Processors
Intel® Processors, Tools, and Utilities
14395 Discussions

11th gen tiger lake undervolting disabled by intel

Quooki
Novice
17,398 Views
Just asking if anyone has any insight as to why intel has disabled undervolting on the 11th gen chips themselves instead of getting OEM's to disable it in the bios as they did for the 10th gen chips. I've got this info from 2 different OEM's senior support staff so I hope I havnt been lead on. At least when it was in the hands of the OEM's there was a tiny bit of choice by using a work around and there was a way of contacting OEM support staff. Now that its on the hardware side we have no control at all over the processor anymore.

It seems to me like they are trying to maximise profits in the mobile chip space by capping the performance. If we arent able to undervolt and tune our performance to get the most out of our cpu, then a higher end (more expensive) H series chip will be more tempting and will end up selling more. It seems like this is contributing to why they are dragging their feet on the plundervolt patch and being so cautious when its only a few niche programmes that cause your device to be vulnerable. And that's if they can gain physical access to the laptop. I just feel like we should have the choice, and of course take on the liability that comes with making the laptop vulnerable to plundervolt.
Labels (1)
19 Replies
AlHill
Super User
17,392 Views

Undervolting is disabled because it is a security risk, pure and simple.

Doc (not an Intel employee or contractor)

0 Kudos
Quooki
Novice
17,371 Views
But it isnt a security risk to everyone running these cpus. If they are trying to cover themselves from legal action I dont see why they cant give the option to people and ensure the user takes on the liability.
n_scott_pearson
Super User
17,388 Views

Intel takes security very seriously and all reported vulnerabilities must be mitigated. For the Plundervolt issue, Intel was unable to identify any mitigation other than having the microcode disable the undervolting capability. While this result was undesirable, it's what they decided to deliver.

It has been argued that the undervolting feature would be completely unnecessary if a proper and adequately-capable cooling solution was provided. Unfortunately, the slimeball laptop providers regularly ship inadequate cooling solutions.

Off my soapbox now...

...S

Quooki
Novice
17,374 Views
Completely agree that a lot of manufacturers really aren't doing their best but we've got to work with what we got. The security risk is not as serious for a home user such as myself, where someone will be unable to gain physical access to my device as a mounted attack is more uncommon and whats more, I dont use SGX. Either way do you know why intel stopped only getting manufacturers to disable it bios and is now pushing the microcode as well? I dont see why it has escalated from the previous generation unless they are just capping performance with the excuse of security.
n_scott_pearson
Super User
17,356 Views

RE: "...we've got to work with what we got"

--> OMG! NOT the right answer. Unless you push back, they will continue to take advantage of everyone.

Look, the situation is what it is. It has nothing to do with the performance of the processors (undervolting lowers performance); it has to do with the security of the system. I also saw the other responses that you posted and then deleted (you do realize that anyone who has enabled notifications receives these, right?); there is no way to support the user taking on the risk that cannot be spoofed, so no, that isn't a possible alternative. 

...S

Quooki
Novice
17,339 Views
I never deleted any of my replies as far as I'm aware, my bad if I did by accident. There arent really any options out there that one can choose in order to push back. All the design and pricing constraints paired with the desire of the user and where they value the product can never really come together into a perfect package. Even the best options will always have a downside, power and thin and light form factor have never got along.
Undervolting on my previous generations of thermally limited laptops (i.e. almost all of them due to the limited ultra book form factor) has boosted performance in synthetic and real world bench marks for me. Its taken away the problem of thermal limitations the resultant downclocking to reduce the heat and allow for the power limit to be the decider of how long my cpu stays at boost clock speeds. Also i find it is very nice to not have the deck or bottom get quite so hot when not doing intensive tasks. I guess I'll just LM the machine after my warranty is up and see if that can make me happier.
Also its really not beyond intel to cap the performance of one chip in order to sell more of another, they have been doing it with their locked chips for years.
TabalugaDragon
17,079 Views

"(undervolting lowers performance)" the heck are you talking about? undervolting INCREASES performance of the CPU due to

1. It working more efficiently due to removal of excess voltage

2. Power limits\thermal limits on laptop CPU-s. 

I got 450 Mhz increase on my i7 8750h after undervolting while keeping it at the same 45 watt TDP, and as a result faster rendering times, and in games now I get stable max 3.9 Ghz all core adversized boost. Undervolting was the Holy Grail of gaming laptops, and Intel took it away from us.

