*OFFICIAL* Alienware m15 R3 / m17 R3 Owner's Lounge

How is the fan noise on m17?

Are the fans off on idle? Can you game without headphone ons balanced mode? Do fans make a whiny noise?

 

I know people don’t like 99c but it is intentional and not a sign of bad cooling par say. Not to say you can’t improve cooling but the big issue is Intel chips. Compared to pre sky lake, they are pumping out a lot more power and heat (eg some go above 90W, on a 45W chip while turbo boosted) which was unheard of before. The issue is so bad, Apple decided to go ARM. Others will likely move to Ryzen (which is offering more performance at less wattage).

Razer for example power limit the CPU to 45W sustained so they don’t go above a certain temp but then you lose performance. Truthfully, if you have adequately designed the system well, nothing is wrong with 100c on these chips. I know some manufacturers in the past (including Apple) have had VRM failures due to not adequately cooling it - but recently these have disappeared so hopefully it means most manufacturers have worked out how to minimise issues.

I’ve noticed these Alienware's can actually hold a very respectable sustained clock speed.

Judging a laptops cooling performance based on CPU temps isn’t really correct, for example there could be an ASUS which is running with same chip but at 90 while you’re is at 95-99, you may think ASUS might be cooling it better but then you realise it’s actually sustaining a far lower clock speed - so it’s an Apple and oranges comparison.

Heat and dust are seen as the enemies of computers so for longevity people might want lower temps, but I don’t think we have hard statistics to show a laptop running at 99c is much more likely to fail than one that sits at 85-90. It might be but I guess no one knows how much of a difference it makes. The true difference maker is likely to be how they have set up the components etc rather than absolute temps - so a poorly made machine which sits at 90 might be more likely to die than a really well made machine at 99.

 

It seems to mainly be Alienware touting the 99c as normal. Almost all other manufacturers are really pushing design to try to get temps down and at a decent operating temp. Doesn’t matter most people don’t keep their equipment longer than a couple years so go ahead and fry some eggs
.99c is the throttle point. Why design a machine to hover the throttle point constantly? But they design the 51m to throttle 10-12c below manufactured throttle point on gpu? Tell me Alienware just doesn’t really know what they are doing on thermals. Which it’s been that way a longtime. Owning a Alienware is fine but it is like owning a temperamental exotic sports car. You have to learn to get it under control and work on it yourself. They are for tweakers and enuthisats to a degree. Almost most enthusiasts won’t buy a welded bga ram and WiFi machine. But then get a throttling firmware limited 51m because of bad design.... Alienware is in transition and lost atm on their direction and leadership and engineering.

 

Never mind, figured it out.

 

I am pretty confident that there are some widespread issues that plague the M15 R3 line (probably even the M17 R3 too).

Having a M15 R2 since November and it constantly overheating I was sent out an engineer on 2 occasions.

• 1st occasion: Replacement of Heatsink/Fan and Mobo (CPU/GPU)
• 2nd occasion: Engineer had to return because he forgot to connect one of the daughter boards to the main Mobo. Sound, Ethernet, Aux Port and a USB did not work. During his "fix" he snapped the Ribbon Cable lock for the trackpad connector and blew the laptop up - It went POP followed by a burning smell. Pretty sure he shorted circuited something.

The laptop was finally replaced with a M15 R3 (i7 10750H/2080 Super Max-Q) unit and I have had issues since day 1.

• Regularly runs at 100 degrees C, when undervolted I can achieve 96 degrees max and that is with a cooling pad also. Never turbos past 4.2GHz however even at 100 Degrees C it stays at 4.2GHz. Temps are the least of my worries and I understand that it is "designed" to be able to run at those temps. Even with the new cooling solution it runs hotter than my m15 R2 ever did
• Graphical Artifcating/Glitching - When using the Out of the Box Video Driver or the one downloaded from the Nvidia Website there is some glitching/artefacts present. I first noticed them when the laptop was waking after coming back on from sleep. Issues mainly occurred in Chromium Edge or Microsoft Word. Literally could stress test with 3d Mark for 6 hours straight with no issues however letting the laptop sleep then waking with either the Edge or Word window being active would glitch and artefact all over the screen.
• Issue probably linked to the waking from sleep issue - When on Battery from time to time when the laptop tries to wake from sleep it doesn't and goes into a coma. Not amount of keypresses or trackpad movement will bring it back to life. Even pressing the power button has zero effect. The fans are on and the keyboard lights are on but the screen stays black. Only a hard reset by keeping the power button pressed then turning back on will bring the laptop back.

