*OFFICIAL* Alienware m15 Owner's Lounge

Wonder what his ambient temps are? Only watched it on my phone, but took screenshots to see temps in certain areas. Does seem to be high, but fans don't seem to be that loud. Wish I could understand Chinese, would be curious what he thinks? Caption says it's evil, lol...

 

How well does undervolting work with the 8750H CPUs?
How much reduction can you except to achieve?
(Reduction by 5*C with the CPU and no downclocking and things would be alright).
1500mhz GPU clock for the 1070 max-q are not too bad tbh, maybe using the msi afterburner and increase it with a 100mhz and you get a normal 1070.

 

Nice temps. And Azor/Dell talked about a possible unlocked i9-8950Hk + 1080/2080Max-Q in this Alienware model. Good luck. I mean in handle the RMA numbers

 

https://www.dell.com/support/home/u...oductCode=alienware-m15-laptop&driverId=0JFJG
I doubt there are many owners in this website, but a new BIOS for the m15 was uploaded.
Fixes & Enhancements:
Fixes:
- Addressed Legacy mode issue with Diagnostics function.

Enhancements:
- Optimized thermal function.
- Added Integrated Webcam funtion in Diagnostics.

 

Now it throttles at 82c. but stays cool! J/K

 

Yes the chip maybe can, but the surrounding supporting components cannot.

 

He updated the description of the video. It did say it was an evil unit and doing PUBG tests just for fun. Now it states: "This M15 is an engineering unit , not for sale, it's not better than the saling system, so just for fun PUBG gaming and thermal test"

Being an engineer unit makes me take some of the temps with a grain of salt. It'll run hot, though - doubt anyone expects it not to.

He's got a burn test up now.

Description of video:
This M15 is an engineering unit , not for sale, it's not better than the saling system I have done this before replace thermal past：ShintEtsu X-23-7868-2D Intel Xtu Undervolting：150mv make D-side little higher aida64 CPU，40 W cpu package，not enough，then FPU，waitng for PL2, then CPU，lasting 5 minutes furmark 1920*1080 no AA

 

Eben if its an engineering sample, i dont think the Consumer ones will behave different.
100c CPU Temp after 20s burn in... even the razer 15 runs cooler...
I am pretty curious what the first user reviews will say.
I really ask myself if they dont quality test those Laptops. I mean, they must see that it will run damn hot... and even if its „within the specification“ its absolutly unuseable because of a ****ing hot surface.

 

They test in a controlled 21C ambient temps so naturally testing goes fine in Lab and in real world it fails either way.

 

Running Furmark is not a scenario we will ever see in gaming.
It’s a little misleading to what the gaming temps actually are especially in a sample unit.
As long as it doesn’t throttle when gaming highish temps are expected.

 

But we see the throttling happening in the video where the guy is playing pubg for 40 min.


We'll see. Sadly the skepticism has been correct more often than not.

 

It sounds like the m15 runs about the same as my 15R3 did before I tuned it up. I was able to manage it some by undervolting the CPU, more aggressive fan profiles and downclocking the CPU (I cringe a bit at that, but the fact is I don't need it for my purposes and many of the games I play would run well at 30+ FPS with the CPU pegged at 1.5 Ghz or so) and limiting the FIPS to 30. I finally broke down and did a repad and repaste with LM (following iUnlock's awesome tutorial) and it runs amazing now (max settings, 60 FPS, maxing CPU/GPU temps in low 60's and hitting low/mid 70's under more extreme tests like Heaven).

While it sucks that I have to do all of this to get the laptop to function at my expectations, I don't see any value in getting mad about it and the other options have too many dealbreakers for me. This form factor really hits the sweet spot for me. I just include it in my expectations that I will have to do some tuning and that I won't be able to do a lot of overclocking, etc. This is not going to run like my M18xR2 did years ago.

