BLACK LIST. Adjustable voltage control/turbo ratio limits are locked out due latest Win Update/Bios

Yeah, you're right. Further debate is pointless. And, the laptops affect were already compromised garbage prior to this, so nothing has really change other than their pre-existing defects are merely amplified and more glaring than before. I feel bad for the suckers that wasted their money on something that was broken before it was boxed at the factory, and for the people that applied firmware updates in the vacuum based on the foolhardy notion that newer automatically means better. Even seemingly intelligent people make that mistake sometimes. They genuinely deserve our sympathy and condolences for their misfortune is in order.

 

Basically our Clevos remain untouched, for now? Phew.

 

Looks like MS updates are getting rid of the undervolt on the Gigabyte Aer 15 XA Classic with an INTEL microcode update (yes from Intel).

 

Umar has finally replied to my tweet regarding undervolting issues. Here is the link. Surprisingly, he's still unaware of it, even though countless forums mentioned it. But now that his team is looking into it, I hope he'll take care of all this, sooner or later.

https://mobile.twitter.com/aw_umar/status/1276541407801606146

 

Don't rejoice too soon yet. Here are the latest updates that are supposed to be rolled out this month.

P9xxED/EF/EN and P9xRD/RF/RN
BIOS 1.07.15
Microcode Update to 0xC6

NB70TH
BIOS 1.07.09
Microcode Update to 0xCA

P7xxTMx-(G)
BIOS 1.07.29
Microcode Update to 0xCA

I don't know if those updates have anything to do Undervolting; I don't have more detailed changelogs.
It will take a few weeks before we start testing these update on P9xx and P7xx series (We don't carry NB70TH).


Info about Microcode Update via Windows Update

The main reason why you have been relatively safe from getting automatic Microcode updates via Windows Update is a lack on WU CHID management in this ODM/LOEM Whitelabel eco-system.
Microsoft can only target drivers and firmware updates on your system if it has permission from the system vendor. Your system can only be identified if certain SMBIOS DMI strings are known.

CHID = Computer Hardware ID

Example:
{22c8188f-30b7-559b-afd2-b77ca7b21d1c} <- Manufacturer + Baseboard Manufacturer + Baseboard Product

The "Manufacturer" string is the one that is read as "System Manufacturer" in HWiNFO64.
Microsoft's CHID management does not allow any CHID idendities that does not include this string.

Consequently, if you have a custom Manufacturer string on your system (from your seller) and if this string has *not* been submitted to the ODM or to Microsoft for WU CHID management, Microsoft won't be able to target your system with platform-dependent drivers and firmware through Windows Update.

At least, this is according to my only moderately well-informed understanding. Exceptions might apply.

By the way, those strings can be easily overwritten by end-users with recent Clevo system with Insyde BIOS if you have access to the EFI Tool called "H2OSDE-Sx64.efi". Typical parameters are:
-SM = System Manufacturer
-SP = System Product
-BM = Baseboard Manufacturer
-SS = System Serial Number
-CA = Chassis Asset Tag

Code Example:
H2OSDE-Sx64.efi -SM "Example Company" -SP "Example Product" -SS 123456789

Tools for other BIOS platforms exist with different names but similar structure.
But beware, some of the strings can be overwritten again by future BIOS updates, depending on system vendor and model. In such cases, you would need to run the DMI Edit Tool again after the BIOS update to customize your strings.


Undervolting Overview across 2019 and 2020 models

For more information on our Undervolting efforts, check out this thread in our sub-reddit:
https://www.reddit.com/r/XMG_gg/comments/h0wuoh/cpu_undervolting_in_bios_setup_for_xmg_and/

Cheers,
Tom

 

Thank you sir. Could you also provide same info in my other thread? Thanks

BLACK LIST 2. Adjustable voltage control/turbo ratio limits locked out - Affected 10th gen notebooks

 

Intel re-enables undervolting for some reference design laptops

https://www.notebookcheck.net/Intel...r-some-reference-design-laptops.481320.0.html

 

Nice to see NoteboookCheck acknowledging Eluktronics, XMG and Maingear for doing the right thing, as well as indirectly calling out the lame OEMs selling competing products that are broken.

