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

Discussion in '2015+ Alienware 13 / 15 / 17' started by iunlock, May 18, 2021.

  1. Terreos

    Terreos Royal Guard

    Reputations:
    1,146
    Messages:
    1,764
    Likes Received:
    2,139
    Trophy Points:
    181
    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. :)
     
    Vistar Shook, Rei Fukai and Normimb like this.
  2. Papusan

    Papusan Jokebook's Sucks! Dont waste your $$$ on Filthy

    Reputations:
    40,612
    Messages:
    29,272
    Likes Received:
    57,551
    Trophy Points:
    931
    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 mobileo_O

    http://forum.notebookreview.com/thr...r5-owners-lounge.835745/page-14#post-11096208
     
  3. Normimb

    Normimb Notebook Evangelist

    Reputations:
    1,733
    Messages:
    547
    Likes Received:
    1,008
    Trophy Points:
    156
    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.
     
    etern4l and Rei Fukai like this.
  4. iunlock

    iunlock 7980XE @ 5.4GHz

    Reputations:
    2,027
    Messages:
    4,519
    Likes Received:
    6,412
    Trophy Points:
    581
    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. :)
     
    Vistar Shook and Terreos like this.
  5. iunlock

    iunlock 7980XE @ 5.4GHz

    Reputations:
    2,027
    Messages:
    4,519
    Likes Received:
    6,412
    Trophy Points:
    581
    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:
    [​IMG]

    w/ Traditional Paste:
    [​IMG]

    Bone Stock out of Box:
    [​IMG]
     

    Attached Files:

    c69k, Terreos, Normimb and 2 others like this.
  6. Normimb

    Normimb Notebook Evangelist

    Reputations:
    1,733
    Messages:
    547
    Likes Received:
    1,008
    Trophy Points:
    156
    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?
     
    Last edited: May 28, 2021
    Vistar Shook, c69k and Terreos like this.
  7. iunlock

    iunlock 7980XE @ 5.4GHz

    Reputations:
    2,027
    Messages:
    4,519
    Likes Received:
    6,412
    Trophy Points:
    581
    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. :)
     
  8. Normimb

    Normimb Notebook Evangelist

    Reputations:
    1,733
    Messages:
    547
    Likes Received:
    1,008
    Trophy Points:
    156
    «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)
     
    Rei Fukai, etern4l and Papusan like this.
Loading...

Share This Page