ALIENWARE m15 Ryzen 5900HX | RTX 3070 - [GAME TEMPS]

Here are some initial out of box gaming temps playing Overwatch. The GPU temps aren't too shabby I must say given its form factor and it being bone stock. That heat sink is doing a ton of work in trying to also cool 8-cores. Tough task as we should have all expected.

This is going to be very interesting between the Ryzen and TG model. It's nice that there options now to choose from...

There is a lot more testing to be done and I'll release them as quickly as I can.

Latest GPU Driver at the time w/ Fresh W10 installation.


Original GPU Driver. Out of box OS etc..


Now time to open it up and try to work some magic...

*I'm reserving my final thoughts until I can thoroughly test both variants. It's still very early and I'm in the process of capturing all the raw data so that I can make it available for you guys.

 

w/ Traditional Paste:

 

w/ Liquid Metal:

 

Pics 3

 

After Pics

 

For giggles... thought I'd just post this here to share.

This is a bone stock Fire Strike run with the gpu driver that it shipped with. For some reason it was acting wonky, but for the sake of the combined score, not too shabby. 13,344.

I'll be running some more benches with a newer driver soon...

5900HX / 3070 Laptop

 

Classic Dell 100C out of the box at any given place. I think Apple M1X or M2 will be running at 120C on dell XPS/AWs even with Liq. Helium cooling 2.0.

 

Thanks for the info on out of the box temps.

My toughts: GPU temps seems fine and even if Time Spy score is good on both CPU/GPU i am still very upset to see that dell sells an overheathing/thermalthrottling laptops like that. In games it will thermal throttle like crazy and even GPU will suffer (my guess on shared heatsink).

Not everyone can repaste/repad, undervoltt etc....to have this gaming laptop doing what it's suppose to do: «Gaming.»
Anyone who buys this laptop for gaming will be disapointed. I understand you will work your magic on this machine: repasting/repading??? not sure but....still i doubt.

When you repaste i suggest you give a try to Thermalright TFX. That paste is extremly thick and cannot be spread like Kryonaut. The X patern seems to be the best way to apply this paste. I am not saying this paste will be the best for this laptop. But if you test with thick and soft paste then TFX is the thickiest of all.

 

No problem. I've been gaming on it in stock form and it's not as terrible as it may seem in relation to the cpu thermals. Of course the cpu runs hot and so does other laptops in this class. They all need some work to some degree.. some more than others.

I've worked on quite a few systems over the years. I'm just reserving my full opinion until I know for sure instead of speculating. .. not expecting miracles, but there's room for improvement with the cpu temps, because one thing that we know for sure is that stock factory paste from any of the brands are not adequate, cheap and break down fast.

 

The 3070 in the ryzen m15 definitely took a hit with older drivers and a lower TGP. My 3070 in the m17 r4 hits just under 32K in firestrike graphics. Looking forward to see what your TGP is in benchmarks and how far you can clock it

 

HeHe. Yeah, Dell continue keep up with their 100C traditions. I wonder when they have the time add in the m15 Ryzen model into their Tech support webpage to try calm down concerned owners/buyers https://www.dell.com/support/kbdoc/...nware-notebooks-purchased-after-2016#High_CPU

M15 R5 (3070, 5900hx, QHD) Thermal camera view

 

You don't need to introduce yourself. You are well known since your repaste guide with 17R2/R3. So i (we all ) know that you can repaste and tweak a laptop to it's limit but....this time....HUM....i don't feel it. Also they reduce TGP of the 3070 (like @HaloGod2012 just mentionned) but still CPU (shared heatsink) core speed seems all over the place because ot temps.

Anyway good luck at making it shine but to me this design is destin to fail. Even with your skills.

 

They had to because they traded the Vapor chamber for the 2 ram slots You can't have it both ways as you know.

Not so sure people will be so very happy with 100C. Some accept it and some move on as he here Re: 17 R5, OEM mobo replacement

 

Good find @Papusan. I was once told by DONOVAN (supervisor at Alienware Costa RIca) that themperatures were normal. He then showed me the exact same page you just linked to us.
But please pay attention to this sentence: « The TCC is activated and adjusts the CPU frequency down by a couple of hundred MHz to reduce the temperature rise. This allows the system’s thermal control mechanism to catch up»
Now what exactly is a: «couple of undred MHz down»
On the area 51M R1 it can go down form 2100MHz to 500MHz!!! I have experienced it. In FPS (Frame per second) while gaming, it translate in going from 70FPS to 15FPS!!!
15 FPS becomes unplayable for a gaming machine.

