AMD to Fix Ryzen 8000G Desktop APU STAPM "Feature" via Motherboard BIOS Updates

[btarunr]

Skin temperature-aware power management (STAPM), is a 2014 feature introduced by AMD for its mobile processors that gets the on-die power management logic to take into account not just the processor's own temperatures (measured via on-chip thermal diodes); but also the physical surface temperature of the laptop itself, by reading off temperature probes mounted on the laptop chassis. This ensures that laptops don't get uncomfortably hot for the user, and the processor could do its bit to bring temperatures down. Every desktop APU released by AMD since 2014 has been a case of mobile processor silicon being adapted for the desktop platform by simply disabling certain I/O interfaces and features irrelevant to desktops, such as battery management GPIO, LPDDR memory interfaces, image processing, sensor suite, etc. One such feature is STAPM.

Gamers Nexus discovered that when creating the Ryzen 8000G desktop APUs, AMD forgot to properly disable STAPM, and this has been impacting the processor's CPU and iGPU boosting behavior under heavy load, where temperature-triggered clock speed throttling is engaged undesirably. AMD confirmed the Gamers Nexus discovery, and stated that it can be fixed through a motherboard UEFI firmware (BIOS) update; and that it will work with its desktop motherboard partners to get these out. The highest performance delta observed by GN between an 8000G processor with STAPM and one with its STAPM disabled (probably using an AMD CBS setting); is 16%; and so those with 8000G processors may want to look out for firmware updates from their motherboard vendors.



View at TechPowerUp Main Site | Source

 

[_Flare]

 

[LabRat 891]

It's neat that this feature is there. OTOH, It's not cool that it's on by default for products that aren't intended for Handhelds or Laptops.

From a 'normie' perspective, one might think their dev teams are overwhelmed/overworked.
'simple' default toggles are being overlooked...

 

[Exilarch]

Thanks, Steve!

 

[catulitechup]

@W1zzard

For this reason and maybe others dont appear in a house reviews?

 

[Ferrum Master]

It's neat that this feature is there. OTOH, It's not cool that it's on by default for products that aren't intended for Handhelds or Laptops.

From a 'normie' perspective, one might think their dev teams are overwhelmed/overworked.
'simple' default toggles are being overlooked...

The R/D budget being few times smaller than Intel, has some drawbacks.

 

[Assimilator]

Once again AMD showing that they don't test for shit.

The R/D budget being few times smaller than Intel, has some drawbacks.

That's no excuse when the bug is this obvious.

 

[L'Eliminateur]

AMD firmware slipups on AM5 have been worrisome from a QA standpoint from the cooking cpu fiasco to this staaahhp'em

 

[kapone32]

AMD firmware slipups on AM5 have been worrisome from a QA standpoint from the cooking cpu fiasco to this staaahhp'em

The Cooking CPU was not AMD

Once again AMD showing that they don't test for shit.


That's no excuse when the bug is this obvious.

If anything it proves how stretched AMD are with what they are producing and able to test. It does seem like a major gaffe to forget to set those APUs to Desktop mode though. That is essentially what this sounds like. It should make the 8600G even more attractive. For me the 8700G is too much for what it is. I would love to know who makes the board for those handhelds so I could make my own. The thing is Mini ITX, even with a 8600G is too expensive for not just the board but case as well.

 

[Squared]

This sounds like a pretty easy thing to fix so my guess is that this is less about a small dev team and more about what kind of checks are in place. There ought to be some sort of development checklist that mentions this, and then the QA process ought to find it if the checklist item is missed.

 

[L'Eliminateur]

The Cooking CPU was not AMD


If anything it proves how stretched AMD are with what they are producing and able to test. It does seem like a major gaffe to forget to set those APUs to Desktop mode though. That is essentially what this sounds like. It should make the 8600G even more attractive. For me the 8700G is too much for what it is. I would love to know who makes the board for those handhelds so I could make my own. The thing is Mini ITX, even with a 8600G is too expensive for not just the board but case as well.

the cooking cpu was AMD and their dodgy bios with hardly any test.

Even if an OEM fumbles their bios, it should be limited to that OEM not to ALL of them, if it affects everyone then the issue is upsteram

 

[Wirko]

PC surface temperature regulation is for cats. They will demand it back.

 

[kapone32]

the cooking cpu was AMD and their dodgy bios with hardly any test.

Even if an OEM fumbles their bios, it should be limited to that OEM not to ALL of them, if it affects everyone then the issue is upsteram

I have been on AM5 since start and it never effected me. The burning was from the MB giving the Memory controller too much voltage. Of course Gigabyte and Asus were the worst as they turn their voltages up anyway but i digress.

 

[Event Horizon]

I like it when reviewers are thorough and critical of a product. We owe these companies nothing.

 

[Nostras]

Interestingly it seems not everything was affected. When asked our local testers reported no such issue on the MSI boards (tweakers).

 

[67Elco]

Great...I get to flash the bios on my new board...again.

 

[evernessince]

the cooking cpu was AMD and their dodgy bios with hardly any test.

Even if an OEM fumbles their bios, it should be limited to that OEM not to ALL of them, if it affects everyone then the issue is upsteram

This is just straight up incorrect. As GN has stated in it's analysis of the issue, it was hard to reproduce and required that multiple things go wrong in order to trigger it. Specifically that you needed a motherboard that exceeded safe SOC limits, that you need to be unlucky and have both low quality RAM and a low quality memory controller on the CPU, and that you needed to have a motherboard that didn't trigger OCP protections.