This also supports my previous comments that the decision to ruin laptops rests in the hands of the OEMs, and it is not Intel telling them they are required to sell screwed up trash with the ability to undervolt removed. That mistake is up to the discretion of the people selling them.

 

Intel is the ODM for these laptops, it was Intel who disabled the undervolting initially and then re-enabled it... It's not XMG or Eluktronics, etc... Perhaps intel listened to them though.

 

It's obvious that Intel cares about it enough to not put their name behind broken garbage. It also remains clear to me that the decision to sell broken trash with undervolting abilities disabled rests on the shoulders of the OEM and it's their call whether or not to sell broken trash. What is being said here supports what I have been saying all along.

"While multiple OEMs have disabled undervolting on new and existing machines..."

"Overall, it is encouraging to see that Intel has enabled undervolting with one of its white-label laptops. With that said, it still does not change the fact that OEMs are either removing the functionality from older machines. Additionally, most OEMs have disabled undervolting on laptops released with 10th Gen Core processors after the publication of the Plundervolt exploit." ​

At any rate, people have the opportunity to vote with their wallets and choose to purchase a product made by a company that actually cares about what they sell and the people that buy what they sell.

 

You left out some of the quote:
"While multiple OEMs have disabled undervolting on new and existing machines, Intel has re-enabled the functionality on its Whitebook LAPQC71. Colloquially called the QC71, the laptop is one of Intel's reference design laptops released under multiple names"

 

Exactly. Intel is not going to sell something they developed that is broken. They would be really stupid to do that.

Further, Intel is also not requiring that the other OEMs disable undervolting and ultimately sell broken garbage. The OEMs are making a decision to do that. Some are even going a step further and retroactively breaking products that they already sold by deploying cancerous firmware updates rather than allowing their customers to make their own decisions.

It is a lack of regard for their customers and taking the lazy way out by the OEMs that are electing to carte blanche and arbitrarily disable undervolting. It's easy to take a Procrustean approach to things by releasing a cancer firmware update with no end-user configuration options in it.

 

I won't deny that OEMs are to blame, but so is Intel. Intel writes the bios for these laptops, and Intel was the one who removed the undervolting initially. It probably took push back from the resellers to actually fix this.

 

Yeah, I think you're probably 100% correct about push-back from resellers having something to do with it.

Companies like Eluktronics, Maingear and XMG actually care about products that bear their names and the people that buy them. Some of other less well-known (not mainstream) boutique brands also care equally... EVOC, XOTICPC, Sager/Schenker and other sub-branded Clevos are the obvious examples.

The people that I feel sorry for are the mainstream consumers that buy from HP, Dell, Acer, Lenovo, etc. that are getting shafted royally. They are buying so-called "gaming notebooks" with their man parts removed, (or being later castrated with a cancer BIOS update,) simply because the people that made what they purchased are making bad decisions or just not caring enough to do the right thing.

 

And another one bites the dust..
Asus FX705 series, forced BIOS update from a capsule and undervolt is disabled..
At least they didn't lock anything, flashed my modded BIOS with FPTW and its all back to working, already disabled firmware updates on the BIOS so this crap doesn't happen again.

 

Is this another one for the list? Or, is it some other HP firmware cancer going on?

http://forum.notebookreview.com/threads/hp-omen-15-and-17-owners-lounge.808312/page-31#post-11035637

Damn. ASUS is doing those Satanic laptop BIOS updates through Micro$lop Updates like Dellienware too now? If so, that's just way messed up. There has to be some way of permanently removing or otherwise damaging that feature in the firmware so it cannot ever be re-enabled again. This kind of crap really pisses me off. Should be against the law to push firmware with Windows Updates without explicit knowledge, consent and full disclosure of the outcome to the owner. They should force scrolling through detailed release notes, receiving an SMS text code to enter, then another layer of 2FA and finally have to read the release notes a second time and then do one of those stupid captcha things before a firmware update occurs over the internet.

But, congrats on fixing it.

 

I can report that the Alienware M17 R3 voltages/clock speed is unlocked in the i9-10980HK variant.

 

That's good. Perhaps because the 10980HK has built-in Plundervolt mitigation?