Seriously beginning to think there are some major QC issues or my luck is terrible! My M15x lasted 10 years and it was a car crash that wrote that off the M15R2 and R3 leave a lot to be desired in terms of reliability.

I had to jump through hoops of fire to get a replacement too as I was not the original purchaser of the laptop My insurance purchased it on my behalf to replace my trusty M15x so no doubt I will have to go through the engineer route again and I am 99.99% confident it will be the same engineer who blew up my R2 (1st and 2nd engineer was same one) that will come and try to fix it. Pretty sure he's the only one available for my region. I don't think a replacement will be authorised and my warranty was transferred over from my R2 to R3 and its ticking down ... honestly I just want a working laptop that will probably last me another 10 years.

Link to one of my issues: https://www.reddit.com/r/Alienware/comments/hqdnd0/alienware_m15_r3_glitchy_graphics_on_screen_wake/

other users reporting similar issues all in the last to 48 hours.

https://www.reddit.com/r/Alienware/comments/hs77hb/for_some_reason_my_discord_and_whatsapp_app_is/
https://www.reddit.com/r/Alienware/comments/hrrsoo/less_than_one_month_m15_r3/

 

I have already seen many AW 17R4 die one after one. 2 or 3 years seems as expected lifespan.

And it won't be any better for these new models once they have been pushed a couple of years... http://forum.notebookreview.com/thr...r3-owners-lounge.833030/page-16#post-11033225

 

Apple has been doing it for years and continues to do so, so did the XPS line recently with the exception of a couple of models due to surface temperature issues. Other manufacturers such as MSI also have had models reach this temperature.

Generally to not reach this temperature you either need to throttle limit it somewhere or have insane cooling which usually means a thick boy. Honestly Intel screwed everyone over. Considering most people use these machines for gaming, perhaps doing what Razer did in 2020 (limiting CPU and GPU wattage) is worth doing as nobody needs to get absolutely the maximum performance out of these laptops for gaming. It might help longevity it might not, but at least the laptop will be less hot in the end and maybe less noisy?

I think most laptop failures occur due to gpu right, hence why the gpu is usually hard limited even on the Alienware. Generally for gaming laptops in general (due to housing a hot cpu and gpu), probably worth a long warranty no matter which vendor you buy from, the relative chance of failure is just higher. With Alienware it could mean in the 3rd or 4th year your machine gets a free upgrade to a r4 .

 

Thinking about going with the 2070s over the 2080s maxq what’s your fire strike score


Sent from my iPhone using Tapatalk

 

Here is the fire strike run I just did. Keep in mind tho fire strike does not scale well. My old laptop had a desktop 1080 in it and the 2070 super gets about 2% higher graphics score in fire strike. But in time spy the graphics score is over 30% higher.

 

So I've just taken delivery of an m17 R3 (i9-10980HK/2080S) and I'm wondering if I have a CPU/pasting problem or if the thermals are just "as expected". On stock settings with thermal profile set to Performance (which I think means CPU TDP of 75W), Cinebench R20 comes up at about 3400-3500, which seems quite low given that I've seen scores elsewhere of up to 4600 (albeit undervolted). Looking at HWiNFO during the run, I can see that two of the cores are permanently thermally throttling, whilst the rest sit in the high 80s. Those two cores also periodically get grayed out in HWiNFO for a few seconds before returning. I tried running the Dell CPU stress test, which found no abnormality with the CPU, but HWiNFO reported critical temperature alerts for all cores during the test and the package temperature went up as high as 107 C. I've been having occasional random reboots, sometimes when testing undervolting but sometimes not, which usually occur at idle rather than during load.

Basically just wondering whether other people are having similar experiences here. Is my CPU defective, is it a bad paste job, or is this just par for the course?