My m15 will be a Christmas present for myself (my son is getting the 15R3) and my wife agreed to give me one night of testing before boxing it back up until Christmas (I'm hoping I can negotiate that with her some more). I plan to do some tests to see how things run and then do a quick repaste with ICD to test non-LM performance. I doubt images will be that helpful, but I will plan to post results to help people get a sense of how good or bad it is.

 

I hope you get a better deal with your wife. Was the LM you applied from thermal grizzly, if so would you suggest ICD over Kryonaut? Thank you.

 

Glad to help. I used Conductonaut on my 15R3. I had used Kryonaut for my first repaste and it seemed to dry out and degrade over time and the results weren't that much better over stock. I've heard some other reports of that on the forum. I'm planning on using ICD for the first non-stock pass because I've heard so many people swear by it and re-sellers recommend it and it it's also supposed to be better at dealing with poor fitting heatsinks (which I may be dealing with). I also had a Clevo years ago that I had ordered with ICD and it had good thermal performance (as much as I can remember).

 

Great, thank you, I will test ICD as well, I thought Kryonaut would have better performance

 

ICD will work with flimsy heatsink very well and temps will be great. Another alternative to ICD if die scratching(only cosmetic) bothers you is to try Coolermaster maker gel nano.

 

Thank you, I didn't know it could result in die scratching

 

The texture of the paste is very thorough and thick so it could only scratch the die but a scratched die wont effect the CPU's performance at all. Just Cosmetic like @Vasudev said. Coolermaster maker gel nano i haven't tried but heard good things from @Vasudev. I would also recommend Gelid GC Extreme. It's also a very high end paste and it's also at the top Good luck

 

ICD 7 didn't have it but ICD 24 aka bigger tube had the issue. You see even Cooler Master Maker Gel uses same composition but it is little smooth and little watery.

 

My ICD 7 from 2 different tubes also caused scratching. or more like ir rubbed some of the shinyness off the chip dies.Bit crappy that. It's so hard to remove the paste without scratching the die.

 

Which is better for these alienware laptops? ICD or Kryonaut?

 

In my testing ICD7 was much longer lasting than Kryonaut. Initial temps where the same, but Kryonaut always degraded on the CPU side because of lower pressure from the heatsink.

ICD7 just lasts and hardly degrades. I had that issue with Kryonaut in every laptop.

 

Phobya NanoGrease. More viscous than Kryonaut and better thermal conductance than both.

 

Thanks! Just bought some!

 

Good advice. That would have been my recommendation for any laptop with sloppy heat sink fit.

Excellent choice.

 

phobya...

 

That's quick.

 

And now I've bought some Phobya too.

Wow where was this **** 3 months ago when I first bought my 15R4

 

You should of read my posts in the other alienware laptop threads. Found out in June this stuff was the bees knees

 

I tested Phobya nanogrease but it would degrade quicker than ICD7. I tested 4 pastes and these is the order from best to worst for these laptops in my experience

1: ICD7, runs well for months, barely degrades. (my to go paste for many laptops past years).
2: Coolermaster makergel nano, similar to ICD7 a bit more runny and thats why it does tend to lose some of it's effectiveness after some time. Still good.
3: Phobya nanogrease, stats out strong, degrades after a month.
4: Kryonaut, doesnt last for 2 weeks, temps go up quite quickly after. (Had 2 tubes from 2 different batches)

 

Eh! Is this right? I think something went wrong. All of these should last at least 4 Month to a year. I Have used Grizly for over 6 months with no hits to performance.

@Mr. Fox @Papusan @woodzstack

Do you notice that Grizly K and Phobia NG last so short?

 

I have used Griz in Aorus X7 for 6 months now and still going like a champ.

 

You really need good heatsink fits with Grizzly Kryonaut.

 

I have found IC Diamond and Phobya to be very durable. IC Diamond is nearly indestructible, but Phobya Nanogrease does a better job for me overall. Kryonaut works extremely well, probably the best if the heat sink fit is excellent. But it is too thin for a sloppy fitting heat sink, and it is subject to rapid pump-out under those conditions. Same is true of Gelid GC Extreme, but it is even more runny and more likely to fail due to pump-out

If the heat sink fit is sloppy, IC Diamond is the only thing that works well for a long time. Using anything else is a waste of time and momey unless you fix the sloppy fit problem.

