This is how the 2017 MacBook Core m3, i5, i7 behave during sustained CPU load: 10+ runs of Cinebench

Discussion in 'MacBook' started by EugW, Sep 28, 2017.

  1. EugW, Sep 28, 2017
    Last edited: Oct 24, 2017

    EugW macrumors 601

    EugW

    Joined:
    Jun 18, 2017
    #1
    I posted this originally in a different thread, but I think it deserves its own thread.

    [EDIT]

    See this post for results of the m3, i5, and i7, with the m3 tested on two different table tops:

    https://forums.macrumors.com/thread...stained-cpu-load.2073415/page-2#post-25271927

    MacBook2017-CinebenchR15-m3-wood.png

    Summary: All machines pretty close to each other in performance, but the i7 is indeed the fastest and the m3 on a wood table is the slowest. However, the m3 on a granite table did better, presumably because the granite wicked away heat better.


    [/EDIT]

    I ran Cinebench R15.038 twenty-five times in a row on my 2017 MacBook Core m3-7Y32. This took about 65 minutes. So essentially it's over an hour of sustained, load aside from the second or so pause between runs. The numbers below are the Cinebench 64-bit multi-core CPU scores, and the speed of the first run is roughly what one would get from an early 2015 Core i5-5250U MacBook Air.

    MacBook Core m3.png

    Ambient room temperature was 23°C. The table top was a granite breakfast bar. The MacBook was running off battery power, with screen dimming turned off. WiFi and Bluetooth were turned off, and Sophos anti-virus was deactivated. Out of all these things, it seems only the anti-virus has a real impact on performance. OS is macOS 10.13 High Sierra. 16 GB RAM and 256 GB SSD.

    Note that part way through the 11th run, I moved the computer over a bit because I was rearranging stuff on the counter. Low and behold, the score went up, and continued to go up the next run too. I noticed the part of the table where the computer had been was a little bit warm (but not hot), so it seems moving the computer over to a cooler spot helped performance. So, right after run 22, I intentionally did the same thing again, and again the score went up.

    Below are the raw scores for future reference in case someone wants to put it into an Excel spreadsheet, but the most valid scores are the first 10 (and then from 12-22) for the reasons stated above:

    264
    261
    259
    258
    258
    258
    257
    255
    255
    253
    254
    257
    257
    257
    256
    255
    255
    254
    252
    252
    251
    250
    254
    258
    258

    You'll note that not even once did the score drop below 250. It may have dropped a bit below 250 had I not moved the computer, but extrapolating that would mean it would taken about half an hour of sustained load to go below 250, which is only 5-6% less than max speed. (Given that 25 runs is about 65 minutes, that means 10 runs is around 25 minutes of sustained load, and 12 runs is about half an hour. I think it would have taken 12 runs to dip below a score of 250.)

    More importantly, compare this to Notebookcheck's 2017 MacBook Core m3 results. I didn't get anything like their weird results:

    Capture.PNG

    Perhaps they got a lesser m3, or else it was hotter in their room, or else their table retained more heat. Their m3 seemed to periodically throttle really hard after sustained high workload, but it's curious because it's very different behaviour from their 2016 models (below).

    Furthermore, the 2017 Core m3 beats all of the 2016 m3, m5, and m7 MacBooks from Notebookcheck's 2016 m3/m5/m7 MacBook face-off:

    csm_cb15_schleife_macbooks_40e6d2367d.jpg

    This 2017 m3 also beats Eldar Gezalov's 2017 Core i7 MacBook under the same test:

    Capture.PNG

    I'm not sure why Eldar's Core i7 is slower than my Core m3. Either it is throttling more or else he has more background processes running. If it's throttling more, then it's a repeat of Notebookcheck's 2016 results, where the m7 ran slower than the m5.
     
  2. EugW, Sep 28, 2017
    Last edited: Sep 28, 2017

    EugW thread starter macrumors 601

    EugW

    Joined:
    Jun 18, 2017
    #2
    Coolio1234 was good enough to post his i7 results in the other thread:

    I put them into Excel and created a comparative graph.

    2017MacBookCinebench.PNG

    Note though in run 11 his i7 received a phone call, and also in run 11, I had moved my computer to a different part of the table.

    Nonetheless, over runs 1 to 10, the two chips seem to slowly converge as the chips warm up. The biggest difference in performance is in runs 2 and 3, at about 5%.
     
