2010 Mac Pro Upgrades

Discussion in 'Mac Pro' started by xWhiplash, Nov 21, 2016.

  1. xWhiplash macrumors 65816

    Joined:
    Oct 21, 2009
    #1
    Hello,

    While I wait for a newer Mac Pro, I would like to upgrade my current one. I find that my 2013 rMBP is able to work much faster in Final Cut Pro X than my 2010 Mac Pro can. My 2010 Mac Pro has a PC NVIDIA GTX 980 card in it which might be causing an issue. Would a Sapphire HD 7950 be better than my GTX 980 for FCPX and things like that? How much better would we be talking about, would it be worth $500?
     
  2. pat500000 macrumors 604

    pat500000

    Joined:
    Jun 3, 2015
  3. xWhiplash thread starter macrumors 65816

    Joined:
    Oct 21, 2009
    #3
    What is strange is that my 2013 rMBP has lower clock speed, fewer cores, and weaker GPU compared to my 2010 Mac Pro. Is the Xeon just THAT old? It is the drivers for my GTX 980?
     
  4. pat500000 macrumors 604

    pat500000

    Joined:
    Jun 3, 2015
    #4
    From the YouTube search, it appears that Radeon series are better with fcpx. I could be wrong.
     
  5. ActionableMango macrumors 604

    ActionableMango

    Joined:
    Sep 21, 2010
    #5
    Your rMBP has an Intel module in the CPU called Quick Sync that is specifically designed to work video, and FCPx makes use of it.
    https://en.wikipedia.org/wiki/Intel_Quick_Sync_Video

    My understanding however is that Quick Sync does trade quality for speed, so it is more "lossy" than when not using it.
     
  6. xWhiplash thread starter macrumors 65816

    Joined:
    Oct 21, 2009
    #6
    Yeah I had a feeling that it was something to do with the processor generation. Does that mean my rMBP will always be faster?
     
  7. ActionableMango macrumors 604

    ActionableMango

    Joined:
    Sep 21, 2010
    #7
    Well first we can check to see if you are in the ballpark or if something is wrong. Can you do the BruceX test and report the result:
    https://blog.alex4d.com/2013/10/30/brucex-a-new-fcpx-benchmark/

    Be sure to follow all steps or the result can be off. And if you want to repeat the test, completely close FCPx and restart it before trying again.

    Then you can compare to other results in this thread:
    http://forums.macrumors.com/threads/fcpx-amd-vs-nvidia.1956128/page-2

    Also, if you are not on ElCap or newer yet, you should upgrade. The switch to Metal will result in cutting times by almost half. For example I have a GTX980 and simply moving from Yosemite to ElCap dropped my BruceX time from 52.33 seconds to 27.20 seconds.
     
  8. xWhiplash thread starter macrumors 65816

    Joined:
    Oct 21, 2009
    #8
    Great, I will give that a try! I am on Sierra now.

    One thing I will mention, my OS X drive started with Mavericks, then went through all the OS upgrades over the years (I never skipped an OS release). Would formatting and installing from a fresh Sierra install help any?
     
  9. xWhiplash, Nov 21, 2016
    Last edited: Nov 21, 2016

    xWhiplash thread starter macrumors 65816

    Joined:
    Oct 21, 2009
    #9
    Well my 2010 Mac Pro takes 53 seconds, but my 2013 rMBP takes 1 minute, 27 seconds.

    I mostly use "Send to Compressor" so I can get the video in an MP4 format. In that scenario, my 2013 rMBP is faster than my 2010 Mac Pro. Is this what is giving me problems?

    On my 2010 Mac Pro using Compressor, with 23.976 fps at 1080p at 12,000 kbps it took 1 minute 8 seconds. On my 2013 rMBP it took 1 minute and 50 seconds. I am not sure why my 2013 rMBP is faster on some projects.
     
  10. ActionableMango macrumors 604

    ActionableMango

    Joined:
    Sep 21, 2010
    #10
    Quick Sync only accelerates H.264.

    BruceX exports to ProRes. The MP is faster than the MBP because its hardware is faster and Quick Sync does not come into play.

    I am guessing for Compressor you are using H.264 codec in your MP4 files. So in this case, the MBP's CPU has the advantage because of its Quick Sync module. This is normal.

    You can try to brute force past the Quick Sync advantage by upgrading your MP, but I don't know what it would take or if it is possible. However, there is definitely room for improvement because your BruceX is 53 seconds and mine is 27.20, both on MP 2010. I'm running a W3690 CPU, GTX980 GPU, and 16GB RAM.

