Alienware Area 51m Undervolt and Overclock Guide

Introduction
The Alienware Area 51m is probably the most versatile desktop replacement laptop from 2019 thanks to it being able to run up to an RTX 2080 and a i9 9900K in a chassis the size of a 15-inch laptop of few years back. Unfortunately, due to some idiosyncrasies from Dell, this device is hampered in its performance and usability for gaming.

Disclaimer
The procedures described in this tutorial may cause instability on your device or it may damage it or void the warranty. This guide is provided as-is and the author exonerates responsibility from the effectiveness of the tutorial, consequences of this tutorial and the software featured in this tutorial.

The frequencies and setting shown are for guidance only and don’t guarantee you will be able to reach or surpass what is described since each laptop is different.

This tutorial is for the Alienware Area 51m, and the procedures described may not work with other devices.

Pre-requisites
The following software must be installed on the device:

• Alienware Command Center.
• MSI Afterburner.
• Throttlestop (Extract the software to C:\Thorttlestop\)
• HWinfo64
• 3D loop application such as Unigine Valley or Heaven.
• Cinebench R20 (available at the windows store).

The laptop must be running BIOS 1.7.3 (at least). vBIOS 1.0.0.5 (180W power limit) is reccomended.

It is highly recommended the laptop be re-pasted and re-padded. Check this thread for guidance:
http://forum.notebookreview.com/thr...-for-alienware-area-51m-9900k-rtx2080.829543/

The use of a cooling pad is recommended for gaming and rendering. Optionally, just raising the back of the laptop by an inch will increase cooling performance.

A second monitor is very helpful so you can run the tests and observe what the device is doing.

Initial configuration
We first need to set up Alienware Command Center (AWCC for short) so it doesn’t interfere with MSI afterburner and Trhottlestop, and unlock the higher thermal limit.

1. Launch ACC and navigate to the FUSION tab.
2. Create a new OC profile. Name it something such as “Default” or “Undervolt”.
3. Make sure that the CPU slider is deactivated. Activate the GPU slider.


4. Click on the GPU Advanced view. Slide the Temperature slider to 78°C. Slide the GPU frequency and memory frequency to 0. Click Test & Save.


5. Go to the home page and select this profile on the Overclock option.


GPU Undervolting and Overclocking (i.e. removing the Thermal throttling)

This is probably the reason of why you are here, you want your RTX 2080 to run at decent speeds but avoid the hard throttling that occurs at 79°C. But first some background.

First manufactured A51m’s featured a poor thermalpad application and less than ideal MOSFETs for the GPU that could cause these burning out. In addition, some MOSFETs where burning out due to the laptop was used with the lid closed, preventing air be taken from the top vents. It also didn’t help Dell releasing a vBIOS that could allow the RTX 2080 to run at 200W. To prevent further laptops failing, power of the GPU was limited at 180W, and maximum temperature was reduced to 79°C. This has been a major issue for users, causing huge disgust in the community since when reaching 79°C the GPU would enter a power down state seriously affecting performance until cooldown.

Undervolting the GPU and limiting temperature at AWCC will allow us to achieve higher GPU clocks while maintaining a power consumption that keeps the GPU under the required temperature envelope. All this will result in an uninterrupted gaming performance and a longer lasting device.

Preparing MSI Afterburner and undervolting
1. Launch MSI Afterburner. Go to settings. Check “Unlock voltage control” and “Unlock voltage monitoring”. Set the drop-down menu to “third party”


2. Go to the monitoring tab and check the “GPU1 voltage”. Set it to show on the On-screen Display in conjunction to your other preferred settings.


3. Go to C:\Program Files (x86)\MSI Afterburner\Profiles . Right click on the shown files and edit them using Notepad++


4. Replace the content of both files with the following:

[Startup]
Format=2
CoreVoltageBoost=
PowerLimit=
ThermalLimit=
ThermalPrioritize=
CoreClkBoost=
MemClkBoost=
[Settings]
VDDC_Generic_Detection=1

Save the files


5. Launch HWinfo64. Check the “Sensor only” on the launch screen.

6. On the list of information, right click on “GPU Power” and select “Show Graph”. Set the maximum value to 180 (if using the 200W vBIOS, set to 200).