  3. Mike Boreham macrumors 65816

    Joined:
    Aug 10, 2006
    Location:
    UK
    #3
    Great data thanks both!

    All we need now is for a 2017 i5 owner to do the same.

    As you say not a lot between the M3 and i7.
     
  4. EugW thread starter macrumors 601

    EugW

    Joined:
    Jun 18, 2017
    #4
    This is what the graph looks like if you show the entire y axis:

    2017MacBookCinebench_allyaxis.PNG
     
  5. taxi_driver, Sep 28, 2017
    Last edited: Sep 28, 2017

    taxi_driver macrumors member

    Joined:
    Sep 26, 2017
    #5
    this got really interesting (thanks @andy9l ) does this essentially make the i7 not worth the cost/premium? guess a MBP is the better buy dollar for dollar..

    adding the i5 in this comparo would be really great to see
     
  6. Mike Boreham macrumors 65816

    Joined:
    Aug 10, 2006
    Location:
    UK
    #6
    This is only one of many benchmarks and there are plenty of folk around who say benchmarking is flawed and not representative of real work and activity. But Benchmarking is interesting and fun so I do it anyway.

    Geekbench still shows a difference between the three, but there are long threads around about how that is flawed.

    When I was working and using Photoshop all day I used to use some of the Photoshop based benchmarks (Retouch Artists, Diglloyd) as I knew they were measuring what I mostly did with the machine (MacPro). I am sure there are similar for video and audio, although anyone doing regular heavy lifting will not be using a MacBook. People who use pro apps on a MacBook occasionally in the field might still be well served by such methods.

    An interesting test would be to ask people to use each of the three for a week without their knowing which was which (and just using them normally, not benchmarking) and seeing if they could rank them correctly.

    I think like many things it is a diminishing returns thing, and probably hard to justify maxing out, and a buying choice is about your personal attititude. Futureproofing comes into as well.

    Personally I choose on Storage, RAM and CPU in that order, but often max out for the peace of mind, and because I know that, if I save a few hundred by not maxing out, once I had actually spent the money, the extra few hundred doesn't seem as much and I wish I had one it. (I didn't max out my MacBook as I got a rare deal to buy a nearly new M5). There is nothing subjective about not having enough storage, or about not having enough RAM if you are hitting memory pressure with VMs.
     
  7. pika2000 macrumors 68040

    Joined:
    Jun 22, 2007
    #7
    I think intel's "i5 and i7" for the retina Macbook is just a rebrand. They are still the same line as the core m3, but intel likes marketing better.
     
  8. Mike Boreham macrumors 65816

    Joined:
    Aug 10, 2006
    Location:
    UK
    #8

    Yes they are all still M processors, but Apple or Intel (confusingly and no doubt misleading some people into believing they have the same i5/i7 processor as their iMac or MBP) refers to the top two as i5 and i7 instead of M5 and M7.

    However I have a vague memory that when I looked at "About this Mac" in a store on a 2017 i5 it actually said M5. Can someone confirm or correct this.
     
  9. EugW thread starter macrumors 601

    EugW

    Joined:
    Jun 18, 2017
    #9
    BTW, a couple of months ago some of us did some yes testing to see throttling speeds. Assuming Intel Power Gadget is reporting things accurately, both the i7 and m3 will throttle to the same 2.5-2.6 GHz speed.

    m3-pegged_zpswuz4xwnn.png

    Assuming this is correct, then this would explain why the graphs above for the Cinebench scores appear to converge over time, presumably as residual heat increases with each run.
     
  10. EugW thread starter macrumors 601

    EugW

    Joined:
    Jun 18, 2017
    #10
    We still need a volunteer from a MacBook i5 owner. :)
     
  11. Brookzy macrumors 601

    Brookzy

    Joined:
    May 30, 2010
    Location:
    UK
  12. Brookzy macrumors 601

    Brookzy

    Joined:
    May 30, 2010
    Location:
    UK
    #12
    Okay, so:

    2017 i5/512/8GB. All apps quit, Wi-Fi and Bluetooth on. On battery power. Mouse wiggled when the screen dimmed. Minimal seconds between each run (i.e. enough time for me to write the number down and hit 'run' again).

    Results to add to the graph:

    264
    264
    262
    260
    256
    252
    251
    250
    249
    249

    So it actually appears to throttle more than either the m3 or and i7, i.e. be the slowest. However these numbers are so close, and the conditions so varied, that it is hard to draw conclusions. Also in my tests the MacBook remained absolutely still (the table got very hot) and the tests were done with no break at all. The fluctuations in the m3 and i7 results makes me very suspicious of the conditions.