There are actually many thermal pastes that do a decent job when the heat sink or water block fits perfectly, but hardly any options that work well when the fit is poor. Sadly, that is frequently the case with notebooks.

 

My phobya is the same temps after 6 months.

 

It really depends on heatsink fit, your delta T and ambient temps.
I haven't repasted my PC for a while ~1.5-2 years and Maker Gel Nano still has good temps and I repadded with arctic pads as well to get hefty temp reduction.

 

They last long only if the heatsink is perfectly flat with strong pressure. If there is a slight gap, Kryonaut just dries out quickly. Dont forget that Kronaut is only halting the drying out process at temps up to 85c, other pastes do not. In general in my experience Kryonaut is in laptops only really good for larger chip dies but not for the small Kaby lake dies if the heatsink is slightly convex or uneven.

Because so many people kept reocmmending it, i thought I might had a tub from a bad batch. So I bought a second tube from a different shop. No change.

I tested some of these pastes while I was in Asia past summer so the temps where higher as well thus accelerating it all. 35c in japan is no joke .

 

In previous post, didn't draw the heatpipes on there in quite the right places. It was too difficult to see where they were, from such a low quality source image.

Anyhow to make a better comparison, we need another heatsink with known good thermal performance. Here again is the tongfang GK5CN5Z / GK5CN6Z ('mech g2').



Source:

To compare against the alienware m15. Unfortunately we don't have good photo of the *underside* of alienware m15 heatsink [yet]. Hopefully one of you guys will post here / help us out.

On the top side, we can see the heatpipes. But not the exact placement of the cores underneath them. Where they line up. But we can already see that the heatpipes are a bit separated though.

 

You can clearly see where the cpu die hit the heatsink cold plate. In the middle od NO-Land

 

Mine has finally shipped! EDD is still showing as 16/11/18 but I can't imagine it will take over a week from China to UK.

Will give it a hammering and let you all know temps/throttling as soon as it arrives. I have had a few 8750/1070 laptops over the last couple of months so can compare against the blade 15, omen 15, AW15r4 and 17r5

 

Thanks for showing us this. Indeed we can see it better now. So most of that die should be covered by the larger centre heatpipe. Perhaps not 2 of the opposite corners of the die. However if you look at the die shot then 6 cores are actually located in the center portion of the die. Which is what is actually generating all of the heat. Right in the middle of it. So lets assume it's fully covered, the cores themselves at least. Which is what actually matters the most.

Then the heatsink's other 2 smaller heatpipes (above and below) are both smaller than the center one. And also both of them are missing the die entirely. I can't help but think this might have been done intentionally now.

It seems that at some point in their design process, they made the assumption that they cannot be straddling the bare die halfway across 2 heatpipes. In the way that the Tongfang is doing. They must have been too scared or something. Or they could not make finer sharper edges to their heatpipes. Or whatever, it doesnt matter.

And in blind faith hoping the metal slug will also help conduct away laterally to the 2 side heatpipes, which are 100% not directly over it. Which it does, but not very much. Because laptop heatsinks are so very flat and thin. You really need some z height for the heat to be able to spread sideways. Which is a luxury that laptops simply do not have. So: single heatpipe then. As much as it really matters for heat conduction.

It can't just be only because the alienware's outer pipes are smaller than tongfangs. The ones that dont hit the die. I recon this could be checked / verified by placing thermocouples on these 3 cpu heatpipes. And measuring the relative temps. Whilst doing a cpu stress test.

