The ThrottleStop Guide

Just a heads up for Asus laptop owners. Some of their laptops like the TUF FX505 have recently developed a bizarre throttling problem. Someone at Asus thought it would be a good idea to lower both of the turbo power limits down to 5W. I am not talking about the 15W low power U series being dropped down to 5W. Asus is lowering their 45W CPUs down to 5W.



Armoury Crate might be involved in this master plan.

If one power limit got changed to 5W, maybe you could walk away thinking that this was some sort of random accident. Oops. Much less likely to be an accident when both power limits get set to the same 5W value. Just enough power so when idle, your CPU might look to be performing sort of normal. As soon as any load is applied, massive power limit throttling soon begins. In 2020, does anyone like gaming when the CPU is running at 800 MHz? Probably not.

Luckily ThrottleStop was able to fix this problem. Default power limits are somewhere around 45W long and 60W short. If you set these values in the TPL window and then also use the Lock option, no more plunging power limit throttling issues. The Lock option prevents bad people from messing with the power limit register.

 

Hello @alfafox, it just happened I read on Twitter yesterday this:
"If you system doesn't have Anti Rollback (ARB) protection (more probably doesn't have) you can downgrade the FW to vulnerable version"
I read link you posted and it's not seem to me like good approach, you should seek some guys on https://www.win-raid.com/, they will help you most probably.
Good Luck

 

You can delete the mcupdate_GenuineIntel.dll (microcode) as already mentioned and it will revert back to last known MC code in bios. But if the bios is updated with help from Microsoft update servers not much can be done.

FYI. Intel didn't push on you the undervolt lock. It's a gift (poison) from the notebook manufacurer. Put the blame where it should be! And Microsoft won't throw out new bios update via it's servers without accept from the notebook manufacturer.

The real dumb stupid, is ASUSTek COMPUTER INC if this is the brand you have. They could offer from both worlds as security or undervolt if they bothered to try make a proper firmware.

Here is a good advice. Never click on a button you don't know what is! And always have the door closed http://forum.notebookreview.com/threads/the-throttlestop-guide.531329/page-1087#post-10980641

I even repeated the same good advice a few months later in TS guide thread http://forum.notebookreview.com/threads/the-throttlestop-guide.531329/page-1160#post-11018866

 

HI, regarding of manufacturing
I seem to have discovered one curious thing. As far as I know, Intel says that overclocking feature is cpu and chipset thing. The chipset status is reported by setting bit 28 of some msr. Well, one guy named TANWare here unlocked the turbo bins of the HM76 after minor modifying of the Bios, but bit 28 is set up already, as he shown by his print screen. Ok, everything is as it should be.
Yesterday, I discovered on the github site a fragment of some HM76 configuration taken from the Clover. The Samsung Board xxxxxxx chipset HM76, state of msr - bit 28 is unset. HUH ??
If I'm not mistaken, does that mean that the manufactory has the keys to the PCH in which are the fused properties of the chipset?
(I think Bios is a little king to change these properties of PCH)

 

Yeep, some of the system manufacturers have their nasty fingers into PCH. Dell is probably the main Bellwether for must have this feature in Gamingbooks.... http://forum.notebookreview.com/thr...-win-update-bios.831450/page-14#post-11005870

A bellwether is an individual who either leads or indicates trends; a trendsetter.

From my link above...
Intel Boot Guard

As defined by Wikipedia: “Intel Boot Guard is a processor feature that prevents the computer from running firmware images not released by the system manufacturer. When turned on, the processors verifies a signature contained in the firmware image before executing it, using the hash of the public half of the signing key, which is fused into the system’s Platform Controller Hub (PCH) by the system manufacturer (not by Intel). Intel Boot Guard is an optional processor feature, meaning that it does not need to be activated during the system manufacturing. As a result, Intel Boot Guard, when activated, makes it impossible for end users to install replacement firmware such as Coreboot.”

 

And, this is exactly why no amount of effort or modding can turn an Alienware slop bucket into a respectable product. It can and will never be more or do more than its corrupt maker wants it to be or do. Beauty is skin deep. Ugly goes to the bone. They castrate their desktops in similar ways. At this point, their only redeeming quality is the case/chassis they use, and even that has declined dramatically. They've got some crazy ugly-looking systems for sale now.

 

Funny how a company with DIY roots is now one of the most anti consumer brands, right behind apple with their soldered parts.

