be pretty convenient for intel if Microsoft just didn't have time to fix a performance regression on Windows 11 big enough to get Alder Lake a great first impression on the reviews cycle, then later on when it turns out it's just basically the same than Zen3 without the unfortunate yet perfectly timed issues Microsoft is throwing at AMD well, reviews are out, orders are made and people's mind are not going to move from the "intel indeed is back!" mentality.

Now getting off that conspiracy theory I think it's just a simple case of AMD being a second-class citizen to Microsoft still: They didn't work with AMD for the initial Ryzen launch, the initial threadrippers had even more problems on Windows than Ryzen 1 and now it seems like it's the same story: intel probably gets a dedicated team of Microsoft employees working around the clock to tune in performance while AMD might get some white papers in an email several weeks too late to react in a "here you go, figure this out yourselves" kind of attitude.

Why not both ?

As for dedicated teams - afaik Intel has so far contributed code to every Windows version, so actual Intel coders (employees) working on the scheduler is very well possible.
 
Why not both ?

As for dedicated teams - afaik Intel has so far contributed code to every Windows version, so actual Intel coders (employees) working on the scheduler is very well possible.
Intel working with MS on the scheduler is something that actually happened. They helped each other for both the windows software scheduler and the on chip scheduler.
 
Intel working with MS on the scheduler is something that actually happened. They helped each other for both the windows software scheduler and the on chip scheduler.
So zero surprise about the ‚side effects‚ then. Seems like some things never change….
 
Intel working with MS on the scheduler is something that actually happened. They helped each other for both the windows software scheduler and the on chip scheduler.

It would also be interesting to know which compiler Microsoft use to compile their apps. I know they produce their own compilers, but it's possible they are using Intel compiler for certain low-level stuff (such as drivers or other kernel modules).

If they do, you can bet that Intel optimizing compiler is "optimizing" the code in such a way to execute slower on AMD. Even if it produces slightly slower code for their own CPUs, but it's important that it hurts AMD more.
 
It would also be interesting to know which compiler Microsoft use to compile their apps. I know they produce their own compilers, but it's possible they are using Intel compiler for certain low-level stuff (such as drivers or other kernel modules).

If they do, you can bet that Intel optimizing compiler is "optimizing" the code in such a way to execute slower on AMD. Even if it produces slightly slower code for their own CPUs, but it's important that it hurts AMD more.
I doubt that we'll see such things since AMD should have access to and be able to audit the compilers. If they are ever found to intentionally gimp the compiler for an OS then it would spell disaster for both Intel and MS.

There are plenty of other, less obvious, ways to optimise code for a certain architecture.
 
Back