7. On AWCC, set the “Thermal” to “Full Speed”


8. We need to determine the maximum offset your GPU can handle. Open your preferred 3D loop application and run it on windowed mode at a low resolution (such as 1280x720). While the loop is running increase frequency until the game is not stable.
9. Close the 3D application and go to MSI Afterburner. Do Ctrl+F to open the Curves setting. You will see something like this:


Now we need to start flattening the curve to stop the GPU from running too high voltages. Start with a voltage such as 968mV and drag all the points to the right down up to the same frequency. Do this until you reach 1100mV since the GPU can’t go higher voltages. Click Apply.


10. Launch the 3D application and let it run for a while. Close it and get to HWinfo64. Check the graph and proceed as follows.
a. Power is pegged all the time/GPU reaches 79°C: Choose a lower voltage point (do not exceed your maximum frequency offset).


b. Power does not reach the maximum limit/GPU runs cooler than 75°C: Choose a higher voltage point (do not exceed your maximum frequency offset).


Re-iterate the process until you find a balance where the laptop runs stable at 76~78°C after a prolonged period of time
11. Proceed to overclock the memory. Stop when you see artefacts and reduce by 100MHz.

Test your settings with games! Some games are more taxing on the GPU and you will need to create different profiles to avoid the GPU from reaching 79°C.

Use G-Sync or V-Sync. Limiting the refresh rate can provide the GPU relief and cool down periods.

Want to run quieter? Use more aggressive undervolts to prevent the GPU from reaching 79°C on other Thermal profiles.

Bonus, extra battery
On MSI Afterburner slide the “Core Clock” and “Memory Clock” to minimum. Save to a profile and activate it when running in battery.

CPU Undervolting and Overclocking
Either if you are running an 8700(K), 9700(K) or an 9900(K), you will find that on some applications the CPU will reach 100°C and it will reduce its frequency to maintain that temperature and keep the CPU from getting damaged. This is not ideal and we can configure how fast the CPU can run and how much voltage it uses to prevent this happening on different scenarios.

To control the CPU, we are going to use Throttlestop. This program allows us to use 4 different profiles, and we are going to use them as follows:

1. Balanced: A profile with good performance for general use. Target ~140W sustained load.
2. Gaming: A profile for applications where single thread performance is important such as games.
3. Render: A profile where all cores are at maximum usage for prolonged periods of time. Target ~110W sustained load.
4. Battery: A profile that limits CPU speed severely to reduce power consumption. Target ~45W sustained load.

Setting up Trhottlestop and finding the CPU maximum frequency.
1. Launch Throttlestop a go to options. Set everything like shown. Click OK.


2. Click on “Turn On”.


3. Click on “TLP” Check that everything is like the picture shown. There is no need to change the Speedshift values, leave them as-is.


4. Click on “FIVR”. Under “Save voltages to Throttlestop.INI”, select “OK – Save voltages immediately”
5. Select profile 2 (Gaming) and make sure that the “Overclock” check box is active.
6. Now change the multiplier of the highest core to something like 50. Apply and Ok.


7. On ACC, set the “Thermal” to “Full Speed”
8. Click on “TS Bench”. Select 1 thread on the drop-down menu, “Size 64M” is ok. Click on start.

a. Test is completed successfully: Proceed to increase the multiplier by 1.
b. Test gives errors: Decrease the multiplier by 1.
​

Repeat the test until you find the highest multiplier that doesn’t give errors. This is your highest frequency the CPU can work with, and this value will be used as a reference.

Creating the different profiles

Gaming Profile
1. Click on “FIVR”. Choose profile 2.
2. Set the multiplier to the maximum CPU frequency you determined earlier.
3. Under FIVR control select “CPU Core” and check “Unlock Adjustable Voltage”.
4. Move the slider to ~ -10mV.


5. Under FIVR control select “CPU Cache” and check “Unlock Adjustable Voltage”.
6. Move the slider to ~ -10mV (The GPU core and cache must have the same offset).
7. Click Apply and Ok.
8. Launch “TS Bench”. Set threads to 4 and size 256M. Click “Start”.

a. Test is completed successfully: Proceed to decrease voltage by ~10mV.
b. Test gives errors: Proceed to increase voltage by ~5mV.​

9. Once you found a voltage offset that doesn’t give errors or crash the computer, set the size to 1024M and continue fine tuning the voltage offset.
10. Optionally, you can run a test with all threads, but overheating will cause throttling down, so is best to revert back to 64M or 256M length tests.

