Simple long CPU benchmark to test m3,i5,i7

Discussion in 'MacBook' started by moshen, Jul 6, 2017.

  1. moshen, Jul 6, 2017
    Last edited: Jul 8, 2017

    moshen macrumors member

    Joined:
    Nov 6, 2005
    #1
    Hi guys, I just got a 2017 MacBook with i7 1.4ghz 16GB RAM to replace my m7 1.3ghz 2015 MacBook. I ran a 20minute benchmark to try to compare speeds in longer CPU tests.

    Using p7zip benchmark mode I find a good 26.5% increase in CPU performance between my 2017 & 2015.

    It'd be useful to get numbers for the 2017 i5 & m3 so we can compare sustained CPU benchmarks. You can do this if you're comfortable in a command line terminal shell.

    1). Install p7zip but first install home-brew (https://brew.sh) by running the first command on the command line.

    a) /usr/bin/ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)"
    b) brew update
    c) brew install p7zip

    2). Close all apps and let your MacBook idle for five minutes or so to cool off from the above.

    3). Run the benchmark on the command line: 7z b 3 -mm=\*

    This takes about 15-20minutes (we can make it even longer by replacing the 3 with more iterations - but I think it's long enough).

    Copy paste the last few lines. Here's what I got:

    2017 Macbook 1.4ghz i7 16GB Ram

    Method Speed Usage R/U Rating E/U Effect
    CRC32:1 1257327 391 2343 9153 83 323
    CRC32:4 3379867 394 1914 7544 68 267
    CRC32:8 4597255 392 1589 6234 56 220
    CRC64 3030705 391 1589 6207 56 219
    SHA256 263625 392 1373 5378 49 190
    SHA1 688859 393 1640 6448 58 228
    BLAKE2sp 454560 393 2544 10000 90 353
    CPU 393 2599 10217
    ------------------------------------------------------
    Tot: 390 2217 8640 78 305

    2015 Macbook 1.3ghz m7 8GB Ram

    CRC32:1 935489 390 1747 6810 81 314
    CRC32:4 2285832 388 1313 5102 61 235
    CRC32:8 3269256 391 1133 4433 52 204
    CRC64 2215723 387 1174 4538 54 209
    SHA256 187292 387 986 3821 45 176
    SHA1 446750 390 1072 4182 49 193
    BLAKE2sp 368354 392 2066 8104 95 374
    CPU 387 1834 7103
    ------------------------------------------------------
    Tot: 389 1752 6810 81 314

    Using the second number: 2217/1752 = 26.5% faster for the 2017.
     
  2. Graham Perks macrumors member

    Joined:
    Oct 2, 2003
    Location:
    Austin, TX
    #2
    A 2016 m5:

    Tot: 386 1944 7510 97 375

    At 1944, close to the middle of your results.
     
  3. moshen thread starter macrumors member

    Joined:
    Nov 6, 2005
    #3
    Cool! Let's get some 2017 m3, m5 models in here..
     
  4. c0ppo macrumors 6502a

    c0ppo

    Joined:
    Feb 11, 2013
  5. Graham Perks macrumors member

    Joined:
    Oct 2, 2003
    Location:
    Austin, TX
    #5
  6. benoitc macrumors newbie

    benoitc

    Joined:
    Aug 18, 2015
    #6
    what anout the throttling ? did tou experiment any "pause" ?
     
  7. moshen thread starter macrumors member

    Joined:
    Nov 6, 2005
    #7
  8. Warhawk15, Jul 8, 2017
    Last edited: Jul 8, 2017

    Warhawk15 macrumors regular

    Warhawk15

    Joined:
    Feb 1, 2012
    #8
    2017 i5 with 8gb RAM

    CRC32:1 1132625 392 2101 8246 85 335

    CRC32:4 3055651 395 1729 6820 70 277

    CRC32:8 4416635 394 1519 5989 62 243

    CRC64 2823604 393 1470 5783 60 235

    SHA256 248260 395 1282 5065 52 205

    SHA1 646394 394 1534 6050 62 245

    BLAKE2sp 401449 393 2245 8832 91 358


    CPU 394 2300 9056

    ------------------------------------------------------

    Tot: 394 2058 8101 83 329


    Not sure what it all means tho...
     
  9. moshen thread starter macrumors member

    Joined:
    Nov 6, 2005
    #9

    Thanks for running it! Looks like in a sustained CPU test the i7 is about 7.7% faster. So faster but most likely the i7 upgrade is not too noticeable in real world use :)
     
  10. EugW, Jul 9, 2017
    Last edited: Jul 9, 2017

    EugW macrumors 68000

    EugW

    Joined:
    Jun 18, 2017
    #10
    How do you read the results?

    Also, aren't some of those actions hardware accelerated?

    As for Home Brew, is there is uninstaller for it?
     
  11. Scott6666 macrumors 65816

    Scott6666

    Joined:
    Feb 2, 2008
    #11
    Original 2015 Macbook 1.2 GHz Core M

    CRC32:1 1017011 381 1945 7404 89 338
    CRC32:4 2649179 378 1564 5913 71 270
    CRC32:8 3616936 382 1285 4905 59 224
    CRC64 2393612 380 1289 4902 59 224
    SHA256 204125 382 1090 4164 50 190
    SHA1 524864 384 1280 4913 58 224
    BLAKE2sp 364572 385 2085 8021 95 366

    CPU 384 2137 8207
    ------------------------------------------------------
    Tot: 384 1855 7127 85 326

    Doesn't seem like the 2017 is much faster. Ran without being plugged in.

    I was thinking of upgrading but maybe its not worth it.
     
  12. Graham Perks macrumors member

    Joined:
    Oct 2, 2003
    Location:
    Austin, TX
    #12
    Compared to my 2016 m5, the 2017 i5 is 6% faster, the 2017 i7 14% faster. At multicore, 8% and 15% respectively.

    As far as I can tell, this is purely a CPU test; won't touch the SSD, GPU etc.

    Homebrew uninstall, from http://docs.brew.sh/FAQ.html In a nutshell, run

    Code:
    ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/uninstall)"
     
  13. moshen thread starter macrumors member

    Joined:
    Nov 6, 2005
    #13

    Interesting - your 2015 1.2ghz benchmarked better than my 2015 1.3ghz. You'll get about 20% more CPU in the 2017 with the i7 upgrade. Also remember, this is sustained CPU load where thermals throttle down performance. You'll get better than 20% in typical shorter loads. My geekbench results are 30% faster in single core and 40% faster multi core.

    Also the 2017 will have massively improved SSD speeds. The value of the upgrade depends on your usage patterns. It was definitely worth the upgrade from my 2015 1.3ghz to my 2017 i7.
     

Share This Page