Problems TPM 2.0 and AMD - performance issues with both Windows 11 and 10?

Published by

Click here to post a comment for Problems TPM 2.0 and AMD - performance issues with both Windows 11 and 10? on our message forum
#5988282
data/avatar/default/avatar31.webp
Astyanax:

Its a cpu hardware flaw, enjoy your increasingly decaying fabric or pcie controller.
I had this issue with the fTPM on my old 3900X. I popped in my new 5950X - same. I told my colleagues about it and they all report having it too. One friend doesn't experience this on either a 5950X nor on a 5900X. Besides staying away from AMD in the future I'm not sure what else is to be learned here.
#5988399
data/avatar/default/avatar20.webp
I'm sure i've got the best solution for this issue: Update Rollback - (KB5009596) 🙄:D
#5988588
data/avatar/default/avatar22.webp
iNerd:

I'm sure i've got the best solution for this issue: Update Rollback - (KB5009596) 🙄:D
Problem also exists in 10. The solution is to turn off fTPM or buy and install an SPI TPM module.
#5988599
https://forums.guru3d.com/data/avatars/m/197/197287.jpg
foxX:

Besides staying away from AMD in the future I'm not sure what else is to be learned here.
With all the Intel issues that people have...i'm not sure how you could come up with this "conclusion" other then trying to bash a company. every product from every company will have issues somewhere down the line, if you wish to have zero issues, never buy a product again of any kind. INB4 "Well i'm not having this issue on intel!" Well, i'm not having this issue on AMD...with a 5900x, an x570 board and fTPM is on. Like honestly, Intels security issues they have had for the past many years and the performance degradation from patching those holes, alder lake CPUs not being able to run specific games due to DRM (yes i understand this is supposedly fixed now, and maybe this will be fixed too, who knows), and various other issues intel CPUs and platforms have had over the years. I just do not understand your "conclusion"
#5988830
data/avatar/default/avatar12.webp
foxX:

Problem also exists in 10. The solution is to turn off fTPM or buy and install an SPI TPM module.
i run into this issue for the first time after this windows 10 optional update: (KB5009596). After rollback or de-installing the (KB5009596) update everything came back to normal. I also read on reddit that there is no guarantee that buying and installing SPI TMP is a solution. And after all - why should throw my own money because the microsoft rolled out a bad update? This would only work if we would be allowed to send the bill to microsoft after buying a SPI TMP Module. and even than - there should not be such an issue and the vendor should bring a fix for this 🙂
#5988839
https://forums.guru3d.com/data/avatars/m/156/156348.jpg
I had a couple of random stutters since i installed windows 11 but honestly it happens like once a week at max and it last for one or two seconds. Really not a big deal from my pov anyway.
#5988875
data/avatar/default/avatar31.webp
MonstroMart:

I had a couple of random stutters since i installed windows 11 but honestly it happens like once a week at max and it last for one or two seconds. Really not a big deal from my pov anyway.
on my pc, windows 10 home, after this windows update the performance from OS was kind feeling like after de-installing gpu and display drivers with DDU... 😀 no other comparisation to describe this 🙂
#5989028
https://forums.guru3d.com/data/avatars/m/166/166706.jpg
Astyanax:

Its a cpu hardware flaw, enjoy your increasingly decaying fabric or pcie controller.
it happens on intel too, its rarer like once per week
#5989310
https://forums.guru3d.com/data/avatars/m/273/273678.jpg
Raider0001:

it happens on intel too, its rarer like once per week
intel also uses a pcie link to communicate to the motherboard which can be subject to flaws, errors and misbehaviors that trigger interrupt spikes. the 99.9 percentile of intel users have no issues with tpm.
#5989372
https://forums.guru3d.com/data/avatars/m/166/166706.jpg
Astyanax:

intel also uses a pcie link to communicate to the motherboard which can be subject to flaws, errors and misbehaviors that trigger interrupt spikes. the 99.9 percentile of intel users have no issues with tpm.
i think we all have it to some degree
#5989855
data/avatar/default/avatar36.webp
foxX:

