AMD fTPM causing random stuttering.

[Guest -Senku-]

discrete chips do not solve tpm stutter, had to do clean install was playing audio while installing a lot of stuff on my 980 pro 2 tb nvme from Samsung, and i had the longest tpm stutter, had a few very short tpm stutters to but haven't had one since i disabled SHA-1 which from if i understood is bassicly the old tpm 1.3 or 1.2 standard, which windows 11 can fall back to, but im not sure google is not exactly useful for searching that kind of info, im not a tpm tech wiz.

 

If you where hoping for a easy solution don't get your hopes up, also i would not call this ftpm stutter, its also happening with non ftpm chips

[caramelgeoff]

53 minutes ago, -Senku- said:

discrete chips do not solve tpm stutter, had to do clean install was playing audio while installing a lot of stuff on my 980 pro 2 tb nvme from Samsung, and i had the longest tpm stutter, had a few very short tpm stutters to but haven't had one since i disabled SHA-1 which from if i understood is bassicly the old tpm 1.3 or 1.2 standard, which windows 11 can fall back to, but im not sure google is not exactly useful for searching that kind of info, im not a tpm tech wiz.

 

If you where hoping for a easy solution don't get your hopes up, also i would not call this ftpm stutter, its also happening with non ftpm chips

Good to know.

 

I am going to start decrypting my drives so I can turn off bitlocker and tpm and cancel my order on a discrete TPM.

[Guest -Senku-]

Also that vivetool fix probably wont fix it, this issue is very rare it can happen once a day to once a week even once a month its very random, you would not even notice it while being afk, the real issue with it is the lack of attention from Microsoft and people calling it ftpm issue, im pretty sure if read somewhere Intel users having issues as well, and Linux users never had issue, no idea tho if linux actually supports tpm

 

I believe those settings like SHA-256 are required if you have a drive that support it or something else needs it, i have no idea what actually uses tpm in windows 11 and i think its about time Microsoft explained this as well why you need tpm 2.0 else i think its about time for a lawsuit, cos why would you pay for something you do not need.

[gymleader]

@solidsniper or anyone else who thinks they have a decent way to show stutters. Can you be my guinea pigs in testing my above idea?

 

Just found out you can get core isolation under device security if you enable amd virtualization. How to on Asus: https://www.asus.com/support/FAQ/1038245/

 

Other manufacturers likely similar. You should now see something like this (while fTPM enabled ofc):

 

Try just this and see if any stutters happen. If they still do, click "Core isolation details" and enable "Memory Integrity". Save changes then restart.

 

See if that does anything. If both still stutter I was completely wrong but always worth a try. You miss 100% of the shots you don't take - Wayne Gretzky, Michael Scott.

 

 

 

[apoklyps3]

none of the above fixed anything for me.

the vivetool command didn't fix it.

I already had core isolation disabled since I upgraded to windows 11.

 

unfortunately I can't seem to be able to disable fTPM from UEFI settings. Only gives me option to toggle between fTPM and hardware TPM

[Antistar]

8 minutes ago, apoklyps3 said:

none of the above fixed anything for me.

the vivetool command didn't fix it.

I already had core isolation disabled since I upgraded to windows 11.

 

unfortunately I can't seem to be able to disable fTPM from UEFI settings. Only gives me option to toggle between fTPM and hardware TPM

I can confirm.

[Guest -Senku-]

On 1/27/2022 at 3:45 PM, djmakk said:

I have test both a 3700x, 3800x, and a 5800x in the same board. Same issue. Both cleaned installs. If its CPU dependent then the fault rate is extremely high. Also dont pcie bus errons get logged in Event Viewer? There is never an entry with this stutter. 

 

Also if this is the case, then why does turning off the fTPM fix it?

 

PCI-E issues ? explain why they are happening only with tpm enabled ?

[gymleader]

Ah well thanks for trying @apoklyps3 @Antistar

 

Back to square one.