Balanced and Render profiles
On these profiles we are going to do things differently. Instead of searching for the maximum speed the CPU can run stable, we are going to look for the power the CPU consumes when working full bore. The indicated power targets are approximate what the Area 51m can sustain for short periods (Balanced profile) and indefinitely (Render profile). These values can be changed to suit particular machines.

1. Click on “FIVR”. Choose profile 1 (Balanced) or 3 (Render).
2. Set the multiplier to a CPU frequency you consider it could meet the power target.
3. Move the sliders (CPU Core and Cache) left ~ -10mV.
4. Click Apply and Ok.
5. Launch “TS Bench”. Set threads to the maximum capable of your device and size 256M. Click “Start”.

a. Test is completed successfully: Proceed to decrease voltage by ~10mV.
b. Test gives errors: Proceed to increase voltage by ~5mV.
6. Once the Undevolt is stable, Launch Cinebench R20. Go to File -> Preferences. Change “Test duration” a at least 600 seconds.


7. Start the test and check the power consumed by the CPU. Make changes to the multiplier and undervolt to meet the power target.


8. Once everything is dialled in, run Cinebench R20 for 10 minutes to validate the stability.


Battery profile
With the battery profile, things a re very similar, but we want to avoid the CPU turboing up. Proceed as follows:
1. Select the Battery Profile.
2. Check the “Disable Turbo”
3. Uncheck “Speed Step”

4. Click on “FIVR”. Uncheck “Overclock”.


5. Move the sliders (CPU Core and Cache) left ~ -10mV.
6. Click Apply and Ok.
7. Launch “TS Bench”. Set threads to the maximum capable of your device and size 256M. Click “Start”.
a. Test is completed successfully: Proceed to decrease voltage by ~10mV.
b. Test gives errors: Proceed to increase voltage by ~5mV.
8. Once you found a voltage offset that doesn’t give errors or crash the computer, launch Cinebench to validate the undervolt.


Some notes:
* Activating Speed Shift – EPP will cause shuttering on games.
* For best results, set multipliers of all cores the same speed.
* If the laptop soft crashes in desktop, increase voltage by 2 to 3 clicks.

References

http://forum.notebookreview.com/threads/the-throttlestop-guide.531329/
https://www.reddit.com/r/Alienware/comments/b2r1rn/my_area51m_started_smoking_i_took_it_apart_and/
http://forum.notebookreview.com/threads/official-alienware-area-51m-owners-lounge.826831/

 

Tutorial updated due to changes on the temperature management system introduced on BIOS 1.8.1.

 

Dude this is great, thank you. Ill rep you son as I can.

 

This has to be a massive amount of tine and effort.

Thank You!

 

Thanks, indeed this took a lot of time to make, but I really thought people needed it.

 

What a legend! Thanks for this mate!

 

Thank you sir!

Enviado desde mi GM1913 mediante Tapatalk

 

1 last thing Can you Screenshot all the settings you have from HWinfo?

 

Settings from HWinfo64 are the defaults

 

Instead of using AWCC to raise the GPU limit, would you recommend or have you tested Asus GPU Tweak in comparison? The reason I ask is that I use ThrottleStop (TS) to OC and UV my CPU, however, in order for AWCC to not mess with the TS settings I need to set the OC profile in AWCC to off. If I follow your guide above, it would require me turning on the OC profile, affecting TS.

I hope this makes sense and I can elaborate if needed.

 

Since the last update of AWCC I found that even turning OC off doesn't prevent AWCC from turning the parameters of Throttlestop and Afterburner off. ASUS GPU tweak took is of no help.

So far the only solution I can find is to manually select a Throttlestop and Afterburner profile a few minutes after you have logged in, and to monitor the parameters using the MSI Afterburners' On-Screen-Display. Should the setting go away, you will have to re-apply manually.

 

Since I tuned off the OC profile from the AWCC home screen, it no longer changes my OC frequency set in TS? Have you made that change on the home screen? As I initially had only disabled the slider within the Fusion section thinking that would stop AWCC overriding TS but it didn't.