    If you have similar hardware then something is wrong and maybe a fresh install is warranted like you mentioned earlier. Your 53 second time is suspiciously close to my 53 second time pre-ElCap.
     
  11. xWhiplash thread starter macrumors 65816

    Joined:
    Oct 21, 2009
    #11
    Is it a problem with Sierra? I have the Intel(R) Xeon(R) CPU W3680 @ 3.33GHz, with a GTX 980 (non flashed and Windows PC version), and 32GB of RAM.
     
  12. ActionableMango macrumors 604

    ActionableMango

    Joined:
    Sep 21, 2010
    #12
    I don't know, but it would appear that something is wrong because our hardware is very similar. If you have a spare drive you could try a fresh install on it without disrupting your current install.
     
  13. xWhiplash thread starter macrumors 65816

    Joined:
    Oct 21, 2009
    #13
    Do you have a SATA 2 SSD, or are you using a PCIe SSD? Would that matter at all?
     
  14. ActionableMango macrumors 604

    ActionableMango

    Joined:
    Sep 21, 2010
    #14
    My 53 second time was on a SATA2 SSD.
    My 27 second time was on a SATA2 2.5" 5400rpm HDD.

    So no, I don't think it has anything to do with the drive. ;)
     
  15. xWhiplash thread starter macrumors 65816

    Joined:
    Oct 21, 2009
    #15
    I really do not mind reinstalling from scratch. Do you think it would be beneficial anyway regardless if it helps FCPX?
     
  16. pat500000 macrumors 604

    pat500000

    Joined:
    Jun 3, 2015
    #16
    Do you use handbrake?
     
  17. CocoaNut macrumors member

    CocoaNut

    Joined:
    Sep 8, 2011
    Location:
    Switzerland
    #17
    I'm also interested in knowing more. I have a 2010 maxed-out MP, with 2x3.33 6c CPUs, 64GB of RAM, Radeon HD 7950, and two 1TB drives, one SSD (system) and one HDD. Handbrake on this machine is quick, but sometimes it does seem sluggish (i.e. slower-than-expected), like moving between apps.

    It's currently still running El Capitan, so I'd like to ask if a fresh install of Sierra would perhaps help?
     
  18. xWhiplash thread starter macrumors 65816

    Joined:
    Oct 21, 2009
    #18
    No, just FCPX and Compressor
     
  19. ActionableMango macrumors 604

    ActionableMango

    Joined:
    Sep 21, 2010
    #19
    I don't know.

    That's normal for Handbrake on OS X. However on Windows it does a much better job of acting low priority.
     
  20. xWhiplash thread starter macrumors 65816

    Joined:
    Oct 21, 2009
    #20
    Well I spent the last few hours reinstalling macOS Sierra. It is still taking close to 1 minute on that test :(. What else should I look at?
     
  21. pat500000 macrumors 604

    pat500000

    Joined:
    Jun 3, 2015
    #21
    You should look into rx 480.
     
  22. xWhiplash thread starter macrumors 65816

    Joined:
    Oct 21, 2009
    #22
    What would be a benefit of that? Would it be better than my 980? It seems like my issues are processor based for some reason :(
     
  23. pat500000 macrumors 604

    pat500000

    Joined:
    Jun 3, 2015
    #23
    It does better with fcpx. I was looking into it via YouTube.

    Your issue is the performance, right?
     
  24. ActionableMango, Nov 22, 2016
    Last edited: Nov 22, 2016

    ActionableMango macrumors 604

    ActionableMango

    Joined:
    Sep 21, 2010
    #24
    Not sure. Basically the switch from Yosemite to ElCap lowered everyone's score by about half, due to FCPx using Metal. Your speed is like my pre-ElCap non-Metal speed.

    If you have a spare drive and can install Yosemite on it to do BruceX there, the resulting time would tell us a lot. For example, if it's 53 seconds there too, then Metal is not working for you for some reason. If the resulting time is closer like 2 minutes, then you are seeing the "cut in half time" but there's some bottleneck occurring that's slowing everything down.

    A much easier way would be to try out that Metal benchmark that people were using here, but I cannot find it right now because I cannot remember the name.

    EDIT: I just found it...
    http://forums.macrumors.com/threads/metal-vs-opengl-benchmark.1957306/

    I'll run it when I get the chance, but family just flew in for the holiday weekend and I'm very busy.
     

Share This Page