    Indeed the Intel Power Gadget shows that it throttled just the same as the others:

    Screen Shot 2017-10-18 at 22.53.49.png
     
  13. EugW thread starter macrumors 601

    EugW

    Joined:
    Jun 18, 2017
    #13
    The significant fluctuations for the m3 and i7 results only happened in the 11th run or later. In the case of the m3 it was because I moved the machine (to a cooler part of the table), and in the case of the i7 it was because it received a phone call during the test. So you can just ignore the results after the 10th run. Your test only includes 10 runs anyway.

    I'll add your results later when I have a chance, but it may take a while to upgrade the graph as my original file isn't with me and I'm currently out of the country. It seems that the results of the i5 aren't significantly better than the m3, even on the first couple of runs. The difference in performance vs the m3 (and i7) is just a couple of percent for the most part. It's not a huge surprise that the i5 isn't significantly faster though, as its top speed for Turbo is 3.2 GHz. It's 3.0 GHz for the m3.

    Perhaps I did so well with the m3 was because my table is granite. It probably wicks away heat better. Maybe at some point I should try it on a wood table or something like that. I suspect if the table had retained more heat, my m3 will would have throttled more over the 10 runs.

    In any case, the point here is it seems overall the best bang for the buck might be the m3, esp. if you want more memory and want to save some money. The different chips are close enough to each other that the differences aren't probably not that noticeable in most usage, but the added memory of 16 GB could make a huge difference.

    BTW, I had turned off screen dimming for my test. No trackpad tapping or mouse wiggling needed.
     
  14. magbarn macrumors 68000

    Joined:
    Oct 25, 2008
    #14
    Wow the m3 is boss. The other’s are basically “marketing speeds”. The extra for a i7 is so not worth it either. Too bad you can’t get a m3/16gb/512gb configuration which would be the best bang for the buck.
    Active cooling would likely allow i7 to stretch it’s legs more,
     
  15. GoldfishRT macrumors 6502

    GoldfishRT

    Joined:
    Jul 24, 2014
    Location:
    Somewhere
    #15
    This is some great info. While there are a lot of variables I think it's safe to say that the MacBook is not built for sustained loads no matter how you slice it. So I imagine after time, things like games, the rendering times for long videos, stuff like that it doesn't seem to make much sense to shell out for the faster processors. Even the m3 becomes limited by the ability of the chassis to move heat away from the processor and without a fan that's severely limited. I guess for certain applications where the burst in speed is necessary then it might be more justifiable? Anyways, I'm not sure if we should applaud Intel for making the M3 such a competent chip or lambast Apple for thinking it acceptable to charge what they do for the i7 upgrade. Maybe both...

    And for the record, my experience coming from the 2016 M3 to the 2017 was pretty night and day. The MacBook went from somewhat performance compromised to totally adequate for all my uses.
     
  16. Mike Boreham macrumors 65816

    Joined:
    Aug 10, 2006
    Location:
    UK
    #16
    I look at all these results somewhat differently. None of the three processors have fallen over thermal cliff edge due to inability to get the heat away. Even after an hour of sustained load the performance is only a few percent down on the cold result.
    The reason why it is not suitable for heavy rendering etc is more to do with its basic speed. People who often do long renders etc should chose a faster machine anyway.
    All three are suitable for people who do occasional long renders and don't mind having an extra cup of coffee. The MacBook is not going to grind to halt because it can't get the heat away.
     
  17. EugW thread starter macrumors 601

    EugW

    Joined:
    Jun 18, 2017
    #17
    Yes, that would be a great option. I went with the m3 / 16 GB / 256 GB option myself, but I don't keep any primary data on this machine for any length of time. It's my travel machine, but my main data gets off-loaded to the cloud and/or to my NAS. (My main machine at home is an iMac with 1 TB SSD.)

    However, had the m3 / 16 GB / 512 GB option been available, I might have gotten that, just because.

    Very true.

    Yes, for my usage, which is admittedly not high performance, I agree. I didn't check out the 2016 much, but I remember being somewhat disappointed by the performance of the 2015 M 5Y31 (which is that year's m3 equivalent). The machine was OK speed-wise, but was noticeably slower than the other Mac laptops, and the keyboard sucked. The 2017 m3 is quite a noticeable jump in performance over the 2015, and has a much improved keyboard as well. While I wouldn't necessarily recommend an m3 MacBook as a primary workhorse, I think for my mainly business type use on the go it does not feel performance compromised. The 2015 felt a bit compromised even back in 2015.