 

They do brag about the voltage regulation on the machine. If I remember correctly, I saw up to about 140W of power throughput on the i9-10980HK processor when power limit was lifted. It couldn't sustain that power throughput because of thermal throttling without extra help from laptop cooling, repaste job I'm sure though. Maybe not even then, I didn't try to push it to the limits at all.

 

I can report the latest BIOS (F.30) on my HP OMEN 17-cb0009nl (11/2019 model, i7-9750, RTX 2070 full) doesn't have the plundervolt update.
I can undervolt just fine with TS version 8.70, like previously.

 

I have 15 inch Omen from 2018 with i7-8750 and undervolting is broken: latest BIOS no go, I went back to BIOS from 2018, I reset the BIOS and I deleted mcupdateintel.dll and still underolting not working. Any other suggestions? Because I'm getting really pissed. I can see microcode is changed to old and it should work, but it doesn't, damn stupid intel and hp.
BTW I hope you guys don't get fooled: on my system only CPU core doesn't work, all other voltage adjustments work fine, including cache and even HWinfo shows offset for core voltage set, except it doesn't sets and even if I set offset to -1 volt computer doesn't crash

 

Pls add to the list:
Dell Precision 7730 - bios 1.13.1 implements CVE-2019-11157

 

Please add Lenovo Yoga 730-15IWL to the list.
The latest BIOS update (version 2.07) has locked CPU voltage control. I cannot return the previous version of BIOS, because Lenovo does not publish archived versions of BIOS on the website for my laptop model.

 

I'm sure it's been discussed but at a very less than desirable method, Asus has implemented an option to undervolt to a max of -80mv in the BIOS of some of their newer ROG laptops, such as my current Asus GU502L.. It's still infuriating because it's just as stable as stock with -80mv and I know there's more I can squeeze out of this i7-10785H to further lower temps and there's no other way to do it.

However, I plan to see if I can dump the BIOS and edit it and find out if there's a way to unlock the BIOS further by disabling SGX or something and or just opening up the undervolt options all together. May be fruitless and I may end up bricking the machine with a re-flash but i'll be damned and determined.

I'm still a fanboy of Intel laptops because of Undervolting all together since there's no easy way or reliable way to undervolt or control TDP with Ryzen stuff on mobile so there's that.

 

It seems you'll never ever get the answers you asked for....


http://forum.notebookreview.com/thr...-owners-lounge.826831/page-2350#post-11053886

 

Something's probably wrong in that AW dugout for sure, seeing many drop outs lately (though Azor was kinda just passionate about getting into AMD).

Also, apart from Alienware, I also tried to talk to one of the guys at ROG Technical Team about removing the -80 mV undervolt limit (on 10th Gen ROG laptops, Asus allows UV limited to -80 mV). He mentioned two things.

1. That -80 mV limit is governed by Intel.

2. When I asked why other manufacturers like XMG and so on impose no restrictions on the UV and others like Asus have so many restrictions, he said:

"Each company has different requests and are able to do slightly different variations under guidance from intel/amd/Nvidia. I’m not sure what that brand's policy is with regards to undervolting but perhaps they are willing to accept the risk of allowing more undervolting. But sure, we’ve been an advocate for allowing users to undervolt and have always allowed this option in bios. The feedback will be forwarded to the relevant groups."

As far as I know, Plundervolt vulnerability is exploitable only through software means. So if UV is allowed only through BIOS (without any restriction) and XTU interface is not exposed to the user, no exploit can work at all. So even if OEMs are so skeptical about UV, are they not intelligent enough to understand such a small observation? And what purpose does that -80 mV serve in the first place? And I'm pretty sure, Intel has no business with that -80 mV limit. That's just a petty excuse IMO to evade questions from customers.

 

Undervolting on Intel Laptops in 2020: this is what you need to know! | XMG Support! Igorslab.com

State of Undervolting in 2020...

Pushback from the community

Naturally, there is some inertia in the industry – so it took months before users of various brands started noticing new Undervolting-blocking firmware updates. Usually those come in the form of an optional BIOS update, but there also have been reports of systems having automatically been patched over Windows Updates. But the biggest push against Undervolting came with the release of a new generation of Intel CPUs. A new generation of high-end mobile chipsets called “Comet Lake H” was announced in April this year. By this time, all OEMs have gotten Intel’s memo and most of them chose to disable Undervolting in new Comet Lake systems from Day 1.

