*OFFICIAL* Alienware m15 Owner's Lounge

Disconnecting the buffer battery (coin like) should reset the CMOS. (after disconnecting wait few minutes, power button pressed without power cord connected of course, to discharge eventual condenser etc...)
If it works, download the latest bios directly from dell site and install it yourself (be careful, if your battery is indeed dead, take care to your power supply connection!! you won't have the battery backing you up in case you loose power)
A piece of advice. Get rid of all the dell junk that try to install update in your stead on your machine.

 

Who has a problem with overheating of alienware m15 rtx 2080 max-q I have a solution for you!

Just a bit of story here my m15 was overheating even after repasting! I repasted it with IC diamond temps got better but not good enough!
So i found that m15 rtx 2080 is 90 watt max-q version and there are some laptops that are 80 watt version's...
As the RTX supports cross flashing vbios i flashed the Razer's 80 watt vbios version https://www.techpowerup.com/vgabios/212116/212116
and temps dropped just as i wanted now GPU is running around stable 70-75C in furmark under long run with my ambient temp of 23-25C.
This is madness that I have to flash and repaste a 3.5K$ laptop with lower clocked vbios and aftermarket thermal compound but this is a solution.

Or may be we should wait winter to use it outdoor in snow?)))

 

..that's far away from the solution, and m15 is not suffering from GPU overheating but from CPU thermal throttling. Basically you gimped your GPU for nothing. I wouldn't recommend that solution to anyone reading this forum.

edit: because i had a bad day

Beside what you have done to your GPU, what about your CPU temperatures?

Also, what about gaming? what are the combined temperatures when both your GPU & CPU are loaded? i guess you did not purchase a 3.5k$ laptop to run furmark, right?

 

I don't care is m15 suffering from gpu overheating but i'm suffering from it's heat output!
In my case gpu was running 85C if someone likes this temps it may be fine... But palmrest and keyboard and bottom panel got too hot to touch! For me it's not good i don't like such temps i don't think that laptop will last long on such temps. I still have almost same level of performance but running 75C on cpu and above 70C on gpu while gaming. No cpu undervolting done! The gpu outputs less heat so the heatsink can handle it and cpu is runnig fine.

Here everyone is discussing the cpu undervolting but no one tried to underclock gpu 2080 gives enough performance but when it's not reaching powerlimits lol))! And on stock vbios i got a gpu powerlimits reaching in stress tests. The heatsink in this laptop cant handle 2080 90 watt heat output so underclocking is a solution which works just fine. Also flashing vbios is not painful and you always can simply roll back if you don't like something.

As i see this is first alienware laptop where same heatsink is for all cards and cpu's(2060 / 2070 / 2080 / i7 / i9)... Previous generations had different heatsinks designed for each gpu to handle heat output but this days are gone.

 

Power limiting your GPU is not really the best solution, because CPU is always limiting factor in laptops. Clearly your mounting pressure or thermal past apply was not good enough. I get 73C with my RTX 2060 at 90W (with stock paste).

 

If you look at the two motherboards you can clearly see that m15 with 2060 is not 90W as it has less power rails to the gpu also missing couple memory chips so it will produce less heat as it uses less power and has less chips to output heat. And i talked about RTX 2080 max-q version of this laptop with i7 8750. Also I found a review of m15 with 2060 and it shows gpu temp around 85C and cpu around 100C with throttling under load. And your saying 73C on stock paste under load which sounds funny. Any way it's good result if your laptop can handle that heat!

 

Yet the m15 rtx 2060 clearly is a 90W version. Also it should not matter if you have 2060 or 2080. 90W is 90W. It should generate the same amount of heat so something is wrong with your laptop.

 

Bizarre. This is the first report of a GPU in the m15 R1 overheating. Try repasting. My 2070 runs at 65C max under full load (Kryonaut), @CptXabaras 's 2080 was even cooler after IC Diamond repaste. There are several 2060 variant owners here and nobody had complained about GPU temps before. In my case the improvement from repaste was about 10C so clearly Dell toothpaste needs to be replaced at the earliest opportunity.

Loading a gimped 80W Razer BIOS is the last thing I'd want to do, if anything, it would be good to get hold of a 100 or 110W BIOS.

 

I totally second this and that would be really, really lovely! given the temperatures at 90W, it should totally be able to handle the extra wattage

 

I have a 9750 + 2060 and of course the GPU is sitting around 65 under maximum load. When I stress the CPU too the GPU will get hotter but just because of the shared heatpipes transferring heat from the CPU to the GPU.

