How to unlock Alienware m16 r1 undervolt for ThrottleStop?

[Alienware_Melting]

Hello all,

I've been using the software very well for years on my old predator helios 300, this software basically made the lifetime of that setup much longer because it does shave off several degrees and that machine was so hot stock.

But now I cannot use it with my new Alienware m16 r1, it is a i9-13900HX RTX 4070. This is another toaster, it get ridiculous hot so fast.

The menu in FIVR does show it is protected like img below

so how do I unlock this thing without screwing up my computer? My firmware is 0.1.14.1, windows automatically force updated it (tho I'm not sure this has anything to do with it)

I already disabled all the virtualization in the Bios, already uninstalled the virtualization crap in windows 11 too. Like img below

Maybe I skipped something? Or do I need to boot that secret menu to disable the undervolt protection like this video:

 

[kurosagi01]

You can probably only undervolt in the bios if it's unlocked by the bios, which is what I had to do on my x16r1.
I mainly use throttle stop to lock the CPU p-core and e-core multiplier to help keep temps down by stopping the CPU from boosting to its highest speed.
I only use it for gaming so I don't need the CPU to try and run anymore than around 4ghz.

 

[unclewebb]

When ThrottleStop reports that Undervolt Protection is enabled, you cannot use ThrottleStop to undervolt your CPU. Undervolt Protection has to be disabled before you boot up.

The advanced menu shown in that video is exactly what you need to access. Disable Undervolt Protection, boot up, and you should not see Undervolt Protection in the FIVR window anymore. Once that is got, you can undervolt.

Everything else looks OK. If you want to test that ThrottleStop is working OK, instead of an undervolt, try applying a small overvolt. Set the offset voltage to +5 mV, press Apply and this should show up in the FIVR monitoring table immediately whether Undervolt Protection is enabled or not.

Post your results if you get this working with the hidden BIOS menu.

 

[Alienware_Melting]

When ThrottleStop reports that Undervolt Protection is enabled, you cannot use ThrottleStop to undervolt your CPU. Undervolt Protection has to be disabled before you boot up.

View attachment 335327

The advanced menu shown in that video is exactly what you need to access. Disable Undervolt Protection, boot up, and you should not see Undervolt Protection in the FIVR window anymore. Once that is got, you can undervolt.

Everything else looks OK. If you want to test that ThrottleStop is working OK, instead of an undervolt, try applying a small overvolt. Set the offset voltage to +5 mV, press Apply and this should show up in the FIVR monitoring table immediately whether Undervolt Protection is enabled or not.

Post your results if you get this working with the hidden BIOS menu.

Hi thank you for the reply. Yes TS now shows it is working, but the temps still run high with current levels of undervolting.

what happened tho is that both benchmarks got a bit better, it was something like 4.7 and 45.xx IIRC.

Should I align the other items to the same undervoltage? Like Intel GPU? I aligned only the Cache so far.

Anyways, I'll notch the offset down bit by bit and test throughout the weeks and see, but I guess I need some sort of fan control too and I'm not sure the thermal paste is good or not (with my previous helios 300 I also changed the stock thermal paste to a better non-liquid metal one), but people are saying those alienwares come from the factory with better thermal paste from Dell already and disassembling them is a known pain in the butt (inverted MB).

 

[kurosagi01]

There are a couple of things you can try which is how I've managed to keep my temps down on both CPU+GPU just under 75c(mainly playing Warzone i've seen these temps, but average around 60-72c) on both chips.
Just for reference, I have a X16 with 13900HK and 4080, some may disagree with this approach as you are "missing" out in performance but we are talking about the balance of longevity and performance for a laptop, we want it to last and minimising the need to replace the thermal paste so often.
The CPU is only configured to undervolt at -40mv(max limit in the bios)
Steps i've taken to keep temps down:
1. Set your RGB light configuration in Alienware command center, then uninstall anything related to Command center
2. Install https://github.com/T-Troll/alienfx-tools (your laptop should be supported), allows you to manually control your fan curve for both CPU+GPU
3. Install MSI afterburner to set a undervolt curve + figure out your GPU default boost clock(for the 4080 its 2280mhz: https://www.nvidia.com/en-gb/geforce/laptops/compare/
4. Allow MSI afterburner, AlienFX-tools to run at start up
5. In Throttlestop(the disagreement will be here), set the P-core&E-core multiplier to something you are comfortable with for daily/gaming use, I have mine set to run at 40x(4ghz), we don't want it to boost upto 5.4 or 5.8ghz whenever it wants to. Which is overkill for gaming.
6.Buy a IET cooling pad

