Hardware efficiency bw Mac Pro, Mac Mini & MacBook Pro for Izzy 2.0

  • Hi All,

    I'm mounting a 2 months installation and having a hard time choosing bw two options: 
    1\. Borrowing an upgraded 2008 Mac Pro (for 350$ friendly rental price)
    2\. Purchasing a MacMini + add SSD (around 900$)

    I haven't got any chance to run tests on MacMini's i7 2012 model. But I'm running tests on the MacPro, the result is worse than my mid 2012 MacBook Pro, even with higher CPU and specs. Which makes me really confused why.

    Here are the specs:

    1\. MacPro (2008 model, upgraded)
    Processer: 2x 3.2GHz Quad-Core Intel Xeon 
    Memory: 3GB 667 MHz DDR2 FB-DIMM
    Graphics: ATI Radeon HD 5770 1024 MB SSD
    OSX 10.9.5
    2\. Mac Mini (2014, which is 2012 model)
    • 2.3GHz quad-core Intel Core i7 or
    • 2.5GHz dual-core Intel Core i5
    • 4GB memory
    • Intel HD Graphics 4000
    • OS X Mavericks

    3. MacBook Pro (mid 2012 model, upgraded to SSD)

    MacBook Pro 15in. (2012) 2.6 GHz Intel Core i7
    16 GB 1600 MHz DDR3
    NVIDIA GeForce GT 650M 1024 MB
    SSD 500mb/500mb
    MacOS X 10.7.5 Isadora 2.0

    For some reason with one of the same testing patch (which is running 9 1280 x 720 photojepg video on Movie Player), my MacBook Pro can run up to 25 FPS, and the Mac Pro can only run to 5 FPS. I've checked and see that on the MacPro, the movie player is running on "AV foundation", and the result was was poor, but in my Macbook pro it was running in "QT" automatically, with the exact same setting (ie. optimized for performance). I can't figure out why. I'm debating on buying a new MacMini instead of renting this Mac Pro, in case the i7 spec might actually run faster on Izzy 2.0 than the Mac Pro (even with faster CPU specs). I know this is getting a bit too detail, but if for any reason the new Mac Mini can run faster than the Mac Pro, I would rather invest $ into it than renting. 

    I'm running 9 (

    832 × 336 ) videos simultaneously with simple efx actors and the mapping function.

    I wonder if anyone would know? even theoretically speaking.


  • Hi Sammy,

    have you licensed the Mac Pro with your Isadora Beta Key ?
    as the Betaversions doesn't support a demo of the new features
    i would be surprised if you could compare the results …
    the Mac Pro is a good one but needs more ram and eventually a better gnu 
    + you shall convert your footage to prores
    even when working in the "old" 1.3 version we had good results with 6 streams of prores lt 720p and CI actors on a Mac Pro 2009
    when working in beta2 prores makes really sense as the native av-foundation format
    but in the end it comes down which effects you want to use
    if they are cpu (video) effects then performance will drop significantly
    hope this helps
    best clemens

  • Hi Clemens,

    Thank you for the help.

    I have an additional USB key, and it was licensed properly (the mapping function works and the background is the new colour, so I'm pretty sure it's working in 2.0 properly).

    I'm using the Slide actor, which is using cpu, I believe. So I'm guessing ProRes wouldn't be a better option than PhotoJpeg?

    What I don't understand is that why would the Mac Pro run slower than the Mac Book Pro with the exact same patch and videos, and why wouldn't it enable "QT" playback engine, which I suspect is the main reason why it plays back in such a poor result.

  • I have an old 8 core Mac Pro 3.0GHz sat in the back of my office and it is easily beaten by the 2012 Mac Mini in CPU benchmarks like Cinebench. It's virtually useless for anything other than word-processing by modern media creation standards.