 

Given the issues you are having it may be worth repasting. If it is an issue of heatsink fitment then it'll do the same thing afterwards. They really glob on the paste from the factory so i would be surprised if that's the problem. @Codec has the same cpu as yours i believe. He will know better than most

 

I don't think there is anyway your cpu is defective, just the paste job is most likely bad. That in combination with the fact dell/intel probably never should have put this cpu in a thin laptop. The power this chip uses is insane, using more power than some desktop cpus. I would check in HWINFO but stock mine was hitting over 100w in all core and over 120w if its an avx load. I put liquid metal on mine if you look at my past posts and it did help a good bit but this cpu will always be hot. Mine will still hit 85-90c in r20 but thats while its using ~150w and at 46-47. As far as the random reboots try using throttle stop and limiting your max turbo to 50. If i got over 50 i can get blue screens if its a heavy single core load with lots of ssd activity. I have thought about having dell change the motherboard to see if it goes away but if your has it too this may be a driver/bios/windows issue. As without the disk activity i am stable up to 54 and i have tried different ssds.

 

Oh nice! So the heatsink assembly etc is compatible with liquid metal? (I will go back and read through your past posts). TBH I'm a bit terrified about wrecking everything with the LM but I think I'd give Kryonaut a crack. Also a good tip about the max turbo, I seem to have random reboots at "idle" which may just correspond to significant disk activity.

 

Hey, I've been following yours and Codecs advice on repasting the gpu and cpu on the r3 models.

I have the m15 r3, days old with the following cpu and gpu;

- i7-10750H
- RTX 2070 super

Do you know what size pads I will need to repad my r3? Also, you mentioned removing the black tape around the HS/CPU section. Did you keep your scotch tape around the CPU dye? I'm thinking about going Kryonaught on the GPU and LM by grizzly on the CPU.

My temps are constantly hitting 100 and thermal throttling at 4.1ghz while running Aida Extreme stress test. The odd thing is AWCC shows the CPU boosting between 4.4-4.7ghz when I'm literally just sat letting the system idle. I don't understand why Aida stress test maxes the cpu to 4.1ghz but at idle, the cpu can boost pretty well and temps sit between 85-87 in hwinfo at those boost speeds.

I've managed a -100mv undervolt on both core and cache. I’ve not had the time to check whether the cpu is stable while gaming. Thanks in advance.

 

I didn't replace any of the thermal pads on mine. I did leave the kapton tape on the gold contacts on the cpu, just make sure you don't have any tape on the metal rim around the die. For me removing the black tape on the heat sink did help, it comes off easy. As far as clocks you may be hitting a power limit or just plan thermal throttling. Looking online shows your cpu should hit 4.3 under all core so if you replace your paste you should hit that.

 

@Codec I saw your sweet R20 scores after repasting, do you remember what your scores were prior to that?

 

I never wrote it down as I saw I was hitting 100c and dropping to about 4.1ghz. After LM it will stay around 4.6 with power limit getting hit and get up to about 90c before the end. I hope someone make something to control the fans on this without AWCC as the fans wont even spin up before the end of the test sometimes.

Has anyone tried making a custom fan curve in AWCC? I tried it on the system before reformatting but it just didn't work. Setting an offset of any value would just lock the fans to 75% and they would never move again until i went back to the default profile.

 

Sheesh, I feel like something is definitely wrong here then. Running R20 I'm topping out at 3.5 GHz on all cores with max CPU power draw 87 W without undervolting. With a -90 mV undervolt it manages 3.7 GHz sustained on all cores.

Edit: Just tried the AIDA stress test and all cores seem to sit around 3.4 GHz

 

Did some more tests, this time just to see what the system can do with the help of two very loud fans. Was able to hit 4.8-5.0 all core with a insane cpu power draw of 180w and a total system power draw of 270w. The r20 test will start at 5.0 but drops some towards the end of the test but Im not sure why as the highest core is 92 though out the test. It may be that I am overheating the vrm inside the laptop as if i try and push 5.1 all core the cpu power usage will spike to 210w and then the clocks will drop way down to 2.3 and stay there until the test is over.