 

[Bogus5]

Can you post a screenshot of the entire FIVR window (without cropping)?

 

[unclewebb]

Should I align the other items to the same undervoltage?

You have a Nvidia GPU so in most situations, the Intel GPU is not going to be consuming a lot of power. Most people do not bother undervolting this. If you do decide to try and undervolt the Intel GPU, undervolt both the Intel GPU and the iGPU Unslice.

Most people do not undervolt the System Agent. Some users have reported that undervolting the CPU E Cache does not accomplish too much. The Core and P Cache are the two most important voltages to adjust.

An Alienware that can pull almost 160W is probably about as good as it gets. You can try using another paste like Honeywell PTM 7950 but 160W might be close to the limit of what your heatsink is capable of. I think some of the Lenovo Legion series have a little better cooling but not much. There is only so much heat that something the size of a laptop can dissipate.

 

[A&P211]

You have a Nvidia GPU so in most situations, the Intel GPU is not going to be consuming a lot of power. Most people do not bother undervolting this. If you do decide to try and undervolt the Intel GPU, undervolt both the Intel GPU and the iGPU Unslice.

Most people do not undervolt the System Agent. Some users have reported that undervolting the CPU E Cache does not accomplish too much. The Core and P Cache are the two most important voltages to adjust.

An Alienware that can pull almost 160W is probably about as good as it gets. You can try using another paste like Honeywell PTM 7950 but 160W might be close to the limit of what your heatsink is capable of. I think some of the Lenovo Legion series have a little better cooling but not much. There is only so much heat that something the size of a laptop can dissipate.

160w is the upper upper limit of laptop cooling. This asus g16 that I have does around 110-120w with a -.100mv undervolt.

 

[Bogus5]

160w is the upper upper limit of laptop cooling. This asus g16 that I have does around 110-120w with a -.100mv undervolt.

Delidding – Kyrosheet vs PTM7950 as TIM

Hey folks, hope everyone is keeping well :) Continuing my adventures on seeing how different things that aren't LM perform as TIM for my 12900k (between the Die & IHS), I grabbed some Kryosheet recently as it was available locally (which is a surprise in itself) and at a reasonable price as...

www.fpshub.com

 

[A&P211]

Delidding – Kyrosheet vs PTM7950 as TIM

Hey folks, hope everyone is keeping well :) Continuing my adventures on seeing how different things that aren't LM perform as TIM for my 12900k (between the Die & IHS), I grabbed some Kryosheet recently as it was available locally (which is a surprise in itself) and at a reasonable price as...

www.fpshub.com

My laptop comes with liquid metal TIM.

 

[Bogus5]

My laptop comes with liquid metal TIM.

Mine too, the problem is that it's even worse because the LM leaks from the CPU kernel onto the laminate and it's very common.

 

[kurosagi01]

All new Alienware laptops from 2023 to now comes with their "element 31", which is a mixture of liquid metal and thermal compound.