So if a custumer ever calls dell Alienware and complain about terrible drops in FPS: Dell has this webpage they can show them to explain this behavior is normal and the system is working as intended (intended = themal throttling).

 

Fire and smoke right after expired warranty is probably worse than throttling and less FPS due the same culprit (100C is just fine). Can become an awful experience if you let the machine run at nights while you sleep

 

Well you’re wrong there my friend. I got into a spat with someone on Reddit that was telling people 100C is normal and totally fine. So there are plenty of bafoons out there that don’t think it’s an issue.

 

Maybe they just like their CPU's the way they like their Women? HOT!

 

Hope they’re ok with the inevitable burnout then. This is why I prefer cute women.

 

Maybe some of the reddit boys/girls just follow what they have learned from Dell tech support? The first I started think about when I saw this article was Dell's statement that 100C is just fine.


Watch Intel's Rocket Lake CPU Burn Up Before Dying

 

There are sheep everywhere man. . .can’t help that some people can’t think for themselves.

 

Thanks @Papusan for the great link/video.

This is the future; 100C all day all the time. Manufacturer couldn't deliver good cooling system but a genious engeneer somewhere had this great idea: «people (customer) are stupid so it will be easy for us to make them believe 100C is normal. This way we make more money»
The problem here is that the engeneer was.....................................right!!!

 

Thanks. I'm fully aware of all the limitations and physics in general, hence why I had mentioned that I don't expect a miracle. The point that I was trying to make is that what comes from the factory across all laptops when it comes to the thermal paste etc... is not adequate and usually very cheap quality with poor contact. I'm most curious to see in these 8 core thin and light models as to how much improvement, even if it's small that can be had. That's all.

 

It still baffles me to this day how some can not see how they are being played by these outrageous temp numbers and thinking that it's normal.

The sad truth and reality is that most people don't even check and have no idea what's going on. This is the herd that becomes the justification for why companies continue to release 100C+ running laptops and somewhat get away with it.

What's even more discouraging is how the recent bios update raised the temp to 105C. I guess that 5C is seen as an improvement? Unreal...

 

Agree. If anyone can improve those system it's you.
But 100C is here to stay... i think,

Thanks for your testing.

 

I'll try my best. With Ryzen mobile it basically boils down to hardware as it lacks anything Intel mobile has on the software side.

With more and more people getting their Ryzen based units, I think the reality has sunk in that AMD was not the saving grace that team red was hoping for. I appreciate the fact that there are options now with different flavors to accommodate each camp, but now that the truth can easily be seen with Ryzen mobile under delivering, I foresee this turning into a blaming game called: Who to blame next?

What Ryzen mobile has going for it is that it sips fuel and can be pretty efficient. However, if you tone down an Intel mobile chip to the levels of Ryzen mobile you can achieve the same lol...

I'll be releasing a lot of data soon...

 

I’m still pretty impressed with how far Ryzen has come in such a short amount of time. And that alone would earn my money. But, even when the first gen Asus G14 was new @B0B pointed out a lot of limitations of the cpu from a tweakers point of view and that once you start putting high end GPU’s in these laptops that they were gonna run hot. And he was 100% right. And with no throttle stop or Intel XTU alternative to undervolt we’re kinda at the mercy of manufactures making good cooling solutions. Which sadly a fair amount of them don’t seem to be doing that.

But, I’m glad we have a choice atleast. And honestly that’s never a bad thing.

 

Going with an AMD solution has some similarities with going down the Linux road. There is an immediate sense of gratification to the extent that you are "sticking it to the man" in a very real sense, but you'll have to resign yourself to the idea of settling for whatever you can get. And, there are a lot of things that you'll do without because there are a lot of things that simply do not exist in the alternative dimension. In some cases, the substitutionary solutions that do exist are half-baked, less than elegant, or too broken to be taken seriously. That might be OK for some people in some scenarios. For other people or different scenarios, it's really not OK.