[ouroesa]

I enabled fTPM in my BIOS, installed win11 and disabled TPM. Havnt had an issue so far. Been through a few substantial updates and hawrdware upgrades (monitor, boot SSD) since then. Maybe disable TPM untill the cause is found and solution implemented - if at all possible.

[djmakk]

On 1/29/2022 at 10:29 PM, Amaranth said:

Interesting, I have yet to see an issue on any of my AMD machines under Windows 11 (running a 5950x, 3950x, and 4750U Pro respectively).

Is there a conclusive test or way to trigger the stutter? I'm curious if I've gotten lucky or if I've just missed it happening.

What motherboard?

 

Also anyone who RMAed a motherboard actually have their issue fixed by this? Asus keeps saying this is the only option.

[Guest -Senku-]

Im not disabling mine unless it starts to become really annoying like daily, i do not believe i fixed it unless i had 2 months without issues, and i already know i did not cos i got it in first hour of doing clean install with discrete chip installed and even configured in the bios

 

If you have no issues with tpm enabled its probably cos its not enabled, by default with older bioses its set to discrete even without chip installed that is the same as having it disabled, discrete is a small chip like this.

 

They can vary in looks location on motherboard can be different to, on my x570-E its right next to the cpu.

[PhoboX]

Hello, first of all apologize for my English.
I was going crazy looking for what could happen to my PC when I got here. I tell you, I have the same problem with fTPM. At first I blamed the Power Reporting Deviation on HWiNFO reports a low percentage until I came to this forum.

 

I have an Asus board, X570 ROG STRIX F and a 5800X, when I change from fTPM to dedicated TPM the system becomes unstable and on reboot it changes to fTPM. So that temporary solution of disabling fTPM does not work for me, since I cannot disable it and I return to the same starting point.

 

Like all of you, I am waiting for a response from AMD and a solution because, although it is not excessively annoying, it can weigh you down at some point.

[gymleader]

14 minutes ago, PhoboX said:

So that temporary solution of disabling fTPM does not work for me

All ASUS BIOS can disable tpm afaik. People are getting confused that the fTPM menu is where you disable it but not true. I can't blame people for this the menus are confusing.

 

Correct way is:

 

Advanced > Trusted Computing > Security Device Support > Set to Disable

 

Press F10 to save changes and exit. fTPM will now be disabled forever unless you update BIOS.

[stage]

On 1/30/2022 at 6:29 AM, Amaranth said:

Interesting, I have yet to see an issue on any of my AMD machines under Windows 11 (running a 5950x, 3950x, and 4750U Pro respectively).

Is there a conclusive test or way to trigger the stutter? I'm curious if I've gotten lucky or if I've just missed it happening.

Can you open Event viewer, and see if you got any SCEP Certification errors?

It should look something like:

 

SCEP Certificate enrollment initialization for WORKGROUP\DESKTOP-7NTSCGA$ via https://AMD-KeyId-578c545f796951421221a4a578acdb5f682f89c8.microsoftaik.azure.net/templates/Aik/scep failed:

GetCACaps
GetCACaps: Not Found
{"Message":"The authority \"amd-keyid-578c545f796951421221a4a578acdb5f682f89c8.microsoftaik.azure.net\" does not exist."}
HTTP/1.1 404 Not Found
Date: Thu, 20 Jan 2022 22:06:17 GMT
Content-Length: 121
Content-Type: application/json; charset=utf-8
X-Content-Type-Options: nosniff
Strict-Transport-Security: max-age=31536000;includeSubDomains
x-ms-request-id: 62e05544-b2a9-4156-8994-0ed5f324efbd

Method: GET(297ms)
Stage: GetCACaps
Not found (404). 0x80190194 (-2145844844 HTTP_E_STATUS_NOT_FOUND)

 

Pretty sure it has to be related since azure has something to do with TPM verification. Anyone else has this errors and also has the stutters?

 

And also registered just to say that I am also experiencing this, with 5600x and Asus tuf  b550 gaming plus motherboard, running windows 10 21h2. Only solution was to disable TPM from the bios. And thanks to stumbeling into this thread, so at least I know am not going crazy and is not something specific to my system.

