AMD fTPM causing random stuttering.

[DreiNeunVierEins]

Same problem here with my 3900X on Asus X570 TUF Plus, BIOS 4021 (newest).

I'll go the way of GameDr04 and get a discreet TPM module. That's a TPM SPI 14-1 pin for me. Just ordered one for around 25 bucks.

Should also help if I ever exchange my CPU for a 5900X. I can imagine that could cause some problems if the fTPM is actually used for anything.

[meemz]

On 10/6/2021 at 5:35 PM, Sete said:

Could you confirm if it is possible to update after doing this? I've seen some reports saying they couldn't update until they re-enable the TPM.

I just got an update and everything worked fine with fTPM off

[GameDr04]

5 minutes ago, meemz said:

I just got an update and everything worked fine with fTPM off

This is definitely welcome news!

Every time I do a clean install of Windows 10 (and now 11), I first unplug my network cable so that it doesn't force me to link my Microsoft account.

If I still only had an fTPM, I would consider disabling that as just one more pre-configuration step for installing Windows. Super-annoying, but we adapt, I guess...

 

On 10/12/2021 at 11:45 AM, DreiNeunVierEins said:

I'll go the way of GameDr04 and get a discreet TPM module. That's a TPM SPI 14-1 pin for me. Just ordered one for around 25 bucks.

It's been a week? I think? I haven't had a single stutter at all with my discrete TPM.

I'm still 10/10 annoyed that I had to go down this path but I'm at least glad to not have the problem anymore.

[AkikoKumagara]

13 hours ago, meemz said:

I just got an update and everything worked fine with fTPM off

 

12 hours ago, GameDr04 said:

This is definitely welcome news!

I've seen theories that Microsoft will allow updating on what they consider unsupported configurations excluding "feature updates". Time will tell, I guess, as far as that's concerned. Still not a big deal to flip on fTPM to install a feature update and switch it back off after completing the update if this is the case. I would rather not have to bother, or for fTPM to not cause this issue at all, but it's not the end of the world if this is how it works.

[CatNis]

I'm glad I've found this post.
I've had the same problem for a while and google search doesn't help unfortunately, it show pages and pages from "trusted" tech websites about Win 11 requiring TPM.

 

Here's a question: what does one do when the board manufacturer forces TPM on you in BIOS?

I'm talking about Asus. I'm on a X570 Dark Hero mb (and R9 5950x) and with the 3702 bios version they introduced compulsory TPM. I've discovered this when I tried to update from 3601 to 3801. What does compulsory TPM means? Well, the way it's supposed to work in this mb's bios is, you set the tpm to "discrete" (default in 3601) instead of "firmware" and if you don't have a discrete one, it will effectively be disabled. Now, in 3801 it's set to "firmware" by default and changing to "discrete" is disregarded - the next boot is reverted back to fTPM.

 

So now, I have the choice of forever staying on an older and older bios version or shrugging of the few stutterings per day. Or buy a discrete TPM chip.

Gotta love Asus for offering more options for their users. /s

[AkikoKumagara]

5 hours ago, CatNis said:

I'm glad I've found this post.
I've had the same problem for a while and google search doesn't help unfortunately, it show pages and pages from "trusted" tech websites about Win 11 requiring TPM.

 

Here's a question: what does one do when the board manufacturer forces TPM on you in BIOS?

I'm talking about Asus. I'm on a X570 Dark Hero mb (and R9 5950x) and with the 3702 bios version they introduced compulsory TPM. I've discovered this when I tried to update from 3601 to 3801. What does compulsory TPM means? Well, the way it's supposed to work in this mb's bios is, you set the tpm to "discrete" (default in 3601) instead of "firmware" and if you don't have a discrete one, it will effectively be disabled. Now, in 3801 it's set to "firmware" by default and changing to "discrete" is disregarded - the next boot is reverted back to fTPM.

 

So now, I have the choice of forever staying on an older and older bios version or shrugging of the few stutterings per day. Or buy a discrete TPM chip.

Gotta love Asus for offering more options for their users. /s

This kind of situation is exactly why this issue needs more eyes on it. This kind of strong-arm is worse than what Microsoft itself is doing, though I do understand it's only being done because of Microsoft's requirement in the first place.

 

I wish we had even some sort of clue why this is happening, or some official response from AMD or any of the motherboard manufacturers... I thought we might be onto something with a recent JayzTwoCents video about AMD and Win 11, but he didn't cover this particular issue.

 

For now, if it were me, I would stay on 3601 until it becomes an issue and leave fTPM disabled when/if you don't need it... at least until this is officially addressed, if ever.

[AkikoKumagara]

Update: I've found that when my stutters happen, in Event Viewer > Application, that it's reported that "The VSS service is shutting down due to idle timeout."

 

This is the service for Windows "Volume Shadow Copy". I went into services.msc to discover that Volume Shadow Copy Service was set to Manual start and was not running. I've set it to Automatic start and enabled it. I will report back as to whether this resolves anything, but it definitely seems suspicious that this service was dying while the stutter occurred, and that it was disabled when I'm fairly certain it should start automatically under normal circumstances.

 

Edit: I should clarify that I don't think this is related, but I'm willing to test and find if anything changes.

[Nyamp]

5 hours ago, Soph3D said:

Update: I've found that when my stutters happen, in Event Viewer > Application, that it's reported that "The VSS service is shutting down due to idle timeout."

 

This is the service for Windows "Volume Shadow Copy". I went into services.msc to discover that Volume Shadow Copy Service was set to Manual start and was not running. I've set it to Automatic start and enabled it. I will report back as to whether this resolves anything, but it definitely seems suspicious that this service was dying while the stutter occurred, and that it was disabled when I'm fairly certain it should start automatically under normal circumstances.

