*OFFICIAL* Alienware Area-51M R1 Owner's Lounge

Discussion in '2015+ Alienware 13 / 15 / 17' started by ssj92, Jan 8, 2019.

  1. Terreos

    Terreos Royal Guard

    Reputations:
    605
    Messages:
    1,133
    Likes Received:
    969
    Trophy Points:
    131
    Holy monkeys. Is that what they’re going for? Might have to keep an eye out.
     
  2. nightingale

    nightingale Notebook Evangelist

    Reputations:
    173
    Messages:
    386
    Likes Received:
    262
    Trophy Points:
    76
    Hi all here, my dad has gone overseas with the area51m i gave him, but i was wanting to put down some benchmark comparisons compared to area51m's with my g703gxr that i got

    Here is the current results ive been getting with the asus, is anyone able to share their benchmarks as well so we can see how much the area51m's are gimped (if at all)

    https://www.3dmark.com/3dm/43554993?
    (i only have 3dmark demo so i dont have access to anything other than timespy)

    Maybe (and im having a big hope for this), dell knows something we dont and even with the 2080 tdp reduction it doesnt impact performance too much in gaming
     
    Fire Tiger likes this.
  3. devilhunter

    devilhunter Notebook Consultant

    Reputations:
    63
    Messages:
    277
    Likes Received:
    178
    Trophy Points:
    56
    In gaming or benching, the average Frequency is what matters, if you boost cuda cores , it would hit thermal limit whatever it is and then it goes back to the average frequency.

    For example, GPU boost to 2100 Mhz, goes for 1 min, temp reaches 85, throttles back to 1850 average temperature 75 and stays this way. Occasionally, it can go high but again gaming session last for 30 mins to hours. One min isnt a big deal.

    But, the question is everyone asking why Dell is limiting the temperatures to 78, why not 85? I believe shifting the average by 50 or 100 mhz at best isnt worth the thermal output and result in less machines sent to dell for repairs.

    But if you undervolt you can get a theoretical 10% bonus IF temperature limits is 85C.

    Would love if the upcoming bios fix things up.
     
    belarusrulez, Lopt and nightingale like this.
  4. Helifax

    Helifax Notebook Guru

    Reputations:
    64
    Messages:
    52
    Likes Received:
    173
    Trophy Points:
    41
    This is what I use for gaming (2-3-4 hours sessions):
    https://www.3dmark.com/spy/10422592
    (Balanced or Performance FAN profile gives little difference, except noise. I haven't yet seen any real difference in temperatures in Performance vs Full Fan Profile, from all my tests :(
    CPU: 69-71 Celsius, GPU with thermal throttle 75 Celsius, without thermal throttle 84-85 Celsius)
    It's a +130Mhz overclock with a custom voltage/frequency curve. If I have the max temp to 75 Celsius it will reach it and then throttle at around 1900-1925Mhz on the RTX 2080 core.
    This is on the latest Bios (1.8.1) with the 200W VBIOS (1.0.0.3). The 180W (1.0.05) VBIOS will give the same results, but it will not allow the card to draw more than 180W at any given time. It will make the card "heat-up slower", but if you plan on going above the 75 Celsius limit, you will not be able to as the power the card can draw will be capped.
    If I set the temperature to 87 Celsius (with the same +130Mhz and custom voltage/frequency curve) I am capping at 82-85 Celsius with a Power Draw of 200W (hence the bios).

    The CPU in the above benchmark is capped at 5.0Ghz on 6 Cores and 4.9Ghz on last cores, with a -100mV offset and a cache ratio of x47.
    (This has been tested over the last 2 weeks, including Throttlestop's benchmark and it is stable).

    The absolute max score I get constantly, on full fan profile with a GPU temperature set at 87 Celsius and 5.0Ghz on all cpu cores with cache x47 is this:
    https://www.3dmark.com/spy/10420758
    + 135Mhz on core + 500Mhz on memory on a standard voltage/frequency curve.
    (However, I cannot recommend this for long gaming sessions, as the CPU will run in games pretty hot around 75-80 Celsius compared to 69-71 Celsius @ 4.9Ghz and the GPU will thermal throttle even at 87 Celsius. Over long periods of time I don't like these temperatures, but they are considered safe).
    In my case, the +135Mhz on the core works fine in some games but crashes in others after some time, I look at Control. (Hence, why I recommend +130Hz which is solid after lots of testing periods.)

    Conclusion: I don't see the Area 51m gimped at all. Sure, there are the new BIOSes that are more protective of the hardware, but if you know what you are doing, you should be fine. This laptop was never meant to be a "plug&play" device like a regular laptop (since is using desktop components in a Laptop case. If anyone thought it would magically work with overclocks and throwing everything at it, than...) I know, I know... we all wanted it "hassle-free", but from my past experience with Alienware from 2011, to-date, it was never so :D


    PS:
    It was actually quite fun finding what works for the i9 9900k in a laptop, as I have the same CPU in my desktop, but is watercooled and clocked to 5.3Ghz.
    The RTX 2080 is roughly 30% slower than the RTX 2080 Ti (which I have on my desktop), so most benchmarks are actually accurate on this regard.