0 Kudos
Mr_Fox
Novice
15,412 Views
@n_scott_pearson  wrote:
It has been argued that the undervolting feature would be completely unnecessary if a proper and adequately-capable cooling solution was provided. Unfortunately, the slimeball laptop providers regularly ship inadequate cooling solutions all the time. 

@n_scott_pearson is absolutely correct. The thermal management of modern laptops is pathetic and inadequate. It is not Intel's fault that they are poorly engineered by people that do not care if they function correctly. Undervolting to compensate for shoddy design is simply a workaround, and the fact that it is necessary is totally unacceptable.


That said, I think the decision to have voltage control should reside in the hands of the computer owner. It absolutely should not be blocked. That makes as much sense as your next door neighbor bricking in the doorway of your residence because it is a security risk (potential point of entry). A proper solution is for the owner to purchase a deadbolt and use it. It should be expected that the firmware will have a menu option to enable/disable voltage control. If I own the computer, that should be my decision exclusively. It should not be pre-decided for me and nobody should have the right to decide what someone else can do with products they own. The default option can be disabled to handle the consumers with no technical abilities or no concern. 

Undervolting does decrease performance. The circumstance under which it improves performance is when it avoids thermal throttling. If you want more performance than a stock processor delivers, overclocking is how you do that. Unless you are only bumping the clock speed by a very conservative amount, you will need to increase voltage. The amount of clock speed increase that is possible running default voltage will vary by silicon quality. Unless you're a winner in the lottery, it's not going to be an impressive overclock using default voltage.

0 Kudos
TabalugaDragon
15,361 Views

Completely agreed with your first 2 paragraphs, but your third one is utter nonsense. Undervolting increases performance, not decreases it. Removal of excess voltage makes a CPU work more efficiently, therefore minimizing the effect of power limit or thermal throttling. True, it doesn't always increase performance(in cases when a CPU already works at max power limit and a laptop is perfectly cooled, which is extremely rare), but it never decreases it either. 

I got +400Mhz on my i7 8750h by undervolting by -150mv, from 3.1 Ghz I get 3.5 Ghz on max load now at the same 45 watt TDP, and got like 10% faster video rendering times in Handbrake and consistently get better Cinebench result. Now you don't have to take my word on that, simply check JarrodTech channel for example, he always undervolts his laptop CPU-s when possible, and it's clear how undervolting increases clockspeeds and benchmarks scores as a result, while in gaming it decreases temperatures without sacrificing performance at all. It's a win\win for gaming laptops, which is why it's necessary like you said. 

Mr_Fox
Novice
15,303 Views

@TabalugaDragon wrote:

Completely agreed with your first 2 paragraphs, but your third one is utter nonsense. Undervolting increases performance, not decreases it. Removal of excess voltage makes a CPU work more efficiently, therefore minimizing the effect of power limit or thermal throttling. True, it doesn't always increase performance(in cases when a CPU already works at max power limit and a laptop is perfectly cooled, which is extremely rare), but it never decreases it either. 

I got +400Mhz on my i7 8750h by undervolting by -150mv, from 3.1 Ghz I get 3.5 Ghz on max load now at the same 45 watt TDP, and got like 10% faster video rendering times in Handbrake and consistently get better Cinebench result. Now you don't have to take my word on that, simply check JarrodTech channel for example, he always undervolts his laptop CPU-s when possible, and it's clear how undervolting increases clockspeeds and benchmarks scores as a result, while in gaming it decreases temperatures without sacrificing performance at all. It's a win\win for gaming laptops, which is why it's necessary like you said. 


It's not a matter of taking anyone's word for it. The difference is a matter of perspective. On one hand you have an avid overclocker and on the other hand you have a thin and light turdbook jockey that is tickled to be able to have some means of avoiding a thermal throttling volcanic abortion. 