 

Indeed. Bad leadership destroys everything. Always does. From stupid minds come stupid products. GIGO.

A little bit of leaven...

 

you forgot to include Intel in your anti-consumer roundup

 

That depends on how you look at it. Yes, Intel makes that kind of crap available. SGX is equally sinister when deployed on the wrong product. But, there is a place for it where things like national security and banking are at risk. It's not Intel's fault that their technology is misused by idiots manufacturing castrated laptops for consumers. Blaming Intel for inventing technology that is misused by control freaks is about as silly as blaming a company that manufacturers baseball bats when a lunatic uses one to beat another person to death. (We can apply the same flawed concept to knives and firearms.)

 

unclewebb, why not reschedule ThrottleStop.sys of "%UserProfile%\AppData\Local\Temp" to "%WinDir%\System32\drivers\" and do not delete it when closing the program.

 

Hi,
you can see here the Dell probably likes Boot Guard more, than some others vendors

 

I do not understand your question. What problem are you trying to solve? I cannot find ThrottleStop.sys anywhere on my computer. Are you trying to add a new feature to ThrottleStop?

Why would I want to not delete something after the program is closed? I am confused. Help me out!

 

Hello! I've an Acer Predator Helios 300 with an i7-8750H and a GTX1060, both recently repasted with Kryonaut thermal paste.

I'm undervolting with a -150V offset, and that alone helped with heating issues, prior to repasting the units. I also use a cooling pad, since I had problems with throttling due to overheating in the past. ThrottleStop 9.2 registers idle CPU temperatures at about 70C and GPU at about 80C (I live in a very hot climate, so that's not surprising). While gaming, both peak at about 93C. However, TS also registers a PROCHOT 95°C red flag, as shown in the image below, even though none of the components seem to be that hot.



I haven't found any indication of throttling even when this flag is active (it sometimes greys out for a while, then becomes red again if I start running a more resource-heavy application), but I'm also not very used to TS, since I've only been using it for a few days now.

Is this something I should be worried about? If so, is there anything I can do to further improve the heat dissipation?

 

At startup ThrottleStop.exe (ver 9.2) the folder "%UserProfile%\AppData\Local\Temp" is placed ThrottleStop.sys.

This is not a problem, but a suggestion.

When you close the program, this file is deleted.

 

@OgUrecheK - Now I understand. ThrottleStop.sys is a temporary file. It is part of the driver that ThrottleStop uses. I did not realize that ThrottleStop creates this file and deletes it after ThrottleStop closes. Thanks for telling me about this.

I think it is safer to delete it each time if it is not being used. I did not write this driver so I cannot change how it works.

Your screenshot shows your maximum CPU temperature was 95°C and the PROCHOT 95°C indicator is red so your CPU was definitely thermal throttling. Intel thermal throttling is very smooth. Many users do not really notice a little bit of throttling.

PROCHOT 95°C is an indicator. It turns red when your CPU does any thermal throttling. If your computer has cooled down and it is finished thermal throttling, you can click on this to reset it. That is how it goes from red back to gray. Clear it before you start gaming and then when you are finished, you can look at this and if it is red, you will know right away that your CPU was throttling.

Why did you set Speed Shift EPP to 150. Use a setting of 80 or 84 for more of a balanced profile. A setting of 150 can limit maximum performance. If that is what you are trying to do, I understand.

Your Acer Predator does not appear to have great cooling. Some users like to disassemble their laptops and replace the thermal paste. Sometimes the factory does not do a great job. Even if you live in a really hot climate, your idle temps seem really high.

Do you know that you can undervolt the core more than the cache? These two offset voltages do not have to be set equally. Set the cache to -125 mV and then you can increase the core offset to -200 mV and still be stable. Run Cinebench R20 to see if your performance improves as you adjust the core voltage more and more. Go in steps of -25 mV and run another test.

Have fun playing with ThrottleStop. There are lots of options. It takes a while to learn what all of them do. I forget about some of the features that I never use. There is lots of useful information in this forum so learn how to use the Search feature at the top right corner. If you cannot find an answer, just ask.

 

Most of these settings were lifted from a post in the Acer forums from someone with the same specs as me, so I haven't really played with TS much.

Great! I'll try fiddling with it to see if I can get a better performance. Thank you!

 

If you want balls-to-the-wall performance set it to zero.

Be careful copying settings that other people use. A lot of the people that post stuff on the web shouldn't be allowed to. Not to say they are intentionally posting misleading or deceptive information like the mainstream news media, most just don't know what they are doing.