    PPS:
    Since I forgot, the actual difference in FPS if the GPU throttles at 75 Celsius compared to 85 Celsius in my tested is around 2-3-4 FPS, depending on the game. In Metro Exodus (Ultra Graphics settings & Ultra RTX for example (which likes to play havoc with your GPU) the difference is 3 FPS at best.
    If you want to sacrifice 3 FPS and get about 10 Celsius cooler with roughly the same performance, is a question you need to ask yourself;) (Raw numbers in the main menu - 2nd chapter 100-105 FPS no thermal throttle, 98-101 FPS with thermal throttle at 75/85 Celsius).

    Hope this helps. If you have any questions about this, please let me know.

    Cheers,
    Helifax
     
    Last edited: Feb 1, 2020
    sasman1964, Jontoad, c69k and 4 others like this.
  5. Papusan

    Papusan JOKEBOOKs Sucks! Dont waste your $$$ on FILTHY

    Reputations:
    28,132
    Messages:
    25,382
    Likes Received:
    45,481
    Trophy Points:
    931
    People forget the R0 stepping on the KS chips. Less performance vs clock speed. Not nice needing 1 or even 2 bin higher clock speed to get same performance vs clock speed.
    Don't compare the results from the MSIbook Gt76. The 3DM Firestrike physics score is heavly crippld. Even 2666MHz ram won't save it. 26500 in 3DM Fire Strike physics is more like from 4.8/4,9GHz.

    Stock clocks (4.7GHz@2666MHz) from my 5 years old Clevo P870... 5.0GHz should be more in the 27,2/27.5K range. Then do the math for 5.2GHz and you see all the showed results fail...
    [​IMG]
     
    Last edited: Feb 1, 2020
  6. nightingale

    nightingale Notebook Evangelist

    Reputations:
    173
    Messages:
    386
    Likes Received:
    262
    Trophy Points:
    76

    Looks like the results youve gotten are quite good, relatively comparable to other similarly equipped laptops so dell doesnt seem to have overtly gimped performance all too much
     
    Last edited: Feb 2, 2020
  7. Lopt

    Lopt Notebook Consultant

    Reputations:
    163
    Messages:
    245
    Likes Received:
    403
    Trophy Points:
    76

    Interesting I honestly never even tried setting the AWCC Thermal limit above 78c because I figured that it was hard coded in BIOS 1.8.1 to 75c what difference would it make ?

    I am now back on BIOS 1.8.1 and VBIOS 200w

    Well AWCC most be doing something I don't understand because all of the sudden after I set it to 87c Thermal I can now complete Timespy with +130 core and +200 Memory ( in the past almost ANY +memory would cause Timespy to crash out )

    Not absolutely sure that's the reason but I can't think of anything else I changed lately that would account for the difference ( other then going back and forth a few times between BIOS 1.5.0 and 1.8.1 / VBIOS 180 and VBIOS 200 )

    Best Timespy numbers I have gotten so far https://www.3dmark.com/spy/10443893

    My Temps after Timespy in HWinfo are max CPU 72c and max GPU 75c GPU hit 2040mhz

    Maybe I will see if I can push it up to +500 Memory in Afterburner :)
     
  8. Rengsey R. H. Jr.

    Rengsey R. H. Jr. I Never Slept

    Reputations:
    1,056
    Messages:
    2,733
    Likes Received:
    930
    Trophy Points:
    131
    yes , it's good prices at the dell outlet. It's warranty is good as if buying new.

    Usually the price in the outlet is up to 45% off plus the all time additional 12% off.
     
  9. nightingale

    nightingale Notebook Evangelist

    Reputations:
    173
    Messages:
    386
    Likes Received:
    262
    Trophy Points:
    76
    how significant is a hundred or so points in the timespy benchmark?

    it seems like the results of the area51m is relatively comparable to other 9900k/2080 or 9980hk/2080 laptops
     
  10. Dbentt

    Dbentt Notebook Guru

    Reputations:
    24
    Messages:
    64
    Likes Received:
    119
    Trophy Points:
    41
    Diff between a couple/few hundred points in synthetic benchmarks for real world gaming performance is marginal. Sometimes not even measurable in FPS in real games.

    Also, working on a bit of a different sample set right now. Will post it when it's finished.
    edit: theyre ready.
    ---------------------------------------------------------------------------------------------------------------------

    I went ahead and did a test to determine the differences in thermals between various fan presets as well as the workaround HWinfo 'presets'. These results are for my machine alone, consider that your machine can deviate from these results depending on your thermal compound, thermal compound application, and environmental factors. Your undervolts are likely also going to be stronger than what I used to run these tests.