The fact of undervolting improving performance is only accurate in the context of applying a workaround solution to minimize malfunction on a poorly engineered and defective product. You are improving results by miminizing malfunction in one area by inducing intentional malfunction in another area. The proverbial robbing Peter to pay Paul. Had it been made correctly, doing that would not be necessary and the performance would be greater than you are capable of realizing through undervolting. More voltage allows for increased wattage, which allows for higher clock speeds, which increases the number of instructions that can be executed, and that results in measurable improvement in performance. If it were not a crippled BGA CPU with locked multipliers you could go even further by increasing the voltage and the multiplier by more than 4 clock bins.

In the case of most turdbooks, doing something that allows the CPU to function as intended by Intel, or closer to it, is a real accomplishment. It's unfortunate that shoddy manfacturing requires it. The outcome would be much better if it did not require undervolting. Good versus great, or better versus ideal. Great and ideal aren't happening on the garbage the notebook manufacturers are peddling. The kiddos think the anorexic look is "sexy" and the people building them don't care how it turns out as long as the kiddos keep drinking the Kool-Aid.

0 Kudos
Mr_Fox
Novice
15,343 Views

It's not a matter of taking anyone's word for it. The difference is a matter of perspective. On one hand you have an avid overclocker and on the other hand you have a thin and light jokebook jockey that is tickled to be able to have some means of avoiding a thermal throttling volcanic mess. 

The fact of undervolting improving performance is only accurate in the context of applying a workaround solution to minimize malfunction on a poorly engineered and defective product. You are improving results by miminizing malfunction in one area by inducing intentional malfunction in another area. The proverbial robbing Peter to pay Paul. Had it been made correctly, doing that would not be necessary and the performance would be greater than you are capable of realizing through undervolting. More voltage allows for increased wattage, which allows for higher clock speeds, which increases the number of instructions that can be executed, and that results in measurable improvement in performance. If it were not a crippled BGA CPU with locked multipliers you could go even further by increasing the voltage and the multiplier by more than 4 clock bins.

In the case of most jokebooks, doing something that allows the CPU to function in a manner that loosely resembles what was originally intended by Intel is a real accomplishment. It's unfortunate that shoddy manfacturing requires it. The outcome would be much better if it did not require undervolting. That good versus great concept. Great just isn't happening on the garbage the notebook manufacturers are peddling. The kiddos think the anorexic look is cute and the people building them don't care how it turns out as long as the kiddos keep drinking the Kool-Aid.

0 Kudos
TabalugaDragon
15,327 Views

The conditions they run at are inadequate, ok, so then what do you suggest exactly? For Intel to drastically decrease max clockspeeds of their CPU-s so they'd run accordingly to their cooling solutions(which would noticeably decrease their gaming performance) or for laptop manufacturers to break laws of physics? Because there is a trend for thinner and lighter laptops and it's not going anywhere, as most people, including me prefer something that is easy to carry around(I however prefer a reasonable balance, such as Legion 5 Pro offers for example). Want a desktop replacement in a laptop for-factor? There are still plenty of super thick and heavy Clevo laptops and such which can give you that, often even coming with desktop CPU-s in them. 

0 Kudos
Mr_Fox
Novice
15,317 Views

That is not accurate. There is only one thick and heavy option with a desktop CPU, which is the Clevo X170. So, yeah, plenty of options if you want that one. That sole option that remains has been compromised by the same flawed perspective. It is thinner and lighter than it should be for best results. So, there is one option left that isn't ideal. That is a far cry from plenty.

There needs to be a better balance. The irrational race to have the thinnest and lightest "gaming" junkbook, versus a decent product that is not focused on form factor shrinkage to the point of being ridiculous has brought us to a place where everything is garbage. I do not entirely fault the OEMs. They could certainly do much better with thermal solutions and stop being control freaks with firmware, but the problem is driven by OCD consumers that get hung up on size and weight to the point of absurdity, expect to pay less for it, but don't hesitate to whine about problems. That has effectively ruined everything for the minority that is not a ridiculous lot with unreasonable expectations.