There's nothing stopping you from degrading or destroying your CPU on other Intel or AMD platforms by pumping up the memory controller voltage, the reason it was an issue here is because ASUS forgot to implement basic safety features and was pumping an unsafe amount of voltage into the SOC simply when enabling a memory profile. This is why the only verified accounts of it happening were on ASUS motherboards, they screwed up big time. There was only one other occurrence of it on a Gigabyte board but that example was never verified. Not that there was a very high total incident rate anyways, there were only some 5 examples in total.


To say that the issue was entirely AMD's fault or that it's even remotely comparable to the example in the article is wholly misrepresenting the facts.

 

[Dr. Dro]

AMD PLEASE stop releasing products with broken BIOS, firmware, power management, memory training, etc. just PLEASE do this

To say that the issue was entirely AMD's fault or that it's even remotely comparable to the example in the article is wholly misrepresenting the facts.

An obscure bug is still a bug and it is their responsibility if it occurs in default settings and was known and entirely avoidable. AGESA bugs that were potentially deadly to the hardware, as well as all the issues and problems that these platforms have had over time, are entirely AMD's fault. Although, I'm willing to be fair: if a fix is made available as quickly as the problem is identified, I'm willing to accept that. It just can't become habit, which it has.

 

[67Elco]

If that be the case then I will leave my MSI motherboard alone unless I run across any problems myself.

 

[kapone32]

AMD PLEASE stop releasing products with broken BIOS, firmware, power management, memory training, etc. just PLEASE do this



An obscure bug is still a bug and it is their responsibility if it occurs in default settings and was known and entirely avoidable. AGESA bugs that were potentially deadly to the hardware, as well as all the issues and problems that these platforms have had over time, are entirely AMD's fault. Although, I'm willing to be fair: if a fix is made available as quickly as the problem is identified, I'm willing to accept that. It just can't become habit, which it has.

You seem to think that every time there is an issue it is entirely AMD's fault. I have never seen an AGESA update that gave me problems. X370 may have been a little wonky as it was a brand new chipset. By the time X470 or B550 came AMD was rock solid. Your statement could be taken a little hyperbolic. In fact people complained about no DDR 3200 support when they bought Corsair RAM and when they looked online they learned that it was not AMD but Corsair. I should know I was one of them. Now we have AM5 and most of the QOL issues like long boot times have been long solved. They forgot to turn off laptop protection for this and though it was an error it does not stop the PC from working and it will be fixed. It is not like this has never happened to the others.

 

[TechLurker]

Free performance is free performance. I'd like to see what the end results will be after some retesting.

 

[Dr. Dro]

You seem to think that every time there is an issue it is entirely AMD's fault. I have never seen an AGESA update that gave me problems. X370 may have been a little wonky as it was a brand new chipset. By the time X470 or B550 came AMD was rock solid. Your statement could be taken a little hyperbolic. In fact people complained about no DDR 3200 support when they bought Corsair RAM and when they looked online they learned that it was not AMD but Corsair. I should know I was one of them. Now we have AM5 and most of the QOL issues like long boot times have been long solved. They forgot to turn off laptop protection for this and though it was an error it does not stop the PC from working and it will be fixed. It is not like this has never happened to the others.

With AM4, it was AMD's fault 99.5% of the time and issues were universal across all motherboards until AGESA got updated (which also requires a BIOS update and that is on the motherboard manufacturers). It's not hyperbolic, and the same repeated with AM5 thus far. Once their AGESA open source replacement rolls out and developers have more agency over it, the situation should improve.

 

[kapone32]

With AM4, it was AMD's fault 99.5% of the time and issues were universal across all motherboards until AGESA got updated (which also requires a BIOS update and that is on the motherboard manufacturers). It's not hyperbolic, and the same repeated with AM5 thus far. Once their AGESA open source replacement rolls out and developers have more agency over it, the situation should improve.

LMAO, you are so pedantic that you make a claim that an AGESA update cannot be applied by updating the chipset drivers from AMD's site. I know you think AMD is garbage and the only reason I am responding to you is that I have been drinking and today was a great day in the Premier league. 99% of what issues? I have built more than 40 AM4 systems with no complaints so try again. With that I have tried every single AM4 CPU and they all work. Now you want to attack AM5? When you feel how snappy AM5 is with PCie 5.0 drives you might understand. Oh wait AM5 is not a good platform. When you see the 9000 CPU reviews and see it is faster than 7000 you will know why it is academic to get into AM5 if you are into computing.

It is really sad that TPU (for some users) is no different than Reddit.

 

[Dr. Dro]

LMAO, you are so pedantic that you make a claim that an AGESA update cannot be applied by updating the chipset drivers from AMD's site. I know you think AMD is garbage and the only reason I am responding to you is that I have been drinking and today was a great day in the Premier league. 99% of what issues? I have built more than 40 AM4 systems with no complaints so try again. With that I have tried every single AM4 CPU and they all work. Now you want to attack AM5? When you feel how snappy AM5 is with PCie 5.0 drives you might understand. Oh wait AM5 is not a good platform. When you see the 9000 CPU reviews and see it is faster than 7000 you will know why it is academic to get into AM5 if you are into computing.

It is really sad that TPU (for some users) is no different than Reddit.

Because it cannot: AGESA are not chipset drivers and the CPU requires this low-level code to boot as it is responsible for a large amount of essential functionality including processor initialization. So if you want to contribute to the quality of TPU, stop defending AMD at every turn as if you have a romantic attachment to them and learn your trade:

AGESA - Wikipedia

en.wikipedia.org