Also tested 5.5 all core but then just ran the single tread test in r20 and it never did hit 5.5. The highest i saw was 5.4 but most of the time it was around 5.2-5.3, the temps with the fans in the single core was only about 60c with a cpu power draw of about 30-35w. Score didn't seem like it scaled with the extra clocks very well at all for the single tread. This cpu is a beast but holy crap can it use some power.

Edit: The reason for the dropping clocks is the fact that I am maxing out my 240W ac adapter. I saw when i drop to 2.3 that i am on battery power lol. A 330w seems to have no effect other than I dont drop to battery power, but it still wont let system power over about 305w according to HWINFO.






The TS setting are for the all core run.





And here is a time spy run, the first two are with the fans. Video card was at +160 core and +1100 mem. Throttle stop set to 55 55 54 54 53 52 51 and the same -90mv. Looking at the graph it maxes out around 53 it seems, never hitting the 55. Avg seems to be at 51.






And here are the same setting but this time no crazy fans and with the built in fans set to full speed. The cpu does hit 100c here right at the last reading taken by 3dmark.

 

That's insane. And amazing. I love it

 

Ok, maybe my results aren't representative of what you can expect as i just got to looking and as of now i have the record for this cpu in time spy. Never got anything like that but i think it may just be not many are playing around with this cpu, as 468 tests isnt a lot, or because of its crazy hot nature. You will have to play around with yours and see what you can do. Dont know if i have a gold chip or if no one is pushing with it. As highest cpu score was just with the fans set to max and no crazy fans going on lol.

Wait a sec, Im in 2nd the 2070 super as well it seem. Looks like I got lucky with my chips. Trying for 1st now with the gpu but at this point Im fighting for just 2-3 mhz core. At 167+ core atm.

 

Yeah that's pretty phenomenal! If only you had the 2080S, you'd be smashing it out of the park.

Meanwhile I'm managing a CPU score of 7974 at stock settings, performance mode. Up to 8375 at max CPU undervolt (75 mV).

I spoke to a Dell support rep who suggested I install the Intel Dynamic Tuning Driver, which has frankly just made things slower. Perhaps they can send a tech guy out to do a repaste and I can slip him some liquid metal.

 

If your not comfortable with putting on the LM I wouldn't ask the tech to do it. I truly doubt you would get a tech that has ever messed with LM before and if done badly it can be far worse than just paste. I had to have my 17r4 worked on 3 times before they sent me this one and all the techs said they work on office stuff more than anything. I would just get some good normal paste and ask him to use that and maybe ask him to remove that black sticker on the heatsink. Looking at pictures from @EepoSaurus his was pushing on the rim hard it looks like. And that gap will just make things far worse. I cant think of any reason to have that sticker there as even if it was to help make sure something cant short since it is copper, you would totally destroy the cpu/motherboard long before it could ever make contact with anything.

 

I’m looking to pick up an m15/m17r3 and I’m wondering if the i9 is worth it over the i7 10750H . Do both thermal throttle badly or is it only the i9 ? Also , how comparable is the performance of 2070S in the m15 to 2080S in the m17 ?

 

Unless you just have to have the extra 2 core or bit of boost I would say get the i7. From what I have seen on here the i7 is easier to cool but like all of intel's current chips, they are very hot. The fact that the i9 is unlocked is moot unless you repaste it with LM. If you like to take stuff apart and tinker with stuff the i9 is as fast as it get for now unless you get a dual power brick beast. You can see my posts for the 2070s side of things but for the 2080s i think its about 15-20% uplift but i could be wrong. Both the 15 and 17 use the max-p if you select that so there should be no performance difference between the 15 and 17 if they have the same hardware. One thing that may be a factor is if you get the 2070s you will get a 240w psu and the 2080s gets a 330w. The 330w brick is much larger if you care about that.

 

Well after your advice. I'm now happily sitting at 81 in hwinfo while running Aida64 stress test. That's without TS running beside Aida64. So totally stock voltage. I seem to be boosting to 4.2ghz. I'm guessing that's just luck of the silicon draw getting to 4.3ghz. I mean, I'm happy at 4.2ghz.

I also tested 2 different offset voltages. -100mv both core and cache, temp never goes above 75. Totally stable at that offset.

However, I settled at -60mv both core and cache as I know these 10th gen intel chips love the voltage. Temps sit nicely at 78.