    The new CPUs provide much grunt for a given GHz than older generation CPUs even with many less cores. Your MBP CPU probably also has quite a healthy Turbo up clock and has a much more efficient data path from disk to CPU to GPU regardless of which playback engine is being used. The Mac Pro is critically low on memory too.
    I'm very interested in what others advise as I'm currently pitching an audio reactive installation and I'd be interested in how well the Mac Mini or iMacs are for Isadora.

  • Hi,
    My home Mac is a 2014 MacMini Core i7 2,3GHz - 16 Go Ram - 256Gb SSD drive, with Isadora 2.0.0b12 licensed installed.
    How many video outputs do you use with your 9 simultaneous movies ?

    If you send me your patch and videos, I can test it for you and give you feedbacks.


  • @Unfenswinger 
    thank you. that's a really good point. I did some research on benchmark comparison bw the two, and Mac Mini did won over the 2008 Mac Pro on CPU, but everything else (GPU, Graphic card, etc) seems like the 2008 Mac Pro still wins. But like you said, what's more important is how Izzy 2.0 uses its power, and if a healthier Turbo up clock in CPU matters more, then that's what i should be investing.

     that's amazing, Philippe. you have the exact model of Mac Mini that i would like to test on.

    I've compiled the test patch and videos here: (should take another 2 hours to upload, the internet i have here is very slow, still on a travel right now)


    please let me know what's your FPS test results, and if you can double check the output video to see how smooth it's actually running despite the FPS result. (there are 4 different scenes to tests on)

    Thanks so much! This community is amazing isn't it?


  • For the moment I only see a .rtf file in the folder of the link...
    an you send me you preferences settings too ? (framerate, resolution of the project...)


  • @fifou
    Hi Philippe,
    I left it uploading while I went to another city, I just got back and I found out the upload got interrupted! sorry about that!

    I've uploaded to another direct link: (I've watched it until it finished uploading this time.. :) )
    https://dl.dropboxusercontent.com/u/16722659/CPU tests.zip

    on preference setting: 30 FPS, 1280 x 720, scale to smallest size

    thank you!!!

  • Here's my score for a 2010 12 core Mac Pro 2.93GHz 24GB nVidia 680 (PC) running on a RAID 0 2 x Caviar Blacks.

    Scene 1, 17-18 fps.
    Scene 2, 26.7 fps.
    Scene 3, 49.8 fps.
    Scene 4, 46-48 fps.
    I recreated Scene 1 using vid-gpu and texture projectors and got a solid 60 fps (target 100). So the moral of this story is to stay on the GPU. The 680 isn't even a great GPU and it wasn't complaining and it also showed the test wasn't disk bound.
    I re-encoded the video clips to ProRes LT, Prores, and HAP and actually the photoJPeg originals performed better in your scene. Using vid-gpu it didn't make any difference which codec was used playback was a solid 60 fps.
    Hope that adds a bit of info to the discussion. I know it's not a Mac Mini but it does show the difference between GPU and CPU quit markedly.

  • Hi Sammy,
    Here are my results :

    MacMini late 2012 - 2,3GHz Intel Core i7 - Intel HD Graphics 4000 1024 Mo - 16 Go 1600 MHz DDR3 RAM - Crucial 256Go SSD - OS X 10.9.5

    Running registered Isadora v2.0.0b12, using HDMI output (19201080) for desktop + Isadora interface.
    And using MiniDisplayPort output to 1360
    768 VGA screen (I know it's an old weird lcd screen !)  Stage 1 output.
    Settings in Isadora are 30FPS - 1280*720 - Scale to smallest.

    Scene One : CPU 720 PJ (tried several times : go blank - go back to scene)
    Cycles : around 8
    FPS : around 4

    Scene Two : CPU 720 +efx
    Cycles : around 14
    FPS : around 7

    Scene Three : 16+2 vid map
    Cycles : 66-96
    FPS : 27-29,9

    Scene Four : 16+1 vid map +
    Cycles : 60-80
    FPS : 29,7-30

    The smoothness of the images is really a subjective thing. I cannot say if the rendering of a scene is smooth enough for you. I would say scenes 1 & 2 are too much lagging. And scenes 3 & 4 are acceptable with little random stutters.

