Photoshop Benchmarks bTB 13 vs mTB 13 vs mTB 15

Discussion in 'MacBook Pro' started by nlupient, Jan 13, 2017.

  1. nlupient macrumors member

    Joined:
    Oct 31, 2014
    #1
    I was curious to know how much of a difference there was between a base TB 13", maxed out TB 13" and a maxed out TB 15". I used this benchmark https://macperformanceguide.com/OptimizingPhotoshopCS6-Benchmarks.html. Attached is a PDF comparing the base TB 13", a maxed out TB 13", and a maxed out TB15" with a 1TB PCIe. I ensured all computers had the same background activity going on and used the same photo for the batching tests.
     

    Attached Files:

  2. Sanpete macrumors 68020

    Joined:
    Nov 17, 2016
    Location:
    Utah
    #2
    Wow. Well worth it to get the 15" for those jobs.
     
  3. jerryk macrumors 68030

    Joined:
    Nov 3, 2011
    Location:
    SF Bay Area
    #3
    Cool benchmarks. Did you run that with CS6? I running the latest CC and want to ensure I can compare my numbers to yours.
     
  4. nlupient thread starter macrumors member

    Joined:
    Oct 31, 2014
  5. campyguy macrumors 68040

    Joined:
    Mar 21, 2014
    Location:
    Portland / Seattle
    #5
    I'm confused, and not being a PITA here. Comparing a new MBP, or any MBP to the Mac Pro with a dedicated GPU and at least 64GB of RAM seems to be a bit of a WTF? comparison with no real basis. That, and the "benchmark" author has no indication of whether a high-speed scratch disk is being used - PS 101 IMHO - and wasting internet space alluding to disabling App Nap. And, PS6 is several iterations old - never mind the whole CC subscription bit that so many tend to gripe about, but the benchmark machine being linked to is a $5k+ Mac, running a 4-1/2-year-old OS. Comparing CC 2017 to CS 6? A head-scratcher, kind of racing a '67 Chevy Impala against a '17 Chevy Impala - same name, different beast… o_O

    Just get a fast external scratch disk, just like Adobe recommends - I've posted this several times here, and stop worrying about benchmarking. Read Adobe's own recommendations - the Mac Performance Guide was and is old, stale junk… Or just ignore this.
     
  6. nlupient thread starter macrumors member

    Joined:
    Oct 31, 2014
    #6
    I am sorry to have inconvenienced you with my lack of Mac OS/Photoshop/Whatever else knowledge. I'm just an ill-informed amateur that was interested in knowing the differences and thought some fellow Macrumorians might also want to know. Please forgive me. :)
     
  7. campyguy macrumors 68040

    Joined:
    Mar 21, 2014
    Location:
    Portland / Seattle
    #8
    Why? I wasn't inconvenienced at the time, it seems as my post here was the only helpful one. I'll be sure to steer clear in the future… :rolleyes:
     
  8. jerryk macrumors 68030

    Joined:
    Nov 3, 2011
    Location:
    SF Bay Area
    #9
    Strangely I could only turn the 1st 3 test you mentioned on my 2015 15" rMBP 512 GB. The others threw errors. Running CC 2017. Here are my numbers.

    speed1 11.49
    Med 14.65
    Huge 154.8
     
  9. yhc macrumors newbie

    Joined:
    Jan 14, 2017
    #10
    Thanks for sharing the data.

    Regards,
     
  10. jerryk macrumors 68030

    Joined:
    Nov 3, 2011
    Location:
    SF Bay Area
    #11
    The benchmarks are a valid comparison between the 3 machines, 2016 13" TB Base, 13" TB Max, and 15" TB Max. All were set up the same. Same version PS 2107 CC, running with a local SSD. Running on a dedicated external drive is not how most people use their laptops, and a USB scratch drive would be much slow than the SSDs in these machines.
     
  11. Sanpete macrumors 68020

    Joined:
    Nov 17, 2016
    Location:
    Utah
    #12
    Your numbers indicate the same thing Diglloyd found, that for his Photoshop benchmarks the new MBP is slower than the old. I've seen different results from other sources doing Photoshop jobs, but no doubt his benchmarks do show something significant.
     
  12. kokhoong0624 macrumors regular

    kokhoong0624

    Joined:
    Aug 23, 2015
    #13
    Thanks for making this :)
     
  13. jerryk macrumors 68030

    Joined:
    Nov 3, 2011
    Location:
    SF Bay Area
    #14
    I had not heard about the 2015 being faster than the 2016 in PS. I was a bit shocked by my results and ran the epoch of 10 runs multiple times, with reboots in between, to ensure what I was seeing was reproducible.

    But, I am still surprised to the see that the 2015 15" with 512 GB system was up to 25% faster than the 2016 15" Maxed with 1TB.
     