This naturally generated some pushback with outspoken community members voicing their concerns over the lack of choice and ownership. After all, it should be up to the end-users if they prioritize security over performance, especially considering the very niche practical usage scenario of the ‘Plundervolt’ exploit. One of the rallying points of this community response was a thread on Notebookreview, where users assembled a list of reports and system model names which got locked out of Undervolting with these new updates.

• BLACK LIST. Adjustable voltage control/turbo ratio limits are locked out due latest Win Update/Bios [Notebookreview Forum]

In April, we at Schenker also reached out to ODMs and to various contacts in Intel, digging for information and advocating for a more measured approach on this issue. Since then, it looks like the situation has become a little bit more dynamic. We managed to convince our ODM to unlock Undervolting for all of our new Comet Lake based systems (with currently one single exception: XMG ULTRA 17) and we have also seen some other system vendors to follow suite. But when you buy a random Intel-based laptop in 2020, there is no guarantee that Undervolt will work. It will be down to each individual system and BIOS version – buyer’s beware!

It still amaze me that noone from Alienware community has rebelled hard and demanded a response from the Dell support team in their own playyard (dell.com/community). 3 pages filled with the same mess and still not a single reply back from Dell support... 17 R5, BIOS 1.11.0, no undervolting

They avoid the topic completely and never reply back. Most likely learned or been told by those on the top. See my post #124!

@unclewebb @Mr. Fox @Ashtrix @jc_denton @Donald@zTecpc @Fire Tiger @Arrrrbol @c69k +++ Nice seeing Notebookreview Forum also is mentioned

 

A person of ordinary intelligence needs only to pay attention to identify the slogan is false narrative. Dell cares about Dell. What customers want only matters when it is what they want.

Hogwash... poppycock... nonsense... choose your own adjective.

 

@Papusan can you clarify this:

"We managed to convince our ODM to unlock Undervolting for all of our new Comet Lake based systems (with currently one single exception: XMG ULTRA 17)"

Does XMG mean that the single exception wasn't necessary since the Ultra 17 already has a K processor? Or is Intel now setting its sights on K processors as well? Meaning you are blued and tatooed if you buy the Ultra thinking you can take control of voltage.

 

Not sure. Maybe because it use desktop processors. Intel doesn't force the manufacturers to disable Undervolting. And from what I have seen... SGX will be removed from coming Rocket lake if this rumors is true. I mean it's still there in 11th gen Tiger Mobile.

A big advantage for those that prefer performance above security. And Plundervolt isn't what you should be most afraid of.

 

@Papusan Installed my BGA AW BIOS 1.11 which has plundervolt uCode DC which killed undervolt and might have borderline killed my display and possible bricking. Update wasn't successful. Had to use Dell Gset to update BIOS and find display was off and no pixels lit.
Downgraded to 1.8 BIOS.

 

Is there a safe way to rollback to 1.8.0 from 1.11.0? Is it even possible?
If you see in the 15 r1/r2 tread I have some problems during boot/restart.
Have you experienced this @Vasudev ?

 

Tried flashing 1.11 through Windows and it was stuck at black screen after reboot twice and felt like sudden re-seat RAM stick when your laptop tripped, but I didn't get any beeps and messages. So, I shut it off twice and plugged in USB recovery Bios and then I came to know, 1.11 was not flashed properly and I used Dell GSet or Dell Flash BIOS utility which is presented when you hit F12 before startup. Put the EXE file in USB drive 2.0/3.0 with older BIOS 1.8 and you will see a message "Downgrade is not recommended" Select Cancel button to flash to older version and use TAB keys to navigate between options. I reverted it back using that. Dell USB recovery option was blocked.

 

Bump.

15 R4, give us back undervolt dell.com/community/Alienware



Dell pushed a new fix to handle the 100C for the newer models.... TCC Offset.