 

Edit: I should clarify that I don't think this is related, but I'm willing to test and find if anything changes.

 

Got a stutter today, I have the same event log on the exact time of the stutter! I'll try that too.

[AkikoKumagara]

4 hours ago, Nyamp said:

 

Got a stutter today, I have the same event log on the exact time of the stutter! I'll try that too.

Doesn't work, unfortunately. Still stuttering.

[dpeter]

Microsoft released a new Insider build(Beta and Release preview) that includes the L3 cache fix for Ryzen, maybe that solves it, im not currently on windows 11 due to these bugs but if anyone is willing to try...

[NimjaIV]

I'm not sure I said this before, but my stutters were happening on win10. So I don't think it's a win 11 issue. I wonder if oem PCs and laptops with amd have the same issue. AFAIK Microsoft demands that they have tpm enabled. Unless they all use discreet tpms.

[CatNis]

Exactly. This is a TPM issue, not an Windows 11 one.

1 hour ago, NimjaIV said:

I'm not sure I said this before, but my stutters were happening on win10. So I don't think it's a win 11 issue. I wonder if oem PCs and laptops with amd have the same issue. AFAIK Microsoft demands that they have tpm enabled. Unless they all use discreet tpms.

 

[DRenars]

I think we established a long time ago that this issue affects Win10 as well.

This is still happening to me but somehow performing a fTPM clear in the bios made it happen less often for me. Still definitely not a fix. I don't think I'll ever go AMD again in my life. Oh and nothing appears in the event log at all.

[dpeter]

25 minutes ago, DRenars said:

I think we established a long time ago that this issue affects Win10 as well.

This is still happening to me but somehow performing a fTPM clear in the bios made it happen less often for me. Still definitely not a fix. I don't think I'll ever go AMD again in my life. Oh and nothing appears in the event log at all.

I mean, anyone tried it with Linux? It could be a windows 10 problem that passed on to 11, but since 11 enforces tpm most people are only discovering it now. To really get to the bottom of this more testing need to be done. The only thing that we truly know so far is that is caused by turning ftpm on and that's it.

[ANGGARAMDAN]

7 hours ago, dpeter said:

Microsoft released a new Insider build(Beta and Release preview) that includes the L3 cache fix for Ryzen, maybe that solves it, im not currently on windows 11 due to these bugs but if anyone is willing to try...

i'm on this windows beta with the lattest update, still not fix ftpm stutter,.

[Battlesheep]

10 hours ago, dpeter said:

I mean, anyone tried it with Linux? It could be a windows 10 problem that passed on to 11, but since 11 enforces tpm most people are only discovering it now. To really get to the bottom of this more testing need to be done. The only thing that we truly know so far is that is caused by turning ftpm on and that's it.

I am frequently using Ubuntu while fTPM is on, no stutters so far.

 

To be fair, updating to the latest bios completely eliminated stutters so far for me even on Win10. I'll let you know here if any of the systems experience stutter once it happens.

[Zips]

9 hours ago, Battlesheep said:

I am frequently using Ubuntu while fTPM is on, no stutters so far.

 

To be fair, updating to the latest bios completely eliminated stutters so far for me even on Win10. I'll let you know here if any of the systems experience stutter once it happens.

You're on the beta update from Gigabyte?

 

That's reassuring. Hopefully it holds true for the Aorus Elite as well.

[fernandob1994]

Tudo indica que a correção da guagueira esta vindo no Update '' kb5006746 '' , conto com isso .

[DreiNeunVierEins]

Just an update. Have been using the discrete TPM module for a few days, and no more stutters, as expected.

[ShockSA]

On 10/17/2021 at 7:25 PM, Zips said:

You're on the beta update from Gigabyte?

 

That's reassuring. Hopefully it holds true for the Aorus Elite as well.

using the latest bios with AGESA ComboV2 1.2.0.4 A completely eliminated the stutters for me.

[dpeter]

11 hours ago, ShockSA said:

using the latest bios with AGESA ComboV2 1.2.0.4 A completely eliminated the stutters for me.

So the expectation is that AGESA 1.2.0.4 fixes this? Good, hoping that MSI releases it as soon as possible for my mobo.

[Zips]

On 10/19/2021 at 12:05 PM, DreiNeunVierEins said:

Just an update. Have been using the discrete TPM module for a few days, and no more stutters, as expected.

 

22 hours ago, ShockSA said:

using the latest bios with AGESA ComboV2 1.2.0.4 A completely eliminated the stutters for me.

Very happy to hear about both of these. I'll continue to keep an eye out for any more updates here. Hopefully that bios update leaves beta soon and I'll try it out for myself then as well and report my own experiences when that time comes.

[dpeter]

AMD just released new chipset Drivers to fix the preferred core issue in Windows 11, anyone tried it yet? Maybe that fixes the issue.

[Battlesheep]

On 10/17/2021 at 6:25 PM, Zips said:

You're on the beta update from Gigabyte?

 

That's reassuring. Hopefully it holds true for the Aorus Elite as well.

No, I meant the regular, non-beta update.

[Zips]

1 hour ago, Battlesheep said:

No, I meant the regular, non-beta update.

Well that's weird then. I'm on the latest official release for my board, F35, and it has the stutters. 

 

4 hours ago, dpeter said:

AMD just released new chipset Drivers to fix the preferred core issue in Windows 11, anyone tried it yet? Maybe that fixes the issue.

This stutter issue impacts W10 as well. It's not restricted to W11. The new drivers from AMD are to address a preferred core issue that is only found in W11.