Another point to make here is that the central pipe is also shared with the gpu. So a cpu ONLY test would also be worth doing. To see to what degree that might help, if anything. Probably not though. In the tongfang, only 50% of it's 2 direct contact heatpipes are shared with GPU. The other one is dedicated for CPU only. Which in some ways, makes more sense than doing one massive vapor chamber on the razer blade. Because then you have an element of isolation combined with the connection to the other half. In fact [in software] if you could make tasks prioritize to a specific core. Then you could go as far as to mark out which of the 3 physical cores which were which. By measuring the temps whilst the GPU was engaged... and use that information to prioritize which cores a game should be loaded onto first. I don't believe windows / intel's firmware is that sophisticated though. At least at the current time.

Anyhow regardless of what reasons. We have already seen in the benchmarks, that the Tongfang's approach is definately the better of the 2, in regards to CPU thermals. With the thermals of the alienware being clearly inferior. By a 'wide enough margin' to be without dispute.

In conclusion, alienware probably did a lot of hard work trying to engineer a better heatsink assembly. Than in their previous designs. However they also probably encountered many difficulties. And ultimately did not hit on the optimum design. So their marketing have decided to spin that aroung as 'new i cool or whatever we're calling it heatsink design'... 'better!'. But not really being specific about anything at all there. Because they were struggling so much to get it right. And the cpu thermals are what I would consider to be 'average', at best.

And no, that does not necessarily mean that the massive vapor chamber of the blade is not necessarily the best possible approach.

I mean... it's not super terrible performance. But it's not really any better than average. Or many other notebooks. If the thermal characteristics are similar to the MSI GS65 stealth. Then I guess it's also worth comparing to that one. As another data point. To see if the same mistakes are going on over there too. i.e. too scared to straddle heatpipes.

Here is my take on it: If you loose a little cooling along the center spine of the dies. These intel cores are still so small that you are kindda getting thermal conduction happening through the hot core itself, so heat is still coming out through the die itself laterally to the outer edge of the silicon die. Because however you look at it, it's only a few mm of lateral thermal conduction to get there. However when you have to laterally transfer heat across to an adjacent heapipe, you are asking for a lateral thermal conduction of a much larger distance, perhaps 10-20mm. Which is 5 times the lateral distance to cross. And the heatsink is so bloody thin. Then that is simply asking too much of the negligble z height. Even through pure copper (or heatpipes embedded in a bonded / welded / soldered copper slug).

In my books they need to just redo it entirely at the next possible opportunity. And go with a design that is just like the Tongfang / mech g2. There really isn't any room for a wealth of different approaches here... getting 2 heatpipes directly over probably works almost twice as well as 1 heatpipe. Well it's 2x the amount of vapour overall. Inside those pipes, ddoing the work of conducting the heat away. Or the other way to look at, is that each heatpipe is only taking heat away from half the heat. Of 3 cores instead of all 6 of them.

Maybe another approach worth considering is some massive vapor chamber, like the razer blade 15 2018, or those nvidia founders edition graphics cards. But presumably that would be more costly / expensive to manufacture. And then you cannot have a dedicated pipe going to each fan. The partial isolation / independance feature.

 

Nice!!! When did you order the laptop?

 

Congrats!
Did you get the GTX 1070MQ model?

 

Ikas - yes I ordered i7/1070/144hz/16gb/256gb in red

Cshawkeye - I ordered it on the day of release (25/10/18)

 

Awesome! Hopefully my order will ship any day as well, I ordered my the same day too.

 

Looks like you'll beat me in getting it! Congrats! Dell did call me yesterday to update that my shipping date is still 11/19 but they believe it may ship sooner. He asked if I had any concerns or anything he could help with. He later followed up with an email summarizing the call. Thought that was a nice touch considering the glitch they had earlier. Did exactly what it was supposed to, help make me feel better about the purchase. These next two weeks are going to go slow

 

Yeah the waiting is so painful. Mine is still showing delivery for the 16th but I am hoping it won’t take that long to get here

 

1080 confirmed to be coming.

https://www.notebookcheck.net/The-A...TX-1080-is-coming-in-early-2019.357821.0.html

 

I wonder what this means for the release date of the RTX 2080 in a M15...

 

It means nvidia is having serious trouble with making a mobile suited rtx 2080, seeing as even their desktop 2080s are randomly dying.