 

[Alienware_Melting]

There are a couple of things you can try which is how I've managed to keep my temps down on both CPU+GPU just under 75c(mainly playing Warzone i've seen these temps, but average around 60-72c) on both chips.
Just for reference, I have a X16 with 13900HK and 4080, some may disagree with this approach as you are "missing" out in performance but we are talking about the balance of longevity and performance for a laptop, we want it to last and minimising the need to replace the thermal paste so often.
The CPU is only configured to undervolt at -40mv(max limit in the bios)
Steps i've taken to keep temps down:
1. Set your RGB light configuration in Alienware command center, then uninstall anything related to Command center
2. Install https://github.com/T-Troll/alienfx-tools (your laptop should be supported), allows you to manually control your fan curve for both CPU+GPU
3. Install MSI afterburner to set a undervolt curve + figure out your GPU default boost clock(for the 4080 its 2280mhz: https://www.nvidia.com/en-gb/geforce/laptops/compare/
4. Allow MSI afterburner, AlienFX-tools to run at start up
5. In Throttlestop(the disagreement will be here), set the P-core&E-core multiplier to something you are comfortable with for daily/gaming use, I have mine set to run at 40x(4ghz), we don't want it to boost upto 5.4 or 5.8ghz whenever it wants to. Which is overkill for gaming.
6.Buy a IET cooling pad

Thank you very much for all the tips here, much appreciated.

I think it does make sense to disable the turbo boost and also make the processor less jumpy to higher clocks by capping the Speed Shift, done that in my previous machine and kept temps cooler on average with acceptable CPU response (128 level - if you put 64 it is already going to push clocks up so fast for almost nothing usually). Also most games won't be that single CPU intensive.

I'1ve ran Cinebench with those same levels of undervolt (-75.2), the system seems stable and these are the results so far:

Not sure if good or bad, have to google the numbers.

Can you post a screenshot of the entire FIVR window (without cropping)?

the other params are stock ones, didn't change anything but offset voltage, the offset was done over the "performance" profile that comes stock with TS.

but here it is (notice that now I just put the intel GPU and iGPU slice under the same undervolting value)

 

[Bogus5]

Yes, they're standard for your computer, but not for mine. Your IccMax as standard is 230 A and I have 240 A, and that difference makes a lot.

Lowering the voltage of the Intel GPU and iGPU more than 50 mV doesn't really make sense, but lowering the IccMax value for the CPU Core and increasing it for the rest of the components makes for real tuning.

 

[A&P211]

Thank you very much for all the tips here, much appreciated.

I think it does make sense to disable the turbo boost and also make the processor less jumpy to higher clocks by capping the Speed Shift, done that in my previous machine and kept temps cooler on average with acceptable CPU response (128 level - if you put 64 it is already going to push clocks up so fast for almost nothing usually). Also most games won't be that single CPU intensive.

I'1ve ran Cinebench with those same levels of undervolt (-75.2), the system seems stable and these are the results so far:

View attachment 335619

Not sure if good or bad, have to google the numbers.


the other params are stock ones, didn't change anything but offset voltage, the offset was done over the "performance" profile that comes stock with TS.

but here it is (notice that now I just put the intel GPU and iGPU slice under the same undervolting value)

View attachment 335622

Can you do a higher undervolt, the processor in my laptop is stable at .120mv, but I use .105mv.

 

[Alienware_Melting]

Can you do a higher undervolt, the processor in my laptop is stable at .120mv, but I use .105mv.

Yes I'll do it over time, I'll just notch it down a bit now and do it and report back, let's see if I can go to your levels of -105. I'll try just because you are telling me you can get even -120, so I'll just put it there and see.

For comparison my Helios 300 with a i7 7700HQ is at -150.4mv, running stable for 4 years. I don't think there will be much improvement in temps if I go down and down, usually things like Speed Shift and turning off boost help better, even in gaming because you usually do not need your CPU at max and responding super quick.

I just pushed for now the lowest undervolting value I've found searching around that is almost guaranteed to work. I do not want to go pushing the OS or I'll need to reinstall the garbage windows 11 and configure everything again. When I was undervolting my helios, below those 150 it crashed and did some nasty things and it took a LONG time to go back and forth finding the sweet spot.

I don't think getting some benchmarks 5% better is worth the headache of unstable system.

 

[Bogus5]

I don't think getting some benchmarks 5% better is worth the headache of unstable system.