 

I think we are getting off track and missing the point. The point of this thread is consumers getting locked out of making voltage changes to circumvent defective engineering.

Voltage is like the fuel mixture on your car. If you set it too lean it will malfunction and performance will degrade. If you set it too rich the same thing will occur. To say "undervolting improves performance" is wrong and misleading unless you understand that "undervolting" means setting the correct voltage because the losers that built it didn't care if it functioned correctly. In the context of the voltage being incorrect to begin with, undervolting does improve performance. But, what you are doing is tuning and applying the correct voltage. If the correct voltage is not applied, undervolting will degrade performance and stability. Being blocked from doing that is a sin and a crime, and it should not be tolerated.

A secondary problem occurs when the turdbook's thermal solution does not allow setting the correct voltage because overheating will occur due to the fact it was poorly engineered. Undervolting to accommodate a poorly engineered product should not be viewed as a good solution. The good solution is to not purchase defective garbage.  If you did, there will be consequences and compromises due to poor judgment. Having an outsider arbitrarily intervene with an OS update without permission and break your ability to apply workarounds is totally unacceptable.

0 Kudos
TabalugaDragon
15,286 Views

"has brought us to a place where everything is garbage. " how is Legion 5 garbage? How is Asus Rog Strix G 17 with it's amazing temps in gaming a garbage laptop? They offer great balance in performance and thickness without trying to be too thin.

 

"If you set it too lean it will malfunction and performance will degrade." nope, only stability, performance by itself doesn't degrade with undervolting, it either causes blue screens or just works, no other option. So no, it doesn't decrease performance. I suggest you to study this topic and actually see undervolting performance results before being so self-assured about what it is without ever actually trying it yourself. 

 

"To say "undervolting improves performance" is wrong and misleading" so raising clockspeeds\improving rendering times and so on is not an increase in performance? okaaaay

 

"because the losers that built it didn't care if it functioned correctly."

all CPU-s are manufactured with different excess voltage, no CPU is 100% the same, which means they can't come with lowest voltage possible out of the box, which is why almost all of them have some excess voltage. 

And if you include an auto-undervolting tool out of the box you'll get an inconsistent product, where one laptops(let's say a review sample) would perform noticeably better than the other(consumer version). And nobody wants this lottery. 

 

"If the correct voltage is applied, undervolting will degrade performance and stability." how will it degrade performance again if efficiency improves and clockspeeds rise? and your point makes no sense, if stability suffers then you applied INcorrect voltage and need to raise it a bit. 

 

 

 

0 Kudos
Mr_Fox
Novice
15,274 Views

@TabalugaDragon - I think we are getting off track and missing the point. The point of this thread is consumers getting locked out of making voltage changes to circumvent defective engineering. 

Our interest are at odds. Your intent seems to be arguing, and doing so from the frame of reference that requires defending compromised products. If your Legion 5 hasn't become a voltage-locked and severely malfunctioning piece of trash due to the draconian practice of blocking user intervention in the name of security, then that is very good for you and you have a lot to be thankful for. If the same is true about the G 17, then good for those than own one. Many are not so lucky. So, maybe there are 2 out of hundreds that are not as defective as the rest because they haven't been locked down (yet) and they are not designed as poorly as the majority of competing products. It is always nice to find exceptions. That has nothing to do with the basis for this thread or the practice of blocking voltage contral that is worthy of condemnation. 

You make assumptions from your limited frame of reference, having no knowledge of the person you are attempting to argue with, so there is no point in continuing the discussion. Perhaps we should end it with an agreement that the practice of blocking voltage control is totally unacceptable and move on. This post will mark the end of my engagement in a pointless debate that is driven off-topic by a niche phenomenon.

"If the correct voltage is NOT applied, undervolting will degrade performance and stability." This typo has been corrected. It is a matter of fact.

"I suggest you to study this topic and actually see undervolting performance results before being so self-assured about what it is without ever actually trying it yourself." Erroneous assumptions about a stranger's experience do not serve you well. Jumping to conclusions based on limited experience from the frame of reference of mitigating engineering defects just makes a person look silly. 