I got the 2060 because I usually game using the AGA with a 1080 Ti and Dell pretended to charge me $400 to go from the 2060 to the 2070 maxQ, and didn't even offered the 2080 maxQ with the 9750, but I'm pretty sure that even the 2080 should not be the blocker in this laptop.

@code9523 Check your pads and make sure that your heatsink is making good contact with the the GPU. Check how much paste is on it after removing the heatsink. There should be almost no paste. I would return my laptop if I have to flash a BIOS from a Razer into my Alienware ... really, that's heresy!

 

totally spot on and agreed with kahuna0k,
@code9523
To put things into perspective have a look a those screenshot below:

This is a firestrike run, taking it as an example since on the graphics test it uses mostly the GPU. Check the temperatures. CPU temps i don't take them as a metric to judge my CPU temps because the test is so short that it really doesn't manage to push the heat dissipation system.


this is a 4h run of Prime95 small fft, performance fan profile on AWCC. here only the CPU is being pushed hard. Check temps.


Finally, a real usage scenario (at least for the purpose i've purchased and use this laptop for, gaming)
Almost 3 hours run on Divinity, Original Sin 2 Definitive edition. Balanced Fan profile. Here you will see how the (stupid) shared heat pipe design affect the CPU&GPU components. I've started HWinfo just after starting the game, so the average temperatures recorded reflect the 2:40m gaming time.


You should really check what Kahuna0k is suggesting about your paste job, an do that poor little thing a favour, and flash back that GPU original Vbios.

 

How were you able to keep the pkg power at those levels with a PL1 of 75W? Were you running mprime with AVX off? Small FFT should use more than 51W. Unless you are doing what I'm doing limiting the current on the cpu with iccmax

 

iccmax is set at 140A, the reading from HWinfo is inaccurate or, better explained, it takes the wattage from the AWCC performance profile which is just a setting, the upper limit of a range in which the CPU is allowed to work at. But given the voltage (fixed) applied to the CPU, the real value is whats shows up on Throttlestop (that's an actual reading that is within the range of the used machine profile).
Oh and PL1 settings on Throttlestop does work, i tested it while running Prime95, changing it live and seeing it changing accordingly on TS windows. To give you an example, with PL1 set at 45W on TS, Prime95, with my voltage settings, will run the cpu at 45W and boost the 8750H at 3.7Ghz. Then from there, i've started increasing PL1 on TS, 50W then 55W and saw that TS with 55W on PL1, was showing the value you have seen on the screenshot <55W and frequency increased accordingly with each increment i did on PL1 in TS, up to the advertised 3.9Ghz on 6 cores.

For that run i've given a fixed cpu voltage of 0,8203 (slightly lower then yours, for testing, i keep it, for that profile at 0,8503 plus what we saw being apparently LLC applying extra voltage, so let say 1.015V aprox) Given that, pushing the iccmax at 140A does nothing imho, because the Amps at that voltage and wattage will be around 51A.

Edit: forgot to mention, yes i had AVX off

Edit2: and i finally found a use for the m15 macro keys. I mapped so that each one correspond to each profile on Throttlestop. Really convenient if you ask me.

 

That explains it. I was looking at the pkg pwr as reported by TS. In my experience it is pretty accurate. When I run mprime small FFT with AVX2 and FMA3 I get to 85W quickly, and temps are a "bit" higher

 

Yep i guess that much

Didn't want to run it with AVX active on a laptop (that has a barely minimum thermal dissipation system that require us, enthusiast, a lot of tinkering and tuning to make it work how it was supposed to be, to begin with) that i use only for gaming , and multimedia content/internet browsing, i reserve that pleasure for my , Overclocked, Overvolted & liquid-cooled desktop

 

Hey guys, new owner here, i just got a m15 r1 8750h + 1070 max q, i have a few questions, my psu is a 180w, should i buy another one? If yes, wich model should i buy? I foubd the dell 330w in my city, but i dont know if i can use that one or the 240w. In my country i could not find the version with 144hz, i got the 1080p 60hz, is it possible to replace the screen? I will be playing cs go most of the time. I have a 240hz Alienware monitor, at home will be perfect, but i travel a lot.

Best regards

 

240W is enough.

 

Have you found a solution to not being able to install the most recent Nvidia drivers for the 1660ti? I have used DDU to uninstall and have tried to install DCH and non-DCH drivers from Nvidia's website, through Geforce Experience, and by using a modded inf and none of these methods have worked. It's frustrating to have to use Dell's drivers since they are months behind. Has anyone else had issues trying to update the 1660ti drivers?