    I’ve redone your first scene with texture projectors (vid-GPU)
    cycles : 190-228
    FPS : solid 30 !

    I would definitely go for the vid-GPU option.

    Hope these tests will help you to make your choice.


  • Hi Philippe,

    Would you be able to repeat the vid-gpu test with the target fps at something like 100 so we can see just what level of GPU performance the Mac Mini maxes out at if you can spare the time?
    Seeing your results I'm hoping that I can specify a Mac Mini for an upcoming project if I win the pitch and would be good to see the max headroom.
    All the best

  • MacMini late 2012 - 2,3GHz Intel Core i7 - Intel HD Graphics 4000 1024 Mo - 16 Go 1600 MHz DDR3 RAM - Crucial 256Go SSD - OS X 10.9.5
    Target Framerate : 120fps

    Scene One : around 4fps
    Scene two : 4-8fps
    Scene Three : 29-32fps
    Scene Four : 30-40fps

    Scene One (vid-GPU) : 55-60fps

  • Thank you guys so much for doing this. This is very very helpful.

    Yes, the vid-gpu seems like the best option for both the Mac Mini & Mac Pro, if we don't require using any video efx actors.

    I am very curious why my MacBook Pro mid 2012 can run both Scene One & Scene Two with such a dramatically better result. (around 27-29 FPS) Perhaps my SSD is 550mb/500mb speed compared to some are 300mb/250mb speed?


  • Izzy Guru

    Just read this thread, first:

    1) Woohoo go isadora community! Really nice to see everyone helping out here, I know it happens a lot but this is really cool. We (izzy team) try to answer as much as we can on here but during busy periods it's hard to keep up. So thanks to everyone for chipping in and helping.
    2) **GPU** is the way to go here. Much more efficient. Also I see you have SSD's also which really helps.

  • I don't understand too why the first two scene have so poor results comparing to my macbook Pro.
    I don't think it's an SSD bandwith limitation since it works well in vid-GPU mode.
    Maybe it is something related to the Intel HD Graphics 4000 ?

    On the first scene, if I switch all players performance from "performance" to "interaction" (reading movies with QT instead of AVFoundation) the fps jumps from 4-8 (performance) to 7-9 (interaction)


  • I think it's easier to explain the performance of your MBP compared to my 12c Mac Pro than it is to explain the difference between the MBP and the Mac Mini. The MBP has a very healthy turbo on more efficient cores which beat an ageing Xeon.

    Looking at Intel's Arc for the processor details the CPUs in the Mac Mini lack the SSE 4.1/4.2 instruction sets that are in the MBP's CPU and it has AVX vs AVX 2.0\. So it seems the MBP has a lot of extra instructions that a re really making a huge difference in this instance.
    Thanks for posting these scenes for testing. As a noob to Isadora this was very interesting.

  • Thank you again @fifou for doing such a detailed hardware, CPU, GPU tests. it's really a huge help.

    @Unfenswinger you really have some amazing knowledge for these technical specs! That really makes a lot more sense then. I'm so glad that it's also interesting for you.

     I just think that it might be helpful for anyone who is thinking of investing computers for their installations or performance work that has specific demands on the processing. The more we know the about them, the more we can save our money and optimize our needs and make the best choices. This can go pretty deep. Sharing them would save more people time and trouble :)

    yes, what an amazing community we're in!

  • You're welcome !

    All the best you all.

  • Hi Sammy,
    Can you drop a quick summary of your discoveries here, at the end of this post? Just to make it easier to people to grab the conclusion you came to.
    I have also had excellent results with a 2012 Mac MIni outputting HDMI 3D video projections, (double-HD top/bottom video streams). When we initially researched the project, Apple infact advised us we couldnt get the results that we did in the end actually achieve from the mac mini. It was impressive. And affordable!