The good news is, even if you copy somebody else's messed-up settings nothing is permanent using ThrottleSop.

 

Hi everybody, I'm having troubles with chaning the offset voltage on my Dell XPS 15 9570 (i5-8300h with microode 0xD6). I suspect i am victim of a BIOS update that disabled undervolting to mitigate the plundervolt exploit. I didn't know about this and updated my bios to 1.17.1 (latest version) and my CPU microcode reads 0xD6. I downgraded the BIOS to the latest version before the plundervolt "fixes" (1.15.0) and did the "reset BIOS to factory settings" trick, but still i am unable to change the offset voltage (greyed out in XTU, +0.0000 in TS). I suspect downgrading the BIOS didn't revert the microcode update. Am i right? Anyone with a Dell XPS 15 9570 that correctly undervolts can confirm their CPU microcode version? Is there any way for me to re-enable undervolting? Do i need to edit the EFI variables with the EFI USB stick method? Thanks in advance.

 

As far as I know, the EFi variables are just hidden sections of Bios setup, which means you clearly can't change version of microcode update from here.
There are many of tools to find version of your current microcode, try cpu - z.
Does your current Bios shows you really 1.15.0?
The Bios update utility's from vendors may not included just the bios image it may include ME image which downgrade can be restricted, so plenty of analyzing here.
The Windows microcode file in system 32 folder could be problem here very well. It should be simple get over it.
The only bright side of the pile of vendor s...t is that at least one can learn new things.
.

 

My microcode version is 0xD6 as reported by TS, HWiNFO and CPU-Z.
My BIOS version is 1.15.0 as reported in the BIOS itself, in HWiNFO, via the

Code:

wmic bios get smbiosbiosversion

command.
I'm at a loss since everyone that downgraded the bios and reset the bios settings to factory defaults managed to undervolt, but it seems that i'm the only one with undervolt still disabled.

The EFI settings editing method i was talking about is explained on this webpage.

I don't know about the ME image or the Windows microcode file. The only thing i know about ME is that when downgrading bios the ME part will fail because downgrading it is disabled so the entire process will report as failed, but that it can be ignored and the BIOS is indeed downgraded correctly. Other people that managed to undervolt also reported the ME part failing but it shouldn't pose a problem.

 

Most of the imbeciles that control firmware distribution on computers block downgrading the ME region. There are hidden/undocumented workarounds to unlock that area for version downgrades and some high-end desktop motherboards openly allow end-user autonomy, but most systems are subject to the retarded wiles of the Nazi control freaks.

 

Can i ask what would be your suggestion? Am i really the only one with this problem? Everyone else did manage to downgrade+undervolt successfully?

 

Dell goes overboard with lockdowns and I am not sure if there is anything you can do at a firmware level. Probably not. You are not alone. I have seen other XPS owners dealing with this heartache.

Have you already deleted the Intel microcode files in Windows? If not, I would try that and see if you get back voltage control. These files are not needed and deleting them causes no harm.

If not, take ownership of and delete mcupdate_GenuineIntel.dll and reboot. You may have to do it again later, after Windows Updates.

 

Hi, it is great those EFI edits are working, just reread link you posted, those 2 locks are generally for different settings, but as unclewebb stated, OEM decided disable OC with UV as well, so it is probably reason EFI edit is working (my pc not include voltage regulator )
Here on page 1231 is link about taking ownership of Windows system 32 folder.
Good luck

 

I took ownership of the file and removed it. rebooted and reflashed old bios, reset factory settings, i successfully went back to microcode 0xC6 but still i couldn't unlock offset voltage. Next idea?

 

If I were you, give try those EFI vars, which is relatively save.
If you will fail again, I don't know. Things could depend on tool which you are flashing also. From my stand point I would probably get whole spi image from same UV working model taken by programmer and searching for differences. The big question is if Boot Guard can be bypassed, but with HW programmer probably yes..

 

When I put any value other than 0 on speedshift I crash very hard ( bios loop laptop tries to boot into windows, powers off and then tries again..) Any ideas why

Sent from my ZTE A2020G Pro using Tapatalk

 

Does it still do it when you don't undervolt?