 

Only thing I still wonder is this AMD, the MB manufacturers or Windows fault, or any combination of the 2/3?

 

 

There is also this interesting microsoft thread:

TPM event logger error after cpu swap, Event id 86

 

Where in the last posts there is this:

Quote

GiuseppeVellucci-9503 • Jan 19 2022 at 10:01 PM | Isaac-3107 edited • 6 days ago

Just for reference, I installed a discrete TPM and the error is gone. Definitely an AMD fault.

 

[PhoboX]

3 hours ago, gymleader said:

All ASUS BIOS can disable tpm afaik. People are getting confused that the fTPM menu is where you disable it but not true. I can't blame people for this the menus are confusing.

 

Correct way is:

 

Advanced > Trusted Computing > Security Device Support > Set to Disable

 

Press F10 to save changes and exit. fTPM will now be disabled forever unless you update BIOS.

Thanks a lot, gymleader. I'm going to try this solution.

[gymleader]

@stageOne thing I will say is that if you take Asus Crosshair boards out of the equation, the other manufacturers flagship boards are widely unaffected. For how popular the MSI X570 Tomahawk is I can't see a single report of that board stuttering (that I remember seeing anyway). Same with the Unify or Godlike lines. The Gigabyte X570 Aorus Master also seems unaffected.

 

On the guru3d forum someone made a compelling case for faulty CPU that makes alot of sense.

 

"the forced/default tpm config in the latest bios is exposing it because of periodic validation colliding with a correctable bus error, which Ryzen cpu's have higher than the industry average of.

since the ftpm fires a locking interrupt IO delays until the interrupt service succeeds."

 

I know it sucks to say a faulty CPU but it's the only logical conclusion now. Sure if like everybody had the problem you could blame software but since most Ryzen users are fine it's probably just a bad batch.

 

 

[Guest -Senku-]

its not hardware stability issue the issue is litterally not there without tpm and even happens with discrete tpm chip.

[gymleader]

That only further makes it a hardware fault though. Both the fTPM and physical TPM have to do periodic validation and if it just happens to happen while a spike of correctable bus errors happen, you get the stutter.

[stage]

16 hours ago, gymleader said:

@stageOne thing I will say is that if you take Asus Crosshair boards out of the equation, the other manufacturers flagship boards are widely unaffected. For how popular the MSI X570 Tomahawk is I can't see a single report of that board stuttering (that I remember seeing anyway). Same with the Unify or Godlike lines. The Gigabyte X570 Aorus Master also seems unaffected.

 

On the guru3d forum someone made a compelling case for faulty CPU that makes alot of sense.

 

"the forced/default tpm config in the latest bios is exposing it because of periodic validation colliding with a correctable bus error, which Ryzen cpu's have higher than the industry average of.

since the ftpm fires a locking interrupt IO delays until the interrupt service succeeds."

 

 

Hi, I am not sure I completely follow the logic here. You are saying is hardware issue but then certain boards "seem" unaffected. So then why make it faulty CPU.

 

I also don't buy that it was a faulty batch, since the users are on many different locations. I am in Eastern Europe and most people in this thread are (I assume) from USA and Canada. Having a faulty batch with such wide spread seems unlikely.

 

So faulty CPU or faulty motherboards? And if the issue did not exist before the ftpm update to the bios how is that not a faulty Windows 10/11 problem? Were there ppl having the same issues under lunux/macos? (tbh I don't even know if there is tpm on those platforms)

[gymleader]

@stage Well I mean I'm just going by memory. To see if that was true you would have to scroll through every single forum/reddit post and log what boards had problem and then make an aggregated list. Just for example I know people with Dark Hero boards who don't have the issue but other Dark Hero owners who do. In that respect you can edge towards a hardware fault. People have tried so many different drivers and BIOS versions that if it was a software issue, some specific combination would fix it.