 

I have never had an issue installing Nvidia drivers on an Alienware laptop, even without DDU. What error are you getting specifically?

 

You get those kind of issues sometimes but not always. Nvidia installer unexpectedly quits saying no nvidia card detected.

 

I get the same message that @hawk1410 got on the attached image. As far as I know, no one with the m15 and 1660ti can update their drivers from Nvidia. I have less than a week to return this so I have been trying to find a solution to this.

 

Have you tried installing the drivers via the GeForce experience app?

 

I had the same issue on my RTX 2080, that was with the DCH driver, then I tried the standard driver and it works although at first it installed all components but it failed to install the USB-C Driver, I tried again without even rebooting and this time it was able to install everything.

 

I have tried installing by using DDU first to uninstall drivers, tried without using DDU first, tried using both DCH and non-DCH drivers from the Nvidia site, tried using a modded inf, and tried from Geforce Experence. Geforce Experience will just tell me that the drivers are up to date even though Dell's drivers are 431.xx.

Edit: I forgot to add if I uninstall the 1660ti then let Windows reinstall it, Windows will install drivers that are even more out of date than Dell's drivers.

 

Sometimes there’s some type of blocking code on the the default driver that prevents using drivers directly from nVidia. Usually DDU uninstall takes that out along with the initial OEM modified driver, allowing you to install the standard nVidia driver. That’s been my experience.

With DCH things get complicated sometimes as even Intel (in regards to their own drivers) has previously mentioned going between DCH and standard can leave residuals that may cause issues in certain cases as it’s not necessarily a clean switch. Maybe DDU isn’t properly getting rid of the residual files.

Worst case option is a Windows fresh install without internet to prevent auto install of the OEM modified driver. Then try to install the default driver before connecting to internet or block auto driver installation. Or at least try installing Windows to a external drive or something to see if this works if you don’t feel like losing your main install.

Lastly at times vendors do use modified cards meaning you are out of luck with standard drivers, that was the case with one of the older Alienware models around the GTX 800 series GPU era. It may have been the GTX880 containing model. Hopefully this isn’t a repeat of this case for you.

 

I decided to stick with my 13r3 and to return this. Not being able to update the video drivers is a deal breaker. Maybe someday when they have another sale I will buy another one with a 2060.

 

so guys wanted to ask how I can further decrease temps on this machine (alienware m15 r1) I have repasted it, raised the back and undervolted by -100mv. How I lower the processor core ICCmax? around how much if anyone has experience? Currently using XTU cause I can't overclock the intel gpu from throttlestop, currently trying to run some experiments with optimus.

 

You can undervolt more if your processor can take it. Mine is at - 150mv and I'm going to test lower tomorrow.

What paste did you use?

What are your peak temps right now?

Edit: Fix Typo (What is undervilt?)

 

Just to chime in, there's easier ways to lower your GPU's power consumption - you can use MSI Afterburner to undervolt your GPU by "flat-lining" the power curve at whatever OC tuner shows is acceptable at the lowest power draw. For example I have my 2070 Max-Q at 1475MHz @ 0.8V, which per HWinfo reduces power consumption from ~92W to 77W. This (in addition to undervolting and underclocking the CPU) is very helpful in keeping the laptop's total power draw <200W, which is the limit on many airline/train circuits.

Also, I'd like to point out that your GPU is by and far the largest power consumption/heat source in the laptop. Since the CPU and GPU heatsinks are joined you can sometimes get the best performance by preferentially boosting performance on one or the other since your total draw is always limited <240W (unless you bought a larger power brick). E.g., in PUBG I give my CPU the preference since it's a CPU heavy game to stay @ 3.9GHz, and my 2070's overclock is power optimized (OC tuner curve, generally 1500-1550MHz, no RAM overclock); which gives me <100FPS @ 1080p. In CPU light/graphics heavy games, e.g. Metro Exodus, I'll actually put my CPU SST to 0 in throttlestop which keeps it at ~2.9GHz and then crank my GPU to its max stable (generally ~1600MHz) as well as overclock the RAM +500MHz. It's all about optimizing your power budget. Note that temps play a big roll here, as the lower the temps the less leakage and further you can push everything.

I'd recommend using ThrottleStop as it has better control and plays nice with other application, but it doesn't matter as long as you get the results you're after.