 

See if you find something in the links below or comments in the Youtube video. Or look into my thread... BLACK LIST. Adjustable voltage control/turbo ratio limits are locked out due Win Update/Bios

https://www.reddit.com/r/Dell/comments/fxfcos/how_to_get_undervolting_back_after_g3/

https://www.reddit.com/r/Dell/comments/fr0j8v/xps_15_7590_psa_undervolting_gone_in_bios_160/

 

@Filipoos - First sign that you have gone too far with an undervolt is when you are stable at full speed but crashing when lightly loaded or when your CPU is near its minimum speed. Proper undervolt testing includes being stable at any speed or any load.

 

But I am stable with 1 and 2 threads running for 10 hours straight benchmarking, I don't get it what more Is there

Sent from my ZTE A2020G Pro using Tapatalk

 

Sadly i did everything in the books that i know. Couldn't find any new information in Reddit, YouTube video or your thread. It baffles me that everyone managed to restore undervolting just by resetting BIOS settings, everyone but me, and i've went way further in trying...

 

I found out what it was but I don't know if I should change it, I'm fully stable with 0 speedshift got 0 bsods, it's the core offset - 250 for my i7 8750h that you guys suggest, (if I get it down same with cache offset I get higher Temps but fully stable to change to all speedshift values, only 0 allows for such a hefty undervolt on core) it does lower the Temps a lot tho so idk any suggestions? I've disabled speedstep from bios..

Lowers voltage with best possible uv from 1.13 to 1.05-1.06 while still maintaining the same performance(cinebench r20 load)

Sent from my ZTE A2020G Pro using Tapatalk

 

You could try the process outlined here in this video:


From the comments, it seems a lot of people have had luck following it for themselves... Still probably can't revert ME though.

 

Repasting the units with high quality paste did lower the temps from an average 76C to about 69C, which is what I'm currently at. I've cleaned the fans and am trying to keep the surroundings clear to ensure proper air flow, but that's about all I can do for now. The GPU is still running at about 77C lately. I'll take the laptop to a cooler climate soon so I can test how much of an influence it is.

Ran it a few times and ended up with an average 2200 points after 10 or so runs. Good news is I managed to not get a PROCHOT flag with the current undervolting (-200mV core, -120mV cache), but I did see a red POWER flag show up a few times, very briefly.

 

If you cannot increase the core voltage significantly beyond the cache voltage, the usual problem is that you have set your cache voltage too high. If the cache is at -125 or beyond, go back to -100 mV and then see how far you can go with the core.

A stable voltage setting will be stable at any Speed Shift value. 0 or 84 or whatever. Did you try running the TS Bench test with the Random MHz option selected? It might help you find a stable undervolt.

 

Bad news. It looks like reading or writing voltage information to an 11th Gen U CPU is not supported. If you try to do this, the CPU returns an error code. It is no longer just Locked like many of the 10th Gen. It seems that CPU voltage control has been turned off within the hardware. Do not hold your breath waiting for a fix. Perhaps Intel is saving this feature for the 11th Gen H series or maybe only the mobile K series will be worthy.

Check out what a fully loaded 10th Gen U CPU is capable of. Good luck finding an 11th Gen U that can compete with this performance.

https://i.imgur.com/NQHgK5G.png

 

who do you think you are, trying to mess around with their CPU? just because you bought it doesn't mean anything. processor-as-a-service. you're just paying for the privilege of using their property. don't dare think you can have any say in what it does or how it works once it's in your hands. just pay up and be grateful it hasn't been embedded yet with microcode to self-destruct after X many cycles of lifetime usage, enter lockdown safemode upon detection of unauthorized access. (seriously though, just wait, eventually we will be renting disposable CPUs with a hardcoded expiration date that lock down and require remote flash manufacturer reset if any modification or analysis is attempted)

 

Micro$lop doesn't have that update ready for them yet.

 

Something tells me we are headed towards cloud computing with thin clients we have not control other than turning the power on/off. Cloud gaming streaming services obviate the need for robust GPUs. If we can play Witcher 3 on an iPad via Geforce Experience, then who's to say we need computers more powerful than a chromebook.

 

The really stupid parts are (1) that's not what we want, and we're willing to spend lots of money on what we want; (2) it takes how many crappy tablet sales to equal one enthusiast desktop or laptop(?) and they're not willing to give us what we want in exchange for that money; and, (3) there are some people that are actually dumb enough to believe that it is a great idea.