    Specs:
    i7 9700k @ 5GHz on all 8 cores; -85mV offset
    RTX 2080 @ an average 1920MHz core 7500MHz memory; 950mV maximum voltage, often utilized lower than that (875 most common)

    The tests:
    Keep in mind these results are almost all within margin of error. The performance difference, especially real world performance, between these results will be minuscule and unnoticeable. The most noticeable detriment to an unstable undervolt/underclock/overclock manifests as worse 0.1% and 1% dips, stuttering, hitching, and microstuttering. This can and normally will be far more of a problem than losing 2-4 average FPS (which is on the high side, for these results it'd probably be less). I did not encounter these issues in my tests but ymmv depending on your settings.

    What does matter in this case, and the main reason that I did these tests, will be thermals and ambient sound produced by the fans. Thermals and noise levels varied extraordinarily between test setups. Therefor, the purpose of these tests is to find a balance between temperature and sound that matches individual comfort and subjective interpretation. There is no objective answer to which of these results is superior to the other. Also keep in mind that these results are completely synthetic. Thermal performance in games will not ever be this relaxed. A majority of modern titles will push your temperatures higher than these benchmarks can, particularly on the GPU, oftentimes they won't push your CPU nearly as high as a synthetic load however.

    But I digress, without further ado... the tests. Starting with vanilla AWCC profiles.

    AWCC Profiles (77C Temp Limit)
    These tests were all performed, as expected, with the built in fan profiles in AWCC. I set the temp limit to 77C via AWCC to prevent hard throttles to 500MHz on the GPU and turned off AWCC's CPU overclocking module, opting instead to use Throttlestop to control frequency and voltage. All other GPU related function was handled via MSI Afterburner.


    Full Speed

    Time Spy
    https://www.3dmark.com/spy/10445773

    Max CPU Temp: 85C
    Max GPU Temp: 69C

    Fire Strike Extreme
    https://www.3dmark.com/fs/21702998
    Max CPU Temp: 79C
    Max GPU Temp: 66C

    Port Royal
    https://www.3dmark.com/pr/213466
    Max CPU Temp: 67C
    Max GPU Temp: 70C


    Performance

    Time Spy
    https://www.3dmark.com/spy/10445899

    Max CPU Temp: 85C
    Max GPU Temp: 73C

    Fire Strike Extreme
    https://www.3dmark.com/fs/21703096
    Max CPU Temp: 79C
    Max GPU Temp: 67C

    Port Royal
    https://www.3dmark.com/pr/213472
    Max CPU Temp: 72C
    Max GPU Temp: 72C


    Balanced

    Time Spy
    https://www.3dmark.com/spy/10446021

    Max CPU Temp: 87C
    Max GPU Temp: 76C Thermal limit.

    Fire Strike Extreme
    https://www.3dmark.com/fs/21703183
    Max CPU Temp: 87C
    Max GPU Temp: 75C

    Port Royal
    https://www.3dmark.com/pr/213476
    Max CPU Temp: 85C
    Max GPU Temp: 76C Thermal limit.


    HWinfo64 Profiles (80C Temp Limit)
    These tests used the recently discovered workaround to bypass Dell's hard temp limit via some kind of crossed wire tech wizardry related to fan profiles. Each of them was set according to the values that you're allowed to set; HWinfo64 is severely limited in scope when it comes to what you can set your fan RPMs to, presumably also thanks to Dell. As a result, these fan profiles sound extraordinarily jarring compared to the vanilla AWCC ones. Outside of the balanced profile, the fans are either off or running full tilt. This has a serious set of drawbacks that I'll touch on later. For now, the numbers.



    Full Speed

    Time Spy
    https://www.3dmark.com/spy/10446789

    Max CPU Temp: 84C
    Max GPU Temp: 71C

    Fire Strike Extreme
    https://www.3dmark.com/fs/21703677
    Max CPU Temp: 82C
    Max GPU Temp: 66C

    Port Royal
    https://www.3dmark.com/pr/213505
    Max CPU Temp: 78C
    Max GPU Temp: 71C


    Performance

    Time Spy
    https://www.3dmark.com/spy/10446673

    Max CPU Temp: 88C
    Max GPU Temp: 73C

    Fire Strike Extreme
    https://www.3dmark.com/fs/21703540
    Max CPU Temp: 81C
    Max GPU Temp: 69C

    Port Royal
    https://www.3dmark.com/pr/213499
    Max CPU Temp: 73C
    Max GPU Temp: 72C


    Balanced

    Time Spy
    https://www.3dmark.com/spy/10446367

    Max CPU Temp: 97C
    Max GPU Temp: 76C

    Fire Strike Extreme
    https://www.3dmark.com/fs/21703391
    Max CPU Temp: 90C
    Max GPU Temp: 73C

    Port Royal
    https://www.3dmark.com/pr/213487
    Max CPU Temp: 80C
    Max GPU Temp: 76C


    Summary
    So what might we draw from these results?