 

As for the batch stuff, I'm not too versed in AMD's manufacturing but don't they only have Taiwan (TSMC)? Then it doesn't matter where the product gets distributed too since it all originates from the same location.

 

I know it sucks for it to be a hardware fault that can't be fixed but sometimes you can't bury your head in the sand expecting a different result. Sure it sucks but could be worse. All I know is I'd rather have this problem than the USB disconnect issue that plagued last year.

[VideoNomad]

Actually had this issue from the very beginning of my Win11 install. Was testing everything that I can, hard drives, GPU, CPU (3900x) etc. etc.
And here it is... fTPM was the culprit! It was stuttering at least 2-3 times a day.

Tried disabling the fTPM with Windows 11 installed and was surprised that it booted normally. I don't use bitlocker though.

No stutters so far. Not sure about the long term effects of disabling fTPM on Windows 11, but we will see. Hopefully AMD/Microsoft fixes this.

[djmakk]

18 hours ago, gymleader said:

@stageOne thing I will say is that if you take Asus Crosshair boards out of the equation, the other manufacturers flagship boards are widely unaffected. For how popular the MSI X570 Tomahawk is I can't see a single report of that board stuttering (that I remember seeing anyway). Same with the Unify or Godlike lines. The Gigabyte X570 Aorus Master also seems unaffected.

 

On the guru3d forum someone made a compelling case for faulty CPU that makes alot of sense.

 

"the forced/default tpm config in the latest bios is exposing it because of periodic validation colliding with a correctable bus error, which Ryzen cpu's have higher than the industry average of.

since the ftpm fires a locking interrupt IO delays until the interrupt service succeeds."

 

I know it sucks to say a faulty CPU but it's the only logical conclusion now. Sure if like everybody had the problem you could blame software but since most Ryzen users are fine it's probably just a bad batch.

 

 

I have both a 3800x and a 5800x that both do this. That is surprising luck. 

[enisand]

Hello. Sorry for my english. Who can tell me how to disable Ftmp on an ASUS ROG Strix G17 G713QM laptop? BIOS is locked, there is no disable function. The problem with stuttering occurs 1 to 5 times a day. Help me please. Thank you.

[PhoboX]

4 hours ago, VideoNomad said:

Actually had this issue from the very beginning of my Win11 install. Was testing everything that I can, hard drives, GPU, CPU (3900x) etc. etc.
And here it is... fTPM was the culprit! It was stuttering at least 2-3 times a day.

Tried disabling the fTPM with Windows 11 installed and was surprised that it booted normally. I don't use bitlocker though.

No stutters so far. Not sure about the long term effects of disabling fTPM on Windows 11, but we will see. Hopefully AMD/Microsoft fixes this.

Me too. I would like to know what are the consequences of disabling fTPM. At the moment, I have not experienced stuttering again.

[DiceGG]

4 hours ago, gymleader said:

@stage Well I mean I'm just going by memory. To see if that was true you would have to scroll through every single forum/reddit post and log what boards had problem and then make an aggregated list. Just for example I know people with Dark Hero boards who don't have the issue but other Dark Hero owners who do. In that respect you can edge towards a hardware fault. People have tried so many different drivers and BIOS versions that if it was a software issue, some specific combination would fix it.

 

As for the batch stuff, I'm not too versed in AMD's manufacturing but don't they only have Taiwan (TSMC)? Then it doesn't matter where the product gets distributed too since it all originates from the same location.

 

I know it sucks for it to be a hardware fault that can't be fixed but sometimes you can't bury your head in the sand expecting a different result. Sure it sucks but could be worse. All I know is I'd rather have this problem than the USB disconnect issue that plagued last year.

While I think this theory is very interesting and plausible, there's still one thing that keeps me from being 100% convinced:
A batch would be a few hundred CPUs. But this persists across generations of CPUs. People with 3000 series, 5000 series and even some with 1000 series Ryzens are reporting this issue.
Not trying to argue with you, just trying to understand / learn.