Every CPU is different, I'd say load up a CPU intensive application or game and keep it running while progressively lowering the voltage (i.e. increasing the undervolt). For example, I loaded up PUBG and kept playing while lower my undervolt 10mV at a time. The game froze at -175mV, so on the daily I run -125mV, which has been rock solid in all applications and testing. An ~50mV stability buffer is what I'd recommend, but others may feel differently. You can also lower the voltage on your Intel graphics core although not by that much. Most resources I've found say -50mV is conservative but safe. Also note that overclocking your intel GPU is going to slightly increase power consumption for your CPU. Also ICC max is just your max current draw, which will just cause your CPU to throttle sooner.

 

It looks like Dell released bios version 2.2.1 on 9-12, anyone try it yet?

Fixes
- Fixed the issue where special characters do not work in BIOS Update Utility when user sets admin or system password using special characters.
- Fixed the issue where the system displays a black screen while updating the BIOS after a fan error occurs.

Enhancements:
- Improved the system fan performance.

 

Thanks for the heads up, i'll guinea pig it

Edit: done, laptop did not explode or else, everything looks normal, this evening i'll check out if there is any difference with fans profile while gaming.

 

I updated from 2.1.1. AWCC still does not seem to respond to any fan profile settings. The fans might be quieter at idle (they rarely dropped below 50% on 2.1.1) but I need to spend more time with it. Not any worse at least.

 

I get same temps and results in benchmarks. Try to run aida64 stability test with only stress fpu checked for at least 20 minutes and furmark on full screen FHD for same 20 minutes.(without any undervolting etc just let it run stock!) And you can see funny things. By the way yours cpu minimal temps are really low what is your ambient temp?

 

Ambient temp somewhere around 22 & 24 °C. I can try Aida64, but i honestly don't like to run furmark at all on any laptop, it just put unnecessary strain on the GPU in my opinion.

Edit: regarding the funny things you are refering to, i'm totally aware, the m15 without undervolt give so much voltage to the CPU that, with those same settings, i could run my watercooled 8700k at home at 5.2Ghz on all cores! . this laptop needs both a good repaste and a nice undervolt to work as it should be.

Well i tested yesterday with this new bios, almost 3 hours of Divinity original sin 2 enhanced edition, on my usual balanced profile from AWCC. What i notice, and didn't like, is that indeed the fans take way longer to ramp up on balanced profile and they did not go full speed to catch up with the temperatures, giving me higher max and average temperatures than the previous bios version. Mind you, i didn't try anything else other than that, and it could have been a glitch or something. Will do some more testing if and when i'll have more time, but from the look of it, i think i'll revert back to previous bios version. Looks like the so called "fan profile enhancement" has been only to make them less noisy, not to improve the cooling performance of the machine. Nicely done Dell, another fail.

On a side note.... how do i roll back to previous 2.1.1 bios? i've tried with the ctrl + esc method (power cord disconnected) then connecting the power cord still keeping the above key pressed and release them once the machine powered up. Got into the option for the rollback. The machine did the procedure but i'm still on 2.2.1.
Am i missing something?

Edit: never mind, as simple as it could get, just run the 2.1.1 installer, and it will flash back the previous bios version. All is good and back as it was before 2.2.1.

So, 2.2.1 does indeed put the fans completely at idle and let them ramp up way less aggressively that it was on 2.1.1 making the machine more silent, but i prefere the previous version over the new one.

Edit 2:

with 2.2.1, CB20 run, balanced fan profile, no optimization for the run (software running in back ground, hwinfo64 running, no priority changed for CB20 etc...), my fixed voltage profile in TS: score 2777.

with 2.1.1, CB20 run, balanced fan profile, all condition as above: score 3021.

gonna stick with 2.1.1

 

I see this too. Too much voltage on cpu. I'm now looking what to do with all this voltage...

 

Get rid of it, as much as you can, also try to get a profile with static voltage, it works really well for under load conditions and allow you to go down further than with adapatative and off-set voltage

check out @kahuna0k post on this thread for refrence, kudos to him
*OFFICIAL* Alienware m15 Owner's Lounge

And here some testing i've done after Kahuna0k findings, plus some TS settings that might be able to help you out
*OFFICIAL* Alienware m15 Owner's Lounge

 

nVIDIA GeForce RTX 1660Ti/2060/2070/2080 Driver v26.21.14.3169 ,A03

Finally!! Dell gives us an updated nVIDIA Driver for the Alienware Area-51m, from the description, it seems like many other GPUs are also supported for Dell/Alienware Laptops in this release:

Download

 

Upon which Nvidia driver version is based this release?

 

431.30

 

Excuse me..... using a vapor chamber is far from a gimmick and can be superior compared to heatpipes filled with flux relying on a phase change. Not to mention, Clevo could only sell the p870TM1-G for examaple because of the vapor chamber they use. Else they wouldn't be able to cool two 1080's running @150 watt, let alone sell the laptop. that's is the reason why the 51m struggles @1080p, they used only pipes.

