15" i7 MBP 2017 vs 2018 temperature?

Discussion in 'MacBook Pro' started by vmflapem, Dec 4, 2018.

  1. vmflapem macrumors 6502

    Joined:
    Dec 27, 2013
    #1
    I'm just wondering, has anyone had a chance to compare the temperature of 2017 and 2018 15" i7 MBP? I see many threads on the i9 model but not the i7. I was okay with my old 2017 model temperature and now I'm considering buying a 2018 model. If it's significantly hotter, then I might skip the 2018 model...

    Also, would Vega 20 run cooler with i7 than 560X with i7? This was probably mentioned before but I could not understand any of the technical terms :(
     
  2. Thysanoptera macrumors 6502

    Joined:
    Jun 12, 2018
    Location:
    Pittsburgh, PA
    #2
    2018 i7 will be hotter (has two more cores and no sensible limits on power draw), Vega will be hotter (uses 40% more power than 560X). Same chassis and cooling capability.
     
  3. maflynn Moderator

    maflynn

    Staff Member

    Joined:
    May 3, 2009
    Location:
    Boston
    #3
    Of course its going to be hotter, the 2018 model has a better, but not cooler GPU (unless you're taking about the Vega), and Coffee Lake based processors have two extra cores, so its going to run hotter. Regardless of what computer you're talking about, dell, asus, razer, or Apple. The coffee lake based processors run hot and people have complained about it.

    The 2019 model may run just as hot. Its apple's fault for not redesigning the enclosure to better manage the heat. If they go with the same case, as with the 2018/2017/2016 models, you're going to be dealing with a hotter computer.

    Why? The 2017 MBP is very fast on its own. What are you doing that you need to upgrade to the next model year?

    Vega does run cool, but apple prevents you from buying an I7 with a Vega, you need to buy the top line I9 and then pay extra for the Vega. The I9 runs a lot hotter and will throttle more then the base I7. In fact there's a number of posts here that show that the 2.2GHz I7 having better benchmark numbers then the I9
     
  4. leman macrumors G3

    Joined:
    Oct 14, 2008
    #4
    I don't see any difference in temperature between the late 2018 MBNP 15" and the two 2018 (i9) 15" I could test. In fact, the newest 2018 model (arrived the other day) I have here appears to be significantly cooler. Right now the CPU is at 35C which is very low for a MBP (and I've spend last two hours working on this machine).
     
  5. PineAppleMac macrumors member

    PineAppleMac

    Joined:
    Dec 1, 2016
    #5
    Nope, you can buy i7 with Vega. Upgrade to i9 is not necessary.
     
  6. maflynn Moderator

    maflynn

    Staff Member

    Joined:
    May 3, 2009
    Location:
    Boston
    #6
    You're right, you cannot upgrade the 2.2GHz with the Vega, you need to select the high end 2.6
     
  7. vmflapem thread starter macrumors 6502

    Joined:
    Dec 27, 2013
    #7
    I sold my 15" MBP and downgraded to a 13" MBP, but the screen is too small I can't get used to it so I'm thinking of upgrading again. So is it correct to say that although Vega uses more power, it still runs cooler than 560X?
     
  8. Thysanoptera macrumors 6502

    Joined:
    Jun 12, 2018
    Location:
    Pittsburgh, PA
    #8
    No, that is impossible, all power is converted to heat, and Vega uses a lot more. Vega is 25-30C hotter than 555x, 15-20C hotter than 560x at the same fan rpm.
     
  9. maflynn Moderator

    maflynn

    Staff Member

    Joined:
    May 3, 2009
    Location:
    Boston
    #9
    That's not what I've read. The vega while outperforming the 555x/560x consumes less power and thus runs cooler.
     
  10. Thysanoptera macrumors 6502

    Joined:
    Jun 12, 2018
    Location:
    Pittsburgh, PA
    #10
    I guess even for a moderator it was easy to miss the single post where someone finally posted the actual values rather than his feelings ;) Vega 20 - 50W, Vega 16 - 45W, 560X - 35W, 555x - 30W. I'd appreciate if you could point me to where you read that Vega consumes less power, for my entertainment. I've seen posts/reviews claiming Vega is cooler, it blows my mind how anybody couldn't notice 20, 30C difference the other way.
     
  11. maflynn Moderator

    maflynn

    Staff Member

    Joined:
    May 3, 2009
    Location:
    Boston
    #11
    Link to the source material please, and yes everything I've read so far shows a cooler running laptop with vega. Why does my status of moderator mean I see every post?
     