  14. leman macrumors 604

    Joined:
    Oct 14, 2008
    #15
    If this is the case then either a) there is something very wrong with Photoshop code b) something very wrong with the benchmark or c) Photoshop hits a bug in the driver code for the 2016 hardware. But again, given the quality of Adobe software, nothing surprises me.
     
  15. Sanpete macrumors 68020

    Joined:
    Nov 17, 2016
    Location:
    Utah
    #16
    It could be that Photoshop is better optimized in some ways for the 2015 CPU. I saw another test where the 2016 was a little faster than the 2015 doing a Photoshop job, but I don't recall where.
     
  16. jerryk macrumors 68030

    Joined:
    Nov 3, 2011
    Location:
    SF Bay Area
    #17
    Or Apple screwed up something in the hardware and software.
     
  17. leman macrumors 604

    Joined:
    Oct 14, 2008
    #18
    Well, given that everything else is faster on the 2016 machine, my bet is on Adobe ;)
     
  18. campyguy macrumors 68040

    Joined:
    Mar 21, 2014
    Location:
    Portland / Seattle
    #19
    You've missed something, not reading my post fully - the post I was addressing was this one link: https://macperformanceguide.com/OptimizingPhotoshopCS6-Benchmarks.html - in Post #6. So, no, they're not a valid comparison. As to how people run their laptops, maybe keep to the people you know and associate with - everybody I know and associate with uses a scratch disk with AutoCAD, VectorWorks, PS (I've been using a scratch disk with AutoCAD since 1995 and with PS since version 3 in the late '90s) and all 53 PCs and Macs in my SMB have attached fast DAS for scratch disks. No offense, most of the people I know use their software properly, and I'm here to extol how to get a leg up or work more efficiently. I've got a 2 1TB SSD RAID 0 (dual Sammy 850 Pros) for my desktop in my main office and an AKiTiO Palm RAID for the road - they're dedicated scratch disks… Cheers…
     
  19. CaptRB macrumors 6502a

    CaptRB

    Joined:
    Oct 11, 2016
    Location:
    LA, California
    #20
    Tests like this don't mean much.

    Being a working photographer, shooting still life or whatever for money and then processing and printing is the real benchmark.

    I'm shooting D810 and D800, 36 MP cameras and using PS and Lightroom along with several plug-ins. I usually process with 27" Dell 4K@60.

    I'm using the 8 GB 13" tMBP, 15" rMBP (2015) and just got rid of a 2011 maxed out quad core i7 iMac. They all do a fine job and have no trouble with software or file size.

    The new tMBP is the nicest and most portable for photography, but I'll likely add the 15" tMBP on the next cycle.


    Cheers,


    Robert
     
  20. evec macrumors regular

    Joined:
    Jun 8, 2016
    #21
    People already thing new CPU will faster old CPU, sorry that not already true.
    The problem is Intel want cost down for sale more, and Skylake is cost reduced chip.
    So in some usage not relative to famous benchmark such as Photoshop custom filer or programming, Intel eliminate some circuit and the speed will come down.
     
  21. jerryk macrumors 68030

    Joined:
    Nov 3, 2011
    Location:
    SF Bay Area
    #22
    What specifically in that link invalidates using the test on other versions of the PS?

    As the post states, it tests "
    • Whether the amount of installed memory is adequate.
    • The effect of one scratch drive versus another (when memory is low or maxed-out).
    • Maximum possible performance (assuming ample memory and fastest scratch drive).
    "

    For all machines tested we are in the 3rd case.

    And speaking of scratch disks:
    The Samsung 850 has a 0.5 GB/sec max transfer rate. The new MBP 15 drive is in excess of 3.0 GB/sec. Over 6 times as fast. Even on it's best day your raid 0 array is around 0.8 GB/sec. Slower than the 1.6 GB/sec in the 15" 2015 rMBP SSD. This is due to a number of factors, including the 850 using the SATA interface versus the new MacBook pros using higher performance PCIe interface. You can see the difference yourself if you compare the specs for the Samsung 850 and Samsung 960 Pro.
     
  22. Miltz macrumors 6502a

    Joined:
    Sep 6, 2013
    Location:
    New York
    #23
    I'll Chime in here with my results. I have a i7 2700k Windows desktop with 16GB of RAM and 1TB Samsung 850 SSD. I ran a few scripts that I normally do in photoshop on both computers and my fully loaded MacBook Pro 2016 15" is faster without question. In some other software it's even twice as fast. I think adobe isn't really taking advantage of the new 2016 models yet. Ultimately people don't realize that Intel's CPU aren't getting faster every year as we were accustomed to. For example Kaby Lake isn't any faster than Skylake, just more energy efficient according to intel. So when Intel is telling you their new CPU isn't faster, why would people expect it to be?
     

Share This Page