That results in the heat bleading the system faster than the pipes can suck up the heat. Vapor chambers are better cause the whole chamber is situated above the DIE where the temperature hotspot usually is. Now that the Die's are getting smaller, the HS should spread the heat better else it will create new problems. A vapor chamber is superior to pipes. Pipes should only be used to transport the heat, not cool the damn thing.

 

After seeing Blade 15 advanced vapor chamber and base Blade 15 pipe heatsink (from notebookcheck review on both 2018 models and real users in several Razer forums), I don’t see Advanced model being any better thermally. Vapor chamber is used to spread the heat exchange across large surface area, but in laptop you're limited by tiny heatsinks, so the main engine of heat exchange - the phase change, still occurs in the small area above the heatsink, the size of a heat pipe, and the whole giant vapor chamber area is wasted. And the cooling system in the Base, consisting of just couple of heatpipes, is just amazingly efficient.

 

A Vapor Chamber ain't worth so much if the rest of the cooling as Grills and Fans can't throw out the heat from the chassis.

Pipes have never been used to cool down hardware in any laptops...
Heat pipe - Wikipedia

 

Yes Blade use also the cassis and users hand to transfer heat

 

Can anyone please share the dumped by gpu-z vbios of RTX 2080 max-q from m15? Thank You!

 

Herewith attached.

Out of curiosity, what are you going to do with it?

 

Toshiba Laptop C55-a is one of the laptops that uses a heatpipe (without coldplate) to cool i.c.m. with the fan: https://linustechtips.com/main/uplo...ides.PNG.b69c841835610299b1a9f7405a114ec4.PNG

and there's more of these fails. Just 1 pipe and that's it. That is true, that's why I used clevo 's example with sli enabled. Those do not even use pipes but the chamber is the heatsink itself.

Pipes are as good as your fans can cool down the radiator, look at this picture:

There are only two pipes and those are for an external radiator but this heatsink cools better than a coldplate/heatpipe/radiator unit where as this is just the vapor chamber body directly connected to where the fans blow air to create the phase change. Because it's more of a unit instead of 3 sandwiched parts the cooling with a VC in small chassis will be alot better.

That's also demonstrated by the razor blade pro with 2080 max-q which performs almost 15% better than the m17 also with max-q but without a vapor chamber. Although the m17 can outperform the razor blade, it does not.

While I do agree with @Joikansai that it's efficient, it's not so compared to a good executed vapor chamber/fan design (without radiator) because the vapor chamber has more surface area than the normal coldplate, it could be used better for cooling different parts not just the CPU and CPU. Look at the picture of the Clevo heatsink design. There is no radiator present only two fans which do the phase change at the back of the VC.

If the whole PCH, RAM, SSD can be covered by the VC you could create desktop like cooling for your VRM's etc. In desktops you COULD use air but in laptops there is something needed that extracts the heat from the components due to hot air that bleeds through the chassis. With a vapor chamber that air will have less if a chance because the VC cools a large area if the mobo itself (if done correctly).

 

Thank You for the dump!

I just wanted to compare my vbios with someone else. I repasted it with LM conductonaut i done undervolt for -150mv the systems scores 18+k in firestrike but temps are still pretty high. The CPU reached for 85c and GPU was bouncing in arround 83-85C in firestrike. Ambient temp is 22c. And laptops idles around 55-60c on cpu and 50-55c gpu (google chrome). I'm thinking may be i have a defective heatsink as contact between the copper and die is superb i checked it couple times and laptop still cant output all the heat. In PUBG it goes to 90C on CPU and 85C on gpu and starts dropping frames down. I can only play if i will turn on cool profile in awcc but it will decrease performance of laptop.

 

Guide: Killer Software / Extension Un-installation and Blocking

 

Does anybody know the part number for the GPU fan? While cleaning with compressed air after some tweaking I failed to secure the fan and it seems that I've unbalanced it, and now it makes a very quiet rattling noise when turning at very slow speed. It is only perceptible at night and with complete silence, but my OCD is getting me crazy

 

So, I updated the overclock controls in the AWCC this morning, should have known better.

Windows defender went nuts and indicated a Trojan:Win32/Fuery.B!cl
amsi: C:\Program Files\Alienware\Alienware Command Center\OCControlService\OCControl.Service.exe

I uninstalled the latest, re-installed the previous and all seems well now, but seriously, this awcc stuff just screams amateur.