I ripped the black sticker off the HS for the CPU. I masked around the dye with 3m tape, being careful to cut inside the metal surround so the tape doesn't interfere with the HS. Kronaught on the GPU - X pattern. GPU doesn't go above 51 at idle.

 

Whats your cpu power usage while in aida64? Good to seem that the cpu look tamed now. The new intel stuff is just crazy hot. Keeping an eye on amds new stuff as i saw a 4800h can score 4500 in r20 while using 45w. Intel is still on top for now but at the cost of more than twice the power usage.

 

I'm not sure brother. But I'll take a look once I'm home. Maybe you could help me dial in some good settings to possibly get some gain.

Question, do you know of any places that sell the copper heat shield for the nvme m.2 drives?

 

Since your cpu is locked I don't think you can adjust anything. For the nvme heat spreaders I would just call dell. Last time I need a part it wasn't that bad. Since the laptop is so new that may be your only option for now.

 

Any one with a m17 2080s that cans send the vbios ?


Sent from my iPhone using Tapatalk

 

Yes I just bought these for an Alienware 13 r3 that had a 970 EVO plus that was running at about 100c during gaming. Installed it yesterday (simple) and now maxing at 80c. Cheap and easy to install:

https://www.amazon.com/Copper-Heatsinks-Cooler-Laptop（2-Pack）/dp/B07XD8M43P?th=1

 

Mine the Koxia branded one from Toshiba if memory serves me well. Thanks for the link, I've been searching for these. But only ended finding plastic type HS with the thermal pad on ebay.

Hopefully this will bring my temps down a bit now.

https://ibb.co/0q7L601

https://ibb.co/T8yZ0qX

https://ibb.co/DCKng7M

https://ibb.co/xs0TyjX

https://ibb.co/5nmYjqh

I've also undervolted my gpu to 1300 clock straight line in msi. This helped drop the voltage and heat when gaming. Just played Sekiro and the cpu was at 86° max and fluctuating between 4.1-45 ghz, this was for 45 minutes play. The gpu was at 1320 constant, 60fps constant - no screen tearing. Hottest the GPU got to was 66°.

I don't think my undervolting is applying, as the voltage offset doesn't show the changes in HWInfo. I'm guessing intel locked me out.

 

Your under volt should show up in HWINFO. Its kinda finicky some times tho, like i will have to restart sometimes after changing it. Throttle stop always gets it set at boot tho. Make sure to set the undervolt on core and cache in throttle stop or it wont work.


I have just kept my cpu locked to 40 for now as it tops out around 65 this way and all the games I have tried so far it makes no difference. Playing Factorio right now and don't need all that until the base gets huge lol.

 

I posted a review of my m17 R3 on Reddit:

https://www.reddit.com/r/GamingLaptops/comments/i0lm7q/alienware_m17_r3_review_hot_or_not/

Ultimately the CPU seems to be well below par so I think the plan will be to return it and go with an Area 51m R2 instead.

 

I just got a replacement m15 in the mail today and not getting any sounds . I’m getting an error that’s says “no speaker or headphones are plugged in” any idea a fix in this ?


Sent from my iPhone using Tapatalk

 

Can’t even get the sticker right. Atleast they gave you the part you ordered. Not sure why you’re only getting around 20K on the CPU In fire strike though. Should be able to hit around 24K. Those graphics scores are very strong though. Blows any max-q machine out of the water.

What would you need the extra CPU horsepower for? The Area 51m is gonna be a lot heavier and less portable.

 

I've just managed to get off from work and fire up my laptop. Your tip worked I had to restart to enable the undervolt.

I'm starting to think that my choice of thermal paste for the GPU was a bad one. I'm hearing that Thermal Grizzly's Kryonaught paste isn't all it's cracked up to be. What is the best thermal paste for GPU's in your opinion?

 

I just use mx4. Or mx2 if its something i don't care much for, like a locked cpu. Main reason i lean to mx2/4 is that i do a lot of tinkering so i use a lot of paste and both of them can be bought in huge 65-100g tubes. Last time I tested some pastes Kryonaught was one of the bests tho. For me it and mx4 did about the same.