 

Beautifully explained.

 

I was in the middle of going all out and doing what I normally do with LM etc... however, I stopped myself before applying that shiny little silver rain drop on the CPU and GPU die, because I figured it would be a good opportunity to show what an average non-conductive paste could do.

Over the years I've collected a nice shoe box full of those little thermal paste tubes that come with GPU water blocks. The one that stood out as being likely the best candidate was GELid so I decided to go with that.

Upon testing the contact pressure on both the CPU and GPU, it was surprisingly pretty good with it a little light on the edges. ie... usually when testing the contact between the die and cold plate it's usually heavy (pressure) on one side and light on the other in a gradient fashion. However, with the m15 and the way the heat sink sits on top of the board, it's pretty secure/flat and actually provides good contact.

For those of you repasting it should work out pretty well for you.

Other than applying the thermal paste, doing a minor adjustment on one of the ends of the heat sink and adjusting the thermal pads (which are pretty decent btw) here is a 1:1 test with playing Overwatch on Epic Settings, same maps, etc...

I'll provide the original out of box screen shot here back to back so that it's easier to compare.

Repasted w/ GELid:


You'll notice that the GPU boosted a tad higher and was less as expected.

The overall averages are pretty good and again as expected. What I mean by "as expected," is that we know how bad stock thermal paste can be. Therefore, this is not surprising at all for a lot of us that has been flying for a while...

The GPU die has more pressure than the CPU side so it's not surprising that the Max GPU temp peak is the same. However, if you look at the average it ran a lot cooler with some pretty respectable numbers.

The CPU's average temp isn't too shabby either with it drawing just a tad bit more power, yet staying relatively cooler compared to stock.

You'll also noticed that the overall CPU clocks were much higher with it repasted, again as expected... "cooler."

Bone Stock Out of Box:


Now time to go onto LM'ing etc... but I hope this gives a quick general idea at to show that there are improvements that can be had, granted you're able to preform a repaste. More to come on that...

Note: After LM'ing I plan to do a 3 game back to back to back (OW, Flight SIM, and COD) test consisting of ~2 hours or so of game play to let the system fully saturate to its peak, along with other realistic CPU and GPU benchmarks. From my past experience, when a laptop is completely saturated (the peak of the heat sinks limits) the average can range from anywhere between +3C to +5C in general so keep that in mind.

Overall, although this is traditional paste, it's a nice precursor to see the potential benefits of LM, which could possibly chip a tad bit more off of the peak... only one way to find out, but I'm pretty confident that there will be some improvements even though if it's small.

I'll keep you all posted.

 

oops post.

 

Not too shabby indeed. Not the greatest but not pegging to 100C all the time is a welcome sight. Does Alienware let you limit the tdp of the CPU at all? Seen a few laptops let you do that. That would be a welcome thing to control since Alienware seems to like forcing a lot power into their CPUs.

How was the tear down? Hoping it’s not too terrible since the motherboard has to be removed to be a repaste. So a few pics would be appreciated.

 

Ryzen 5000 series Cpu in the Alienware m15 Ryzen Edition R5 don't do it so well vs. previous gen 4000 Ryzen mobile.
Even beaten by Dell's own G branded models with the older Ryzen mobile
http://forum.notebookreview.com/thr...r5-owners-lounge.835745/page-14#post-11096208

 

Great info thanks.

So your CPU improved 4 bin on maximum but max temps still close to 100C at 97.5C and thermal throttling. I am affraid LM will improved your CPU bin again but temps will stay high.
Would be nice if you ran CBR20, only for CPU testing and showed us temps and thermal trhottling info.

 

The tear down itself is pretty average, but the part that needs a lot of attention is the intricacy that goes into the overall process. I'll have details on this soon. As for the awcc it's absent of anything useful cpu control wise. You could use ryzen controller, but that's more of an imaginary GUI in this laptop with ryzen mobile to be honest. It's meh...

I'll have benches up very soon now that it's finally LM'ed! Perfect timing.

 

Here's the moment a lot of you have been waiting for. Thanks for being so patient.

I'll post all three screenshots here so that it can tell the story.

Stock vs Traditional Paste vs Liquid Metal.