" TCC offset will reduce maximum CPU temps but may also slightly reduce CPU performance in certain applications". He even put out statements that the new feature will works more like <It's like cruise control for your CPU temps>.

 

I remember people on Alienware subreddit recommend to disable turbo and be happy about it.

 

some are desperate enough and just want a working machine. quite sad really.

 

We can get back undervolt on Dell laptop following this thread : https://brendangreenley.com/undervo...hermals-battery-life-and-speed/#cpu-undervolt
Got my 10875H to -150mV and was able to run an all core 4.3GHz boost during benchmark

 

Hey bright I have a PC50 and I'd like to know: What microcode revision are you on that lets you undervolt?

 

@Papusan Undervolting is locked here on my Mi Notebook 14(alt. Redmibook 14 in other markets) running on a Core i3 10110U.

 

FYI - Undervolting-Options are greyed out since latest BIOS Release on ASUS Zephyrus S GX531GXR.

To get access again you have to roll back to the previous BIOS-Version and before turning your rig on you have to do a complete reinstall of WinGoofs!!!
Otherwise it will update automatically to the newest BIOS-Version.... again! What a crap!

Gesendet von meinem LE2123 mit Tapatalk

 

Hi. Core i3 10110U is 10th gen and belong in the BLACK LIST 2. Could you please post it there as well? Thanks.

Thanks. ASUS Zephyrus S GX531GXR added in OP post.

Isn't it amusing? Up to the OEMs lock you out from Undervolt or not. Often let you have undevolt unlocked it if you paid the higher price for the unlocked SKUs. Does that mean the more expensive laptops is more secure bro @Mr. Fox ?

For 11th gen Tiger lake (n_scott_pearson - former Intel employee). Missed the point what undervolt can do for your Cpu but was correct on the OEMs behalf regarding crippled cooling.
https://community.intel.com/t5/Proc...e-undervolting-disabled-by-intel/td-p/1245398

 

https://community.intel.com/t5/Proc...ed-by-intel/m-p/1300043/highlight/true#M52665

 

Good post although I'm not clear on why undervolting decreases performance, given a particular clock target. All we are doing with UV is saying "run as you would but at a lower voltage", rigtt? If the CPU can run at a particular clock speed, albeit at a lower than default voltage, that's a good thing insofar temps and energy utilisation are concerned, with no downsides.

Perhaps you meant "undervolting does not increase performance, exept through avoidance of thermal throttling"?

 

No, @Mr. Fox is correct. Undervolting does decrease the performance.

The clocks are not just the clocks, the platform needs power to perform.

A properly designed chassis and cooling system is what is sorely needed.

 

What is the exact mechanism behind this loss of performance? Let's take the case of non-overclock able CPUs, to simplify the argument. If such a CPU can run a benchmark such as CB20+ at full clock speed under full load with a -150mV UV, are you saying that removing the UV would lead to further increase in performance, despite the same clock speed (assuming sufficient cooling)?

The platform does need power. Undervolting simply reduces the amount of power needed to perform a given computation within a set period of time, I'd think. The reason why it's not done by the manufacturer is the instability risk. The default voltages are high enough to ensure all CPUs, no matter how bad the silicon, will be stable across what's probably a wider range of conditions than needed by most users.

Better cooling solutions are sorely needed in most laptops for sure, but that's to enable lower temps, higher clock speeds, or higher nominal power CPUs and GPUs.

 

I can tell you that setting the lowest stable voltage almost always lowers my benchmark scores. If firmware is set to allow the CPU to dynamically boost within a temperature-limited scenario, if temperatures are low enough having the voltage lower will limit the maximum boost clock. (This applies to GPUs as well) Lowering voltage also limits how far you can overclock, which in and of itself is a performance limiter.

The main reason laptop owners associate undervolting with performance improvement is specifically due to how poorly made laptops have been for several years and taking drastic measures to control the temperatures helps circumvent malfunction (as outlined in the post over on the Intel forum).

 

@Mr. Fox, most people don't realize the distinction.

Just like you said, lower voltage = less performance. What bothers me is the snappiness that simply vanishes.