Here is what I mean, as I'm useless at explaining. This is from the 'Home' tab in AWCC.



Here are the profiles within 'Fusion' untouched.

 

With the latest version of AWCC, it doesn't matter where you disable the OC, for some reason it still reverts the changes made by Throttlestop an Afterburner at logon. Previously this wasn't the case.

 

How do you find out your version of AWCC, I assumed I was on the latest as I got an update yesterday. Hoping I miss it, if its going to start messing with TS.

 

Checked this at the Microsoft Store

 

Thanks brother, let me check mine. Need to get some time to properly go through your guide and report back, thanks again

 

I have checked if there is a difference between Asus GPU Tweak and rising thermal limit with AWCC, and it's the same. Really, best is to have less bloat on the laptop.

 

Do you guys see your 2080 hitting 2100mhz? Or would this require overclocking the GPU? Mine only ever seems to reach 1900-1950mhz, can't say I've seen it go higher?

I haven't undervolted yet, I'm not sure what you recommend seeing on the HWInfo64 graph, that bit has confused me (not a hard task when it comes to this)?

Also, what are your idle temps for the GPU?

 

Due to GPU Boost 2.0, GPU frequency is not a theoretical maximum fixed value. When you apply an overclock, what you are telling the GPU is to function at higher speeds for the same voltage, but depending on the temperature and power constraints, the GPU will choose a lower voltage to function. This entitles a lower frequency.

We modify the curve, not only to increase the GPU frequency for any given voltage, but also to avoid the GPU from choosing voltages that consume too much power. Intelligently, the GPU will choose the highest possible frequency at the minimum available voltage.

The power consumption graph tells us if we are limiting the GPU too much or if we should choose a lower voltage point to reduce the throttling. Basically, you want to see if the GPU is hitting 180W all the time or if there is still room for higher voltages and frequencies.

At Balanced fan mode my GPU idles at ~37ºC after a prolonged period of time.

 

Excellent, thank you. It makes more sense now. But the fact I have never undervolted or overclocked the GPU should I not see it hit 2100mhz?

My GPU idles at around 47c, I think a repaste is a must here and something I need to stop talking and instead actually doing it.

 

Thermal and power constraints. If you want to see 2100MHz, you need to put a very light 3D application when the GPU is at less than 40~50°C.

Double check what fan profile are you on or if the computer is not clogged up. 47°C is too much at idle. A re-paste might be very beneficial for your device.

Enviado desde mi GM1913 mediante Tapatalk

 

If I set my oc to +150 core I see 2100mhz until Temps rose of course but I'm on driver 441.08 anything higher just made core clock crazy low about what, you are saying. What driver are you on?

Sent from my GM1917 using Tapatalk

 

I only received mine on the 19th of September and I had it open about 10 days ago to add the new M.2 heatsink and the fans looked fine and clean.

I ran a few Unigine tests earlier testing both 180w and 200w vBIOS and used performance then max fans and still haven't seen 2100mhz.

Those idle temps are on the balanced profile for fans.

I'm on the latest driver but can't honestly remember ever seeing 2100mhz. I haven't attempted any sort of OV/UV on the GPU yet, still new to this game and only really built the confidence to mess with the CPU thus far.

On my testing with Unigine earlier, the 200w vBIOS performed a lot worse than the 180w vBIOS. During the 200w runs when the GPU hit 79c it would downclock to around 500mhz and FPS would hit 40-60. It remained like this until the GPU hit 50c, then it would start to increase both FPS and clocks.

On 180w runs, the clocks didn't drop to the same levels and FPS remained decent. I can post the comparisons scores later if needed.

 

I mean you are on bios above 1.5.0 you need to downgrade it. Then 200w will work better

Sent from my GM1917 using Tapatalk

 

Yeah I'm on 1.7.3, I'd prefer to keep G-Sync for now and I'd no doubt brick my laptop following the downgrade process.

 

If you follow it right you should be fine and I don't use gsync and all games are fine.. No need for gsync with 2080 tbh, I thought It was going to be something special.. But I quickly realized it looks no different..

Sent from my GM1917 using Tapatalk

 

I might give it a go at some point, I haven't saw anything concerning whilst gaming to make me feel the need to downgrade yet.