 

Latest LAN/WLAN/Bluetooth Drivers 06-Aug-20

Spoiler: LAN/WLAN/Bluetooth Drivers versions

Killer E2X Ethernet Driver v9.0.0.50

Killer E2K Ethernet Driver v10.038.1118.2019

Killer E3000 Ethernet Driver v10.039.212.2020

Killer E3100 Ethernet Driver v1.0.0.12

Killer 1550/1650/Intel WLAN Drivers v21.110.1

Killer 1202 WLAN Driver v10.0.0.355

Killer 1435-1535 WLAN Driver v12.0.0.927

Killer 1525 WLAN Driver v12.0.0.448

Killer 1550/1650/Intel Wireless Bluetooth Driver v21.110.0.3

Realtek LAN Driver v10.35.510.2019

Don't forget to unblock the ZIP file before unzipping it otherwise all the extracted files will have a BLOCKED status since this is a file you downloaded from the internet and Windows puts a BLOCKED status on any downloaded file which causes incomplete or erroneous installs. For how to do this and to stop this automatic blocking of downloaded files, see:
"Are you sure you want to run this file?" [Yes or No]

 

Not bad. I seen Spartan mentioned in the Area-51m r2 thread, that Kryonaught degrades real fast. He suggests Phobya I think.

Thanks as always Spartan.

 

A very stylish review. I'm surprised you didn't mention (I think, based on a skim) the soldered, and likely vwey slow, RAM as another downside of the laptop. Uneven core temps have been a bane of Alienware laptops, with the possible exception of 51m. My m15 R1 suffers from this too, repasting never really resolved it. It's very nice of AW to have offerred you an opportunity to repaste the laptop yourself, if only they didn't design it with an inverted motherboard whihh makes this process much more difficult!

 

Don't necessarily need right now but a CPU that's almost twice as fast for multithreaded tasks is great for dev work and I'm after a machine that is going to last me a few years. On reflection, the m17 R3 seemed like too much of a compromise for my use case. For most of what I'm doing a desktop replacement is fine from the point of view of being a bit heavier. When I need to be light and nimble a tablet should be able to do everything that I need. For less money in total, I've ordered an Area 51m R2 i9-10900K/2080S/32GB *plus* a new 34" 3440x1440 144 Hz screen.

Thanks! I didn't mention the RAM but from my perspective 32 GB is more than enough and Alienware don't have much of a reputation for supporting fast RAM in upgradeable machines in any case. Didn't end up repasting in the end.

 

Even if you currently think 32GB of RAM should be enough (if you changed your mind in the future for any reason, you would be out of luck with this laptop), the fact remains that the memory is likely slow, and any RAM issue will entail motherboard replacement. The former issue is uncorrectable, and the latter is very inconvenient to prohibitively costly out of warranty.

I think it is mainly the ridiculous RAM support in Area 51m R1 that could provide grounds for this "reputation" claim (OK, that and no memory overclocking support, but which major laptop brand offers such luxury?). All the earlier models up to m15 R1, supported fast RAM upgrades, and even the new 51M R2, reportedly works great with HyperX Impact RAM. It's just a singular idiotic design decision taken in the dumbed down m15 R2 and so far carried forward, unfortunately.

 

Do you know what it means if the Alienhead/power button is yellow? Does that mean it's charging perhaps or something else?

 

If it change color then it starts charging.

 

Sorry, I don't mean to sound funny, but in both my 15 R2 and m15 R1, I can make the alien head button any colour I want, and customise this for different power states. Check out the AlienFX section of AWCC.

 

Right, just wanted to make sure as I couldn't find any reference table for the power button colors like you have for the beeps. Just wanted to make sure that's normal.

 

Not sure if mentioned here but the new SKUs have the i7 10875H in place of 750H that was available before

 

The median CB20 score is a joke, right?

https://www.notebookcheck.net/Intel-Core-i7-10875H-Processor-Benchmarks-and-Specs.458968.0.html

 

Oh that's interesting. And maybe poor timing. I'd be interested in having the base spec 10875H m15/m17 R3 with an AGA and desktop 3080. I really like the idea of uncoupling CPU from GPU and only needing to cart round the former when you're on the go. I wonder how long until they're released.

Sent from my Pixel 2 XL using Tapatalk