    First and foremost, the workaround introduces a significant increase in heat and/or noise. This is to be expected as the fans are not following a smooth curve, but rather a binary on or off. It is excessively jarring to go from 0 or 1300 RPM up to 2000, 3000, 3200, etc.. This leads in to my next point...

    Idle temperatures with the workaround are going to go up by at least 33%. In my experience, my idles were going from 30-40C on the CPU and 30-35C on the GPU up to a pretty steady 50C-60C on both. The problem here is, you will either have your fans at 0RPM, causing the entire machine to be passively cooled by the heat pipe array until it hits your target thermal point (50C, 60C, 70C, etc.) and the fans kick on at full blast. This will happen over, and over, and over, and it's terrible. Therefor the only real solution is to have the fans running basically full time, which can be loud and obnoxious when you don't have the sound from a game to drown out the constant fwoosh of mid to high RPM fans. This is less severe on the Balanced mode, but I'll get to Balanced shortly. Alternatively you can cook your laptop and have the fans set to 0RPM up to 60 or 70C which is terrible for idle temperatures in any machine, but that's your decision.

    As far as thermals themselves are concerned, once the fans are on and everything is running the difference between stock and workaround becomes less pronounced. You'll notice the noise less during gameplay, especially with a headset, but there are still some very notable differences. The Balanced fan profile in particular performed very poorly via HWinfo in some cases. This is unfortunate because Balanced is... well, the most balanced. it is the only (in my opinion) usable fan profile that offers 3 steps in HWinfo (0, 1300, and 2800). You can thus have the fans run at 1300RPM while idle, and 2800 under load... however I found 2800RPM to be inadequate in some tests, and it would be even worse in a lot of real world scenarios. But ymmv on this. Full Speed and Performance ended up being very close between AWCC and HWinfo, however there was the serious drawback of the fans turning themselves on and off constantly, which is very distracting in a quiet room, or running full blast at all times at a monstrous 3000/3200 RPM for Performance and 3600/3800RPM for Full Speed. Each side of this particular coin can be ugly and bad depending on your preferences. If you don't care about these problems though, we get on to the final part of my summary...

    The entire reason that you would do this at all is to prevent thermal throttling from Dell's dumb BIOS. Unfortunately, these tests were incapable of representing this as none of them stressed the GPU hard enough for long enough to hit my 80C thermal limit, nor did we go above the original thermal limit to prevent throttling with a standard AWCC profile when using the HWinfo workaround. That said, when playing actual games, my temperatures can absolutely exceed 76C, and I do not experience the dreaded 500MHz emergency mode throttle that you normally would (as expected; it was never a question whether or not the workaround actually did what it was supposed to do as I trusted the original source already). So ultimately, you must choose between somewhat worse thermals and much worse sound management vs. a lower thermal limit for your GPU. Does this provide any actual benefit? Sort of. Being able to go higher than 78C on your GPU will allow you to have higher clocks for longer, but that doesn't necessarily mean you will get the highest clocks out of your GPU at all times.

    GPU Boost 3.0 will still begin throttling your card well before you hit 87C, and there is no way to turn that off (nor would you ever want to). Are there gains? Sure. But they probably aren't worth the additional heat. At least not for me. The primary draw to the workaround for me, then, is the confidence in knowing that I won't randomly get blasted into chromebook territory for having the audacity to use my gaming desktop replacement notebook... for gaming. I don't know if this is for everyone, but for me, I sometimes experience thermal throttling before hitting that 78C mark. It is random and unpredictable, but it happens. Removing this issue is almost worth it for me, but ultimately the irregularity of the fans' noise level, and the higher idle temps, will force me back to running AWCC by itself at a lower thermal limit and crossing my fingers that I don't encounter any problems by doing so, or more realistically, back to 1.5.1 BIOS where I lose my coveted GSync functionality in exchange for the peace of mind that my machine won't become a glorified paperweight until it stops having its mood swing. $3,500 computer ladies and gentlemen.

    As for that point, which fan profile is best? For me, I'll be sticking with Performance. Noise levels are completely managable, and there's a marked difference in thermal performance compared to Balanced. Full Speed does have the best thermals out of all three, but I'm not running my fans at 100% at all times to achieve that benefit. If I'm running something particularly demanding and notice that I'm hitting my thermal limit all the time, then I can always turn on Full Speed and then turn it off after.



    Anyway, thanks for hopefully not reading my way too long and largely pointless post about fan profiles and temperature.
     
    Last edited: Feb 2, 2020
    Lopt, MogRules, sasman1964 and 2 others like this.
Loading...

Share This Page