I had this issue with the fTPM on my old 3900X. I popped in my new 5950X - same. I told my colleagues about it and they all report having it too. One friend doesn't experience this on either a 5950X nor on a 5900X. Besides staying away from AMD in the future I'm not sure what else is to be learned here.
I think I also avoid AMD CPUs as well. I have already decided to never buy their gpu's again, especially not after 3200G, and both of my AMD systems have issues with tpm.(I think that CPU USB ports can even damage headphones with tpm on). I still don't get it why some people insist on faulty hardware. It was same story with 1st gen GCN after AMD bugged dxva upscaling/downscaling in 2016 and card would be stuck in UVD clocks. As long as I had 270x, bug was never fixed, it probably went bugged into legacy. This is simply just another proof that AMD needs to fire whole software department and hire new people. AMD can't make scheduler to work properly in time (but we can always blame intel, ms, nvidia...), they can't make hardware to function properly and bugs needs a long time to get fixed (unless reddit explodes), sometimes even eternity is not enough...
#5990458
https://forums.guru3d.com/data/avatars/m/281/281980.jpg
I have a 3700x + B450m and have absolutely no issues, currently using win 11… build 22000.469
#5990482
data/avatar/default/avatar26.webp
Astyanax:

Its a cpu hardware flaw, enjoy your increasingly decaying fabric or pcie controller.
Bullshit but hey, whatever, It is beautiful to see people using their stable hardware investment, every single day. 🙂
#5990538
https://forums.guru3d.com/data/avatars/m/259/259654.jpg
MonstroMart:

I had a couple of random stutters since i installed windows 11 but honestly it happens like once a week at max and it last for one or two seconds. Really not a big deal from my pov anyway.
I have the same experience, but I'm not cool with things like that in hardware so expensive.
Astyanax:

intel also uses a pcie link to communicate to the motherboard which can be subject to flaws, errors and misbehaviors that trigger interrupt spikes. the 99.9 percentile of intel users have no issues with tpm.
If it matters at all, it started happening to me in Windows 10, after enabling the fTPM. As mentioned above, it's very rare to happen, but it does. My memory is not out of spec (3200CL14, 1600 IF clock, learned not to play with bus clocks a long time ago), but even resetting my CPU curve undervolting and turning off PBO, made zero difference. There's something about this, I wouldn't be surprised if it needs a CPU firmware fix.
#5990690
https://forums.guru3d.com/data/avatars/m/287/287596.jpg
TPM grab hashes of your hardware and software state at particular stages and uses. Those value are called PCR and are stored for state integrity and correlation if done correctly. If there an issue with TPM it mean there an issue with some variable within the BIOS setup. Changing hardware state may trigger the issue on it's own, do as follow (do not post logs.txt around). Launch PowerShell with admin privilege and within the PowerShell applet type: 
tpmtool getdeviceinformation

tpmtool gatherlogs
tpmtool parsetcglogs -validate >parsetcglogs.txt
Copy the following NET method and paste the code within the PowerShell applet 
Add-Type -Assembly System.Windows.Forms
$state = [System.Windows.Forms.PowerState]::Suspend
[System.Windows.Forms.Application]::SetSuspendState($state, $false, $false) | Out-Null
Exit from the sleep state by pushing the board power button and within the PowerShell applet type again: 
tpmtool getdeviceinformation

tpmtool gatherlogs
tpmtool parsetcglogs -validate >parsetcglogs.txt
If there are changes TPM will signal the changes with different hashes. If those PCR changes, or cannot be validated, one by one exclude any non default BIOS menu configuration and communicate to your board manufacturer the offending value to deal with. Alternatively if PCR are all the same and can always be validated it could be a totally different issue. -no support will be offered anyway.
#5991233
data/avatar/default/avatar24.webp
I deinstalled that one update causing this issue and today few hours ago this optional update dissapeared from the update list and instead of that i've recieved this update on windows10: https://support.microsoft.com/de-de/topic/february-8-2022-kb5010342-os-builds-19042-1526-19043-1526-and-19044-1526-bd5c4434-bec4-42e9-991d-5810d4ec52d9 and since this one my system is running so much smoother / better / more stable and no stutter at all. so i can only recommend it to you. i believe, but maybe i'm wrong, that this will work for few ppl out there. if not? well you can still remove it. But from my point of view this runs really good until now. Take a look / download CVS file. it's a pretty long list.
#5991392
https://forums.guru3d.com/data/avatars/m/273/273678.jpg
this is a security only update and has no fixes or changes for anything performance wise.
#5991504
data/avatar/default/avatar25.webp
not only according to other resources i‘ve just read online.
#5991667
https://forums.guru3d.com/data/avatars/m/273/273678.jpg
iNerd:

not only according to other resources i‘ve just read online.
then you read nonsense, the performance fixes are always released in the end of month cumulative, unless an out of band patch is published with KB article to back it.
#5993149
data/avatar/default/avatar39.webp
i will now test this one: 2022-02 Vorschau auf kumulatives Update für .NET Framework 3.5 und 4.8 für Windows 10 Version 21H2 für x64 (KB5010472) + 2022-02 Kumulatives Update für Windows 10 Version 21H2 für x64-basierte Systeme (KB5010415)

Previous Page

Next Page