  12. leman macrumors G3

    Joined:
    Oct 14, 2008
    #12
    I've been comparing the my newly arrived i9 with Vega Pro 20 to the maxed out 2016 I have here and I don't really see any difference in the temperatures. Under load, the temps are the same, both GPUs max out at 80C and the CPUs max out at 100C.Din't check out the fan speed yet. In low-power operation (normal work with text, browsing etc.) the new machine is considerably cooler than any other Mac I've used before — its consistently 33C with 22C ambient temperature. All other Mac laptops I had idled at at least 40C or higher.

    In the end, I think that the question doesn't make much sense. Temperature is dependant on what one is currently doing. So when one asks me "which laptop runs hotter", I can't answer that question generally.
     
  13. Thysanoptera macrumors 6502

    Joined:
    Jun 12, 2018
    Location:
    Pittsburgh, PA
    #13
    Yes you can, just set the fans to the same value, do the same task and you will see a huge difference. And saying that Vega is cooler implies that it has lower power requirements.

    @maflynn - here are the values:

    Does the Vega run hotter than the 560X?
     
  14. maflynn Moderator

    maflynn

    Staff Member

    Joined:
    May 3, 2009
    Location:
    Boston
    #14
    Thank you, I was hoping for specifications rather then a single user's experience. I'm not discounting what he's experiencing but I think actual specs to refute what's written would be helpful as well.
     
  15. leman macrumors G3

    Joined:
    Oct 14, 2008
    #15
    I can't say that I agree. You are reducing the question to power consumption under high load. But what about power consumption under light load (e.g. when driving an external monitor, but not doing anything intensive)? Or about the overall laptop temperature under load?

    Anyway, I'll run side-by side graphics benchmarks later and look at the fan speed. If Vega produces more heat, the new laptop should have its fans revving higher.
     
  16. upandown macrumors 6502

    Joined:
    Apr 10, 2017
    #16
    I highly doubt the computer is always 20-30c higher, even at load. Yes it has a higher a TDP but it wouldn’t be able to operate normally if the temps were 60c-70c at idle. The reason no one has complained about these temps you’re suggesting is because they don’t exist.

    We are dealing with the same TDP numbers from the 2000s and 1900s. Ie 35w, 45w etc. yet the performance is exponentially better. Using the same TDPs. The Vega is different hardware and better optimized software = better results within the same constraints. So it’s completely possible the Vega runs faster and cooler. It’s the progression of technology.
    This forum over does it when it comes to heat.

    Edit: my 2018 13 i7 quad core runs noticeably cooler than my 2012 13 i5 dual core. Progress.
     
  17. Thysanoptera macrumors 6502

    Joined:
    Jun 12, 2018
    Location:
    Pittsburgh, PA
    #17
    Yes, I am absolutely reducing it to question of power under load. Why wouldn't I? They should be identical at idle. The author of the screenshots I posted says his Vega is 13 degrees cooler at idle. His idle temps with Vega are 43C, which makes his idle temps with 560X - 56C. However you slice it - it is not normal, and I would get rid of such laptop. And your fans - they are maxed out at load, actually - Apple raised the rpm limit on Vega.

    Apple has tuned the fan curves to be quiet, and allows both CPU and GPU to be hot which I hate, so I can kind of see that without looking at fan speeds one could be under impression that the heat dissipation is similar. But still, it looks like they raised the GPU temp setpoint on Vega to 80C from 70C on RX, so you will have higher temp on GPU even on default curves. Match the fan speeds and you will see 20 - 30C temperature drop on RX.

    Absolutely Vega is more efficient than Polaris, but it is not groundbreaking 60% but rather something in the range of 10-20%. But because it has higher TDP it will be hotter, progression of technology didn't break so far the first law of thermodynamics. What I have a beef with, is we were led to believe this is going to be 35W GPU, Nvidia killer. Instead we got something that trades blows with Ti MaxQ in XPS15 while having higher TDP and being two years late.
     
  18. leman macrumors G3

    Joined:
    Oct 14, 2008
    #18
    So, I've just tested my 2018 (i9, Vega Pro 20) vs. the old 2016 (i7-6920HQ, Pro 460), side by side. I don't have a 2017 machine to hand, but the 2016 should be a reasonable approximation. It's essentially the same GPU anyway. What I did was running Unigine Valley in a small window for a while, while watching the fan speed and temps.