 

Eh, for me undervolting increases performance (have hundreds of benchmarks that show this) because the cooling solution sucks on most laptops so it allows me to maintain higher clocks. Again, voltage doesn't have anything to do with performance (otherwise we would be overvolting our desktops and not adjusting clocks), but it does have everything to do with stability. If your undervolt is decreasing your performance, then you are not totally stable.

And yes you need more volts to run at higher clocks, that's a given (again to increase stability, not performance). Most laptops are not overclockable so undervolting will increase performance. Within the given clock ranges that a laptop can run, undervolting will improve performance.

 

You're missing the point.

Given a superior cooling system, undervolting is not needed. And will, in fact, decrease the performance of the platform.

A workaround for garbage designed chassis' is greatly appreciated (million thanks to @unclewebb), but a better designed chassis and cooling system is what's missing from today's powerful systems (and missing for a long time now).

 

Yes, that is spot-on, deadly accurate. It is one of the most misunderstood concepts as well. People draw conclusions based on engineering defects and poorly designed laptops. The harsh reality is, they are dealing with consequences... squeezing lemonade from their lemon.

An important point of clarification in this type of discussion to enhance understanding is the difference between VID and Core Voltage. @Falkentyne is one of our resident experts on this topic, so he can correct me if I state anything wrong.

Many software monitoring tools do not correctly report core voltage and some report only VID. VID basically doesn't matter. It is what the CPU is "requesting" from the motherboard/firmware. Core voltage is the actual applied voltage. Input voltage (which is often not adjustable on laptops) is the total amount of voltage available at the socket to the CPU package. It is generally MUCH higher than VID or core voltage because it takes a lot more voltage to run the CPU package than it does at the core and cache level. If a person doesn't understand this they may struggle with tuning and they may create problems for themselves by losing sleep over a high VID value and being unable to control it to the point that it shows them a number they think they want to see.

For example, for 5.2GHz on all cores I have my core voltage set at 1.325V in the BIOS. With LLC factored in, the final voltage is about 1.359V. This is what the motherboard reports through the onboard LED and it is what HWINFO64 reports for core voltage. HOWEVER, the Core VID is 1.550V and if I look at CPU-Z or Core VID in HWINFO64 it is misleading and alarming if I do not understand what I am looking at. Core VID can be ignored because it is not the voltage the CPU is actually using. It's only the amount of voltage the CPU believes it needs based on the algorithms baked into microcode which are agnostic to silicon quality. Although, to some extent the ASIC may affect the Core VID, it is not relevant to tuning. It's only relevant to what the CPU might be capable of in terms of overclocking. If the turdbook has a locked CPU then none of this matters. You tune it for thermals and the performance is what it is... you'll have to live with it.

This all comes into play with undervolting and overvolting. If it is not understood, a number of mistakes are possible:

• If you look at Core VID you can wrongly assume your CPU is a junk silicon lottery loser
• If you try to adjust based on Core VID (rather than core voltage) you can end up with an unstable system or one that underperforms because the CPU is starved for voltage
• You can lose performance from having voltage too low if you are trying to adjust based on Core VID, as well as have WHEA errors, because you are aiming at the wrong target

It adds insult to injury when OEMs and firmware developers do not distinguish between the two and treat them as synonymous.

At the end of the day you have to tune for performance. What you are able to achieve is ultimately going to be determined by how cool (or hot) it runs and silicon quality. Using the RIGHT voltage is the key to higher performance. Too little degrades performance. Too much degrades silicon, degrades performance and degrades thermals. Owning a turdbook that does not have adjustable voltage is messed up no matter how you want to slice it. It is a crippled piece of crap that is going to run too hot, thermal throttle, potentially degrade and provide an overall miserable user experience. If you tune for performance on a laptop with a lackluster thermal solution (which means almost all of them) you will lose performance (not gain) because the system can't handle it. So, you end up tuning for the lowest common denominator on the turdbook (the weak thermal solution). You may gain performance versus letting it run at firmware defaults. No question there. But, you won't achieve optimal performance because giving the CPU the voltage it needs to perform at its peak results in a performance loss due to thermal throttling. Toss in the TVB wildcard and it's anyone's guess what you're going to end up with. Thermal paste, ambient temperatures, fan speeds, etc. will affect the outcome even if you have the voltage dialed in.