    At this point with the m3 performance, I'm good. The main concern for me is memory and storage. Having a whole bunch of business applications open at the same time and Safari and email doesn't actually necessarily need a ton of performance, but having more memory is beneficial. I don't need a lot of storage though (as mentioned above).

    Reasonable assessment. I also am pleased that it doesn't hard throttle like the unit that Notebookcheck reviewed. I'm glad it seems their model may be an anomaly.

    I'm currently at a conference with a bunch of co-workers, and almost all of them have MacBook Pros. A few of them have the latest TouchBar MacBook Pros and for what we do the performance advantage of the MBP is pretty much irrelevant. The main advantage of the MBPs is that they have larger screens, but those machines, mainly 15" models amongst my colleagues, feel very awkward to me.

    BTW, the other great feature of this MacBook is the very good battery life, along with the fact that I can use an iPad charger with the appropriate cable to keep this thing charged. In fact, for most of this business trip I haven't even bothered using the full-sized charger because I haven't needed it. Or to look at it from a different perspective, although the full-sized charger is considerably bigger than an iPad charger, it's still quite small for a charger, and that is made possible by the low power requirements of the MacBook.

    In fact, tomorrow when I go to the conference, since it is just a half day I'm not going to bother bringing any charger at all. Battery power will be more than enough for my business applications over that period.
     
  18. Mike Boreham macrumors 65816

    Joined:
    Aug 10, 2006
    Location:
    UK
    #18
    I don't go to conferences any more but I do go away for weekends to family. I take an Anker RavPower USB-PD spec battery (ie will charge MB at 30w). At night this recharges the MB. During day it is recharged by Anker 5 port USB-PD charger. Same charger charges wife's iPad and two phones at night. Anker charger can fast charge the MB directly as well. Unfortunately the Apple 29W won't fast charge the Anker, but the Anker 5 port will. Getting bit off topic here.
     
  19. kiranmk2 macrumors 6502a

    kiranmk2

    Joined:
    Oct 4, 2008
    #19
    I too lament the lack of an m3/512 GB config. The price difference between m5/256 and i5/512 of £300 is even more scandalous when you consider that Intel charges the same price for both the m3 and i5 chips!
     
  20. Loge macrumors 68030

    Loge

    Joined:
    Jun 24, 2004
    Location:
    England
    #20
    Still trying to understand the need for 16GB over 8GB on this product. I can see it either for apps that process large datasets, or music apps with a large number of virtual instruments for example, or for VMs. However for simply having a number of applications open at once, surely the fast SSD and macOS memory management takes care of that?
     
  21. EugW, Oct 20, 2017
    Last edited: Oct 20, 2017

    EugW thread starter macrumors 601

    EugW

    Joined:
    Jun 18, 2017
    #21
    SSD is still way, way, way slower than memory.

    Sure having a fast SSD will mask slowdowns from disk swap, but it's still faster not to have to swap. Plus with lots of memory, you minimize the overhead of memory compression and decompression.

    Put it this way: I'd take a US$1499 m3 with 16 GB RAM over a more expensive US$1549 i7 with only 8 GB RAM any day of the week.
     
  22. EugW thread starter macrumors 601

    EugW

    Joined:
    Jun 18, 2017
    #22
    Screen Shot 2017-10-20 at 10.50.43 PM.png
     
  23. ZapNZs macrumors 68020

    ZapNZs

    Joined:
    Jan 23, 2017
    #23
    It's very interesting to see how different the results we are getting are compared to NoteBookCheck (awesome website I trust very much). I've run the same R15 test with my 2017 core m3 and, like the original test results here, the system has NOT exhibited the extreme thermal throttling that some other reviews indicated.

    I wonder how much variation there is among the individual CPUs? Or if some sort of extraneous variable(s) is at work?
     

    Attached Files:

  24. kiranmk2 macrumors 6502a

    kiranmk2

    Joined:
    Oct 4, 2008
    #24
    Anyone got any comparable data from e.g. a 2013-onwards MBA (i5 or i7) to compare?
     
  25. EugW thread starter macrumors 601

    EugW

    Joined:
    Jun 18, 2017
    #25

Share This Page