Perhaps a 5% performance gain is not much, but when you spend 3 months "carving" an expensive computer to find all the scenarios useful for your work and you keep hitting the 35000 wall in Cinebench R23 and then find a setting that allows you to easily overcome this magical limit, is it a success or a failure?

Significantly, you are doing this against the recommendations of the author of TS with the help of TS.

In the Asus G16 (2023), an 18% increase in performance was caused by replacing memory modules from DDR5 4800 MT/s to DDR5 5600 MT/s, and this 5% increase from recent changes is only a bonus.

 

[Alienware_Melting]

Can you do a higher undervolt, the processor in my laptop is stable at .120mv, but I use .105mv.

Ok so I've done -100 and it seems -120 works, but strangely the single core benchmark didn't chance at all. This is with -120

The same tests with TS Benchmark gave slightly better results in 120M, but worse results in 960M

There are a couple of things you can try which is how I've managed to keep my temps down on both CPU+GPU just under 75c(mainly playing Warzone i've seen these temps, but average around 60-72c) on both chips.
Just for reference, I have a X16 with 13900HK and 4080, some may disagree with this approach as you are "missing" out in performance but we are talking about the balance of longevity and performance for a laptop, we want it to last and minimising the need to replace the thermal paste so often.
The CPU is only configured to undervolt at -40mv(max limit in the bios)
Steps i've taken to keep temps down:
1. Set your RGB light configuration in Alienware command center, then uninstall anything related to Command center
2. Install https://github.com/T-Troll/alienfx-tools (your laptop should be supported), allows you to manually control your fan curve for both CPU+GPU
3. Install MSI afterburner to set a undervolt curve + figure out your GPU default boost clock(for the 4080 its 2280mhz: https://www.nvidia.com/en-gb/geforce/laptops/compare/
4. Allow MSI afterburner, AlienFX-tools to run at start up
5. In Throttlestop(the disagreement will be here), set the P-core&E-core multiplier to something you are comfortable with for daily/gaming use, I have mine set to run at 40x(4ghz), we don't want it to boost upto 5.4 or 5.8ghz whenever it wants to. Which is overkill for gaming.
6.Buy a IET cooling pad

I've tried some simple Afterburner adjustments but it seems this GPU is way more sensitive than my last one (1060 6GB). I could push around 175/200 MHZ in the Curve (more on mV below 900) and 870 MHZ more for the memory. This one won't let me push a over the board 190 without crashing, and frankly the improvement is almost null in cinebench (11700s vs 11500s, zilch). Not sure about other tests.

This 4070 GPU seems to run cool on those tests, the heat problem is 99% the processor. I'll not be gaming 4k, I'll be actually doing 1080, maybe 1600. This should let me run over 60 FPS in all games I suppose.

 

[Bogus5]

I have no idea if anyone's criteria put the total performance gain of 11.3% achieved with the ThrotteStop app a lot or little, but that's what it looks like in the Cinebench 2024 test that was cited here.
A completely cost-free increase, not counting the time wasted on these experiments.

 

[Alienware_Melting]

Ok so I've done -100 and it seems -120 works, but strangely the single core benchmark didn't chance at all. This is with -120

View attachment 335690

The same tests with TS Benchmark gave slightly better results in 120M, but worse results in 960M

View attachment 335691


I've tried some simple Afterburner adjustments but it seems this GPU is way more sensitive than my last one (1060 6GB). I could push around 175/200 MHZ in the Curve (more on mV below 900) and 870 MHZ more for the memory. This one won't let me push a over the board 190 without crashing, and frankly the improvement is almost null in cinebench (11700s vs 11500s, zilch). Not sure about other tests.

This 4070 GPU seems to run cool on those tests, the heat problem is 99% the processor. I'll not be gaming 4k, I'll be actually doing 1080, maybe 1600. This should let me run over 60 FPS in all games I suppose.