 

Hi @unclewebb
I have a lenovo y510p with a 4700mq and a slightly modified cooling system.
I've been using throttlestop for a while, I untick BD prochot, set a non turbo ratio of 1 (non locked), undervolt of 50mv core, cache and gpu, and overclock to 3.4 GHZ all cores.
I tried running benchmarks like TS bench and cinebench R15, and my temps rarely exceed 93c, by that time (after 28s) PL1 kicks in and the CPU is throttled to 3.2 GHZ.
But when running cinebench R20, my CPU immediately throttles to 3.0 GHZ, this time it's not the PL1, but the the EDP current limit, which becomes red from the start and for the duration of the benchmark.
I wonder if you could help me figure out why is that, and how can I fix it.

 

unclewebb,

I wanted to thank you hugely for all of the stunning work you do on Throttlestop. Since my days of the mighty XPS 9550, when I was desperately trying to figure out how to tame those fans, stumbling over TS (and a great tool written by someone else to access the embedded controller of those fans), I have been using TS since then - always very successfully.

The latest success is of a nature which I barely cannot believe on my own, though the benchmarks seem to speak a clear story: I just purchased a Matebook X Pro 2020 with the i7-10510U CPU. That notebook is in reviews known to perform well below what it can, due to Huawei's doubtful cooling system leading the thing to thermal throttling quickly.

But TS helped on more time! Thanks to someone very smart in the reddit forums (u_Cat_Turbo), they found a way to write into the UEFI to unlock the locked voltages which Huawei closed with one of their BIOS "upgrades". Hence the path for TS was open, and here are the stunning results. I can only encourage anyone with an i-7 U CPU and the ability to unlock the voltages to try this too:

- CPU Core: -219.7 Volt
- All other: -50.8 Volt
- Turbo boost long set to 25 (was 18)
- Turbo boost short disabled (put the time limit from 28 to 20 which in spite of disabling that did have an impact on the scores)

Results:
- Geekbench from a poor 3140 to a whopping 4255 in multi-core (also a comparable strong increase in single-core)
- Cinebench R20 from 1327 to 1698
- Fans are even when plugged in not kicking in when doing normal work. And the cooldown speed is significantly faster.
- Rock solid so far. Ran 12 hours in high and light loads.

I do get throttling still (of course P2 due to the 25W cap, and at times short spikes of thermal throttling), but I decided to stop optimizing and now start to have fun with my new machine

So with that: THANK YOU to you and all of the incredible magicians making us get what we paid for!

Thanks & stay safe & healthy,
Thomas

 

@fanlessfan, great post. But give something back to @unclewebb... post some screen snips for him! He can't get enough of the success stories they tell.

 

With pleasure Three screenshots:

Spoiler: 1) Main interface

Spoiler: 2) Undervolting interface

Spoiler: 3) Cinebench post and pre undervolting.

 

@fanlessfan images not showing.

 

Struggled a bit with embedding them. Are they visible now?

 

You have to use a direct link to the image, not the web page hosting the image. You can usually right-click directly on the picture and choose "Open Image in New Tab" from the context menu and then use the direct link to the image that opens in a new page placed between the "IMG" code as shown in the box below. Like this:

Code:

[img]https://www.bilder-upload.eu/upload/8fb871-1604175077.png[/img]
[img]https://www.bilder-upload.eu/upload/f3c321-1604175113.png[/img]
[img]https://www.bilder-upload.eu/upload/b87574-1604175145.png[/img]

Which will produce this...

 

You are welcome. It is always great to see a success story. When fully unlocked, the 10510U is a powerful 4 core CPU. It is definitely one of ThrottleStop's success stories. This is one of my fav screenshots but it looks like you got this one beat.

https://i.imgur.com/NQHgK5G.png

Most people will need to adjust their glasses when they see your humble 15W CPU peaking at over 56W. Congrats. The last guy started getting concerned that maybe he had gone too far. I do not think so. These are more like 45W H series CPUs that someone at Intel put in the wrong box.

Have you tried setting your long term turbo power limit higher than 25W? Did you enforce a 25W cap or does your laptop enforce 25W long term? I would definitely go higher to see what is possible.

Edit - I was able to see your three pictures.

By stopping at 25W, you might have left 300 Cinebench points on the table. I hope you can go further. It looks like you need to increase the PP0 Current Limit and the IccMax value. I like 256 but 100 is a nice compromise without going too crazy.

Unfortunately the PP0 Current Limit is locked to 85 A in the Y510P. Hard to get by that limit without a modified BIOS. Have you tried using the PowerCut feature on your Y510P? That might be one way to get over this hurdle.