Few people have done as much performance tuning over the past 15+ years, but I do not limit my frame of reference to niche products built with severe limitations and engineering defects that require end-user mitigation. The sole focus of my passion is tuning to achieve the best overall result, but I do not devote much effort to products that are compromised toys. https://hwbot.org/user/mr._fox/  A person that doesn't know what they are doing does not make it into the world's top 3% of performance tuning by chance. More than a decade of the journey has involved the use of laptops exclusively, but modern laptops are no longer capable of supporting that passion. That journey does not include following popular YouTube creators whose content is focused on low-cost portable gaming devices. There's nothing wrong with that, it's just not something that interests me.

Profile.JPG  League_Career_Rank.jpg

0 Kudos
TabalugaDragon
15,270 Views

I am arguing because you said what is completely untrue "undervolting decreases performance". Just because you've been in this field for a long time(so have I by the way, not on Intel site though) doesn't mean that you're right about everything. You've yet you show a single example of undervolting decreasing performance. While I can send you like a hundred links of performance improvement after undervolting was done, and among those hundreds of proofs I've seen I've yet to see a single one where performance got worse after undervolting.

 

And no, neither Legion 5 or Strix G17 offer undervolting, they have Ryzen CPU-s in them, I was talking about excellent temps out of the box in these laptops. 

0 Kudos
AlHill
Super User
15,267 Views

Gentlemen, you both have beaten this to death.   It is time to step back.  Or, you can take this argument to another forum, where those members may be interested in such a discussion.

This is a technical support forum.  If you have a technical issue, please post it.

 

Doc (not an Intel employee or contractor)
[Windows 11 is the new Vista]

Mr_Fox
Novice
15,261 Views

Agree. I will not post anything else and moving on. Take care.

I saw your post after my last one.

0 Kudos
Mr_Fox
Novice
15,263 Views

@TabalugaDragon wrote:

I am arguing because you said what is completely untrue "undervolting decreases performance". Just because you've been in this field for a long time(so have I by the way, not on Intel site though) doesn't mean that you're right about everything. AGREED AND NOR IS IT CLAIMED TO BE THE CASE. THE PRESENTATION MERELY ADDRESSES THE ERRONEOUS ASSUMPTION OF IGNORANCE THAT WAS PRESENTED. You've yet you show a single example of undervolting decreasing performance. While I can send you like a hundred links of performance improvement after undervolting was done, and among those hundreds of proofs I've seen I've yet to see a single one where performance got worse after undervolting. ALL RESULTS ACHIEVED WITH COMPROMISED PRODUCTS, WITH THE SOLE PURPOSE OF THAT EXERCISE BEING TO MITIGATE THERMAL PROBLEMS. THE NATURAL END RESULT IS THAT PERFORMANCE IMPROVES. THE MALFUNCTION IS REDUCED. THIS IS TRUTH FROM A LIMITED PERSPECTIVE. NO PROOF IS REQUIRED FOR EITHER POSITION AND OFFERING IT WOULD SIMPLY WASTE TIME FOR THE PROVIDER AND EVERY PERSON VIEWING IT.

 

And no, neither Legion 5 or Strix G17 offer undervolting, they have Ryzen CPU-s in them, I was talking about excellent temps out of the box in these laptops. I DID NOT KNOW THAT BECAUSE THEY ARE NOT IMPORTANT TO ME PERSONALLY. SO, THEY ARE AMD POWERED, WHICH MEANS THEY ARE NOT RELEVANT TO THE THREAD TOPIC (LOCKED VOLTAGE CONTROL) AND THE PROBLEMS THAT IT CAUSES. WHAT THEY DO ILLUSTRATE IS THAT AN OEM ACTUALLY CAN TAKE PROPER STEPS TO ENGINEER A NOTEBOOK CORRECTLY IF THEY WANT TO. THAT IS OUTSIDE OF THE NORM. THEY SELDOM EVER DO.


Note: the use of all capital letters is not "yelling" but merely used to identify the response.

0 Kudos
Reply