Ok so for anyone searching this later on: I cannot put -150 undervoltage, windows freezes and it doesn't boot (actually doesn't log since I've put TS as a task to initiate upon logging, NOT upon system startup), I needed to reset undevoltage lock to boot windows. So far -120 seems to be stable. I gamed a bit, ran cinebench, downloaded things, used browser and restarted windows several times.

 

[PHVM_BR]

Ok so for anyone searching this later on: I cannot put -150 undervoltage, windows freezes and it doesn't boot (actually doesn't log since I've put TS as a task to initiate upon logging, NOT upon system startup), I needed to reset undevoltage lock to boot windows. So far -120 seems to be stable. I gamed a bit, ran cinebench, downloaded things, used browser and restarted windows several times.

Do you have -120mV for core and cache?

Try maintaining -120mV in cache and lowering the core.
It can probably reach -140mV or -150mV.

Alienware m16 with RTX 4070 or below comes with factory default thermal paste.
Element 31 only in versions with RTX 4080 or higher.

If you apply PTM 7950 you will obtain a gain of up to ~15% in thermal headroom.

 

[A&P211]

Perhaps a 5% performance gain is not much, but when you spend 3 months "carving" an expensive computer to find all the scenarios useful for your work and you keep hitting the 35000 wall in Cinebench R23 and then find a setting that allows you to easily overcome this magical limit, is it a success or a failure?

Significantly, you are doing this against the recommendations of the author of TS with the help of TS.

In the Asus G16 (2023), an 18% increase in performance was caused by replacing memory modules from DDR5 4800 MT/s to DDR5 5600 MT/s, and this 5% increase from recent changes is only a bonus.

What memory modules works on your g16. I am still trying to find better ram than the stock. I tried 4800mhz cl38 but didnt start with that ram.

Ok so I've done -100 and it seems -120 works, but strangely the single core benchmark didn't chance at all. This is with -120

View attachment 335690

The same tests with TS Benchmark gave slightly better results in 120M, but worse results in 960M

View attachment 335691


I've tried some simple Afterburner adjustments but it seems this GPU is way more sensitive than my last one (1060 6GB). I could push around 175/200 MHZ in the Curve (more on mV below 900) and 870 MHZ more for the memory. This one won't let me push a over the board 190 without crashing, and frankly the improvement is almost null in cinebench (11700s vs 11500s, zilch). Not sure about other tests.

This 4070 GPU seems to run cool on those tests, the heat problem is 99% the processor. I'll not be gaming 4k, I'll be actually doing 1080, maybe 1600. This should let me run over 60 FPS in all games I suppose.

Here are some of my. The -.155mv in unstable so I tried highering it until I got to -.120, I keep it at -.105mv for stability

I think this is good for a mobile processor
View attachment 334726


With a .155mv undervolt
View attachment 334746

 

[Bogus5]

What memory modules works on your g16. I am still trying to find better ram than the stock. I tried 4800mhz cl38 but didnt start with that ram.

I used 2×16 GB Kingston Fury memory modules.

Kingston FURY Impact DDR5 SODIMM Memory – 8GB-64GB/6400MT/s - Kingston Technology

Plug N Play kit capacities up to 64GB, Kingston FURY Impact DDR5 Plug N Play modules overclock to the highest frequency published, up to 5600MT/s, without adjusting BIOS settings. Modules featuring XMP 3.0 profiles can overclock to the highest frequency published, up to 6400MT/s, by selecting...

www.kingston.com

If you apply PTM 7950 you will obtain a gain of up to ~15% in thermal headroom.

PTM works better in desktops and less powerful laptops, but in powerful laptops with hot CPUs they lose all advantages.

 

[PHVM_BR]

PTM works better in desktops and less powerful laptops, but in powerful laptops with hot CPUs they lose all advantages.

You are wrong.
M16 R1 with well applied PTM 7950 can sustain ~160/170W.
Only LM can deliver more (~180/185W).

 

[Bogus5]

You are wrong.
M16 R1 with well applied PTM 7950 can sustain ~170W.
Only LM can deliver more (~180/185W).

Have you heard of the Grizzly KryoSheet thermopads?