 

Ah okay ^ ^ yeah as soon as I got mine I just ripped 1.5.0 on it lol... And yeah I was shaking lel..

Sent from my GM1917 using Tapatalk

 

So instead of a 9900ks I might be getting a 8700k or 8086k instead and I was curious can I use this https://rockitcool.myshopify.com/products/copper-ihs-for-lga-1150-1151 with it in the 51M? All I do is game and having the 8th gen cpu means cooler temp aswell.

Sent from my GM1917 using Tapatalk

 

@S.K used an aftermarket IHS and it caused issues. Not sure of the brand but you will be able to search in the A51m owners thread.

 

Is there an app or way to limit FPS to 144? I'd imagine this could help with temps, or is there any benefit to the GPU pushing out more than 144fps if that's all our screens can handle? Sorry for the noob questions, just not too confident on the GPU side.

 

I've completed a repaste but still haven't touched the UV guide for the GPU and I'd now idling around 50c which is higher than before the repaste.

In some bench tests though its either taking longer to reach the throttle point or isn't where it was before the repaste.

I might open it back up again and try repasting again, it seems my idle temps are worse but, under load temps are better......

 

You will get onto a habit with this thing.. I've taken mine apart like 20 times now? Lol... I have it open and close within 25min now

Sent from my GM1917 using Tapatalk

 

I was terrified last night but it wasn't as bad as I thought haha (my first time taking it or any laptop apart). But yeah I'm starting to become obsessed with checking temps since the repaste. I think I need to just chill and enjoy it before I break something.

 

The Fujipoly pads are too stiff or too thick. Use the Artic ones since it's much better to have good contact of the heatsink with the GPU and CPU rather than better thermal transfer of the other components.

Enviado desde mi GM1913 mediante Tapatalk

 

Look up nvinspector, it allows to cap fps for individual applications.

 

I left the Dell pads on, howeve, I think the pads will be my next upgrade.

Is there any benefit to letting the GPU throw out more FPS than our screen can actually handle?

 

What fan profile are you on? The laptop should idle on high 30s low 40s on Balanced fan profile.

If your laptop doesn't do this, check the re-paste. Maybe you have bad contact between the heatsink and the CPU - GPU or not enough coverage of the paste

Enviado desde mi GM1913 mediante Tapatalk

 

I'd say it depends, for competitive titles perhaps. But even then, most of us lock framerate to refresh for consistency in mouse movement/muscle memory.

 

I always used balanced unless playing a demanding game. I think I'll give the repaste another go over the holidays. If anything I probs put a little too much being overly cautious with it being my first time. I'd have expected the viscosity of Phobya to mitigate any heatsink imperfections. As mentioned it does seem better in terms of terms under load but the idle temp just seems off.

The temps aren't a concern for now and seem better under load after the repaste, but I might play around and limit most things to 144 FPS and see if I notice any difference to temps. Thanks again bro.

 

Don't leave it because you are risking this:

 

What this can happen with my GPU hitting 50c at idle? I'll do it over the coming days then. I won't be heavy gaming before re-applying the paste.

 

That CPU had a liquid metal application that went bad due to shipping. Idle was much warmer, and temperatures under load where weird. After weeks of using the laptop like that, when I finally went around doing the re-paste... well, not much else to say.

Enviado desde mi GM1913 mediante Tapatalk

 

My CPU seems to idle around 38-40c on balanced fans, not sure how that compares to yours?

 

That's normal and OK.

Enviado desde mi GM1913 mediante Tapatalk

 

Bad how? Was it applied on a non-nickel plated copper heatsink and ended up diffusing into it?

 

The full story.

http://forum.notebookreview.com/threads/life-and-death-of-my-aorus-x7-v8.830543/

 

That link doesn't lead anywhere

 

Here... http://forum.notebookreview.com/threads/life-and-death-of-my-aorus-x7-v8.830543/

The old lady from 2015 and 25 Google tabs...

 

My GPU is now sitting around 35-40c at idle on balanced. I think the issue was that I had the Nvidia Settings set to max performance so the GPU was always showing around 1515mhz in HWInfo and never downclocking. I have it set to adaptive at the minute and clocks idle around 300mhz.