As many of you already know, Liquid Metal is the only way to bring out the peak potential in a laptop. It's great stuff, however, it's not for everyone and I highly recommend doing a lot of research for those who are new to repasting in general. If you're completely new to this stuff I would advise staying away from liquid metal all together and sticking with traditional paste. Now with all of that out of the way let's talk numbers.

As you can see from all three graphs the improvements scale nicely:

• The overall averages are higher.
• GPU maintains higher clocks.
• CPU Max is much higher.
• 26.1% drop in temps for the CPU Avg, which is right about where I expected it from past data from the many LM'ed laptops that I've performed.

The overall gaming experience felt much better due to there not being any stuttering, skips or jitters like it was before at certain times during heavy scenes.

Now with the thermals in check, this changes my perspective completely about Ryzen mobile, despite the lack of any tuning controls. I had mentioned before that I was reserving my full opinion on Ryzen mobile until I can actually test, validate and verify for myself. For the given chassis, non vapor chamber heat sink, smaller fans and with an 8-core CPU paired with a laptop 3070, the best case scenario numbers are pretty respectable.

Now that I've had the opportunity to drive all three scenarios, my final thoughts for the main review are shaping up pretty nicely.

If you have any questions I'll do my best to try and reply asap.

Thanks guys. Enjoy...

w/ Liquid Metal:


w/ Traditional Paste:


Bone Stock out of Box:

 

You sure did a good job improving temps on this laptop as we all expected. Pressure paper might helped (did you bent the heatsink for better contact or did you applied more torque to some screw, lapping etc... to improve contact?) but i find that your LM really only improved your GPU temps over the previous GElID repaste you just made. In fact, i think your biggest improvement on the CPU was achieve with your first repaste with GELID and i find it confusing to all i have ever experienced with repasting a GPU.
In general the GPU rarely benefits from LM over a good non-conductive thermal paste. On the Area 51M it performs worst. But on your system it improved temps by 7C on max GPU temp and 1C on the average. (i personnally consider that « Max temp» is the most important temp to pay attention to for thermal throttling and performance drop issue, »Average Temp» being secondly important for overall comfort of the user).

I also question on how demanding is the game: «Overwatch» How consistent is the power requested for a gaming session and does all 25 minutes gaming session requires the same amount power?
We certainly need more benchmark to conclude anyting on this sytem. The CPU remains a big question here. Why does it throttles all the time? Is it normal behavior for AMD? Can it sustian all 8 cores at max clock speed like Intel?

So far i would say it prooves 2 thinkgs:
1- Dell's paste job is really, really bad and it's disgrace to ship such an expensive system to a normal user.
2- You are one of the best at repasting laptops

Thank you for testing

Edit: Your CPU max temp also improved 9C over GELID wich is a good normal difference between LM and NON-LM paste. Could it be why the GPU drop that much?

 

Thank you. The heat sink surprisingly sits very level and makes good contact (above average) from the laptops that I've worked with. There were only minor adjustments needed ...but more on that in my upcoming detailed review.

GPU temps typically don't improve a lot with LM in certain laptops and that's usually due to poor tolerances between the gpu and cold plate. Traditional paste often works better and this has been known for many years already.

You can look at max temp, but that shouldn't be your primary data point as it excludes a lot of details that really matter in the overall picture. That's why it's important to look at all of the data including the averages etc.. If you look at the graph, the stock system was a flat line in redline zone vs the upper 70C's with it LM'ed. This data is important and a very big contrast in showing the before vs after.

I like to use Overwatch as it is a pretty balanced game in taxing the CPU and GPU, especially on Epic settings. It's a widely played game as well, very popular so more people can relate. The bigger picture here is consistency and because I have a library full of data with OW, along with knowing the game characteristics, behavior etc...inside and out, it is what I choose to use for "consistency."

AMD CPU is apples to oranges compared to Intel. They are two very different CPU's with their own strengths and behavior. More details in my review...

Stock factory paste from any manufacturer has always been known to be not very ideal. I chose GELID for the traditional paste to represent an average paste. Even with using an average paste it is much better than cheap factory paste, which continues to prove how these laptops are hindered by that one factor.

I have a lot more info to come...stay tuned.

 

«This one can forget about onsite /in-Home service repair»
DELL- Chris M.

(That is a joke guys. I invented this quote....just for fun)