    Result: I'd say no difference, really. The 2018 machine ramps up the fans earlier but the 2016 is louder in prolonged operation. Assuming that the fans on both systems operate with the same efficiency, I'd say that the last fact means that (if anything) the 2018 i9 Vega Pro 20 runs more efficiently under load than the 2016 i7-6920HQ Pro 460

    Detailed result:

    - Both GPUs almost immediately get up to 80C and stay there
    - The 2018 model ramps it fans up quicker, its in the 4000 range after a minute, while the 2016 one still sits at around 3000, but
    -
    after 10 minuets of running, the 2018 model I still in the high 4000 range while the 2016's fans are revving like crazy at 5500. Thats where the equilibrium has been reached since I couldn't observe any changes in fan speed from there on
    - Vega Pro 20 was 30% faster when benched. The difference is this small since it was running in a small window, so API and data sync overhead was more noticeable. When run at fullscreen (1920x1080, Ultra, AA off), the Vega Pro is 70% faster in this particular benchmark.
    --- Post Merged, Dec 6, 2018 ---
    Not necessarily. There are two "idle" modes after all — one when the dGPU is completely off and one when its on, but not doing anything demanding. The second one is active when you are connected to an external monitor. And that is exactly the situation where running cooler is beneficial, since you don't want your laptop to be loud when all you are doing is driving an extra monitor. I haven't had the chance to test the new laptop with an external monitor, but according to many other reports, it is indeed much cooler than the previous GPUs.

    That is absolutely normal when running an external monitor. That is the temperature I saw with both the 2016 and with my previous 2018 (560X).

    The max RPM on my the Vega machine is 6000, on the 2016 machine its 5900. I don't know if that is such a large difference...
    --- Post Merged, Dec 6, 2018 ---
    At the same time the very same chassis performs better under stress testing and overall power consumption hasn't changed. And, at least in my experience, I haven't noticed that the 2018 machine runs any hotter than any other MBP I've tested.
     
  19. Thysanoptera macrumors 6502

    Joined:
    Jun 12, 2018
    Location:
    Pittsburgh, PA
    #19
    Thank you for taking time to do it, I really appreciate it but I think there was something wrong in the test, it doesn't look like the Vega was fully loaded in this small window test, that 30% kind of gives it away, it was bottlenecked by something. There are massive improvements in Unigine on Vega. 4000 rpm thats around what it takes to dissipate 35W. If you want to compare, in the other thread I have a screenshot of my Heaven run on 555x, fans set to max, extreme preset, run 4 times and screenshoted iStatMenu sensor page at 250 sec mark of the 4th run.

    And two year old laptops tend to accumulate a ton of crap in the heatsink reducing its performance, plus that 6th gen CPU may contribute.

    Maybe, I don't know. Maybe they changed how the external monitors are driven, from the iGPU, like on Windows machines. Again - looking at the power consumption would answer that question. The great invention of eGPU allows me not to worry about it, I'm only concerned about the dGPU while it is being used in mobile settings.

    They both (left and right) go to 6000 now, before it was 5900 and 5500. Small difference, true, but still.
     
  20. vmflapem thread starter macrumors 6502

    Joined:
    Dec 27, 2013
    #20
    I heard that the internals of MBP have been re-designed for Vega 20. Is it possible that maybe this is why some people are claiming that Vega runs cooler than 560X? I'm so clueless I wish I had some knowledge in this field lol.
     
  21. Appledoesnotlisten macrumors member

    Appledoesnotlisten

    Joined:
    Dec 2, 2017
    #21
    I own same model - the coolest laptop I have owned.
    And I am returning the 15" 560X 2018 because it's too freaking hot even at idle.
     
  22. leman macrumors G3

    Joined:
    Oct 14, 2008
    #22
    So, I did some more testing and I think you are right. I can confirm that Vega Pro 20 is a 50 watts GPU, up from the 35 watts of the Polaris series. Thats what the iStats reports and that is also what I see as total system power consumption.

    The power management now seems to distribute the available power dynamically between the CPU and the GPU depending on actual demand. Would also explain very high improvements in the GPU-only benchmarks — the CPU is not doing that much work and the system can dedicate all the power to the GPU. Real-worlds improvements will depend on the application, if high CPU uptime is required, the GPU will be tuned down correspondingly.
     
  23. Thysanoptera macrumors 6502

    Joined:
    Jun 12, 2018
    Location:
    Pittsburgh, PA
    #23
    Good to know that the other user result wasn't a fluke. I'm usually afraid of situations where GPU takes over almost all cooling capability of a given chassis, but looks like Apple did a sensible thing and put a 30W hard limit on GPU throttle leaving the rest for CPU preventing huge spikes on down clocks often present on Windows boxes with inadequate cooling. I'm disappointed about the 50W TDP though and I can't figure out where the HBM2 power savings went to.
     
  24. vmflapem thread starter macrumors 6502

    Joined:
    Dec 27, 2013
    #24
    So if 560X is hot at idle as some people are saying, that means Vega 20 is going to be even hotter at idle? damn...
     
  25. Thysanoptera macrumors 6502

    Joined:
    Jun 12, 2018
    Location:
    Pittsburgh, PA
    #25
    No, at idle they should be the same. There are some claims that Vega may actually be cooler (as in consume less energy) at some tasks, like driving an external monitor that Apple hardwired to it and forces its use. Vega will be hotter/louder at full load.
     

Share This Page