To help people work out which Macs work best with Final Cut Pro X, it is useful to refer to a standard speed test.
In the past we've had the 'Far Far Away' test, in which you time how long it takes to render the 10 second long standard 'Far Far Away' title. This was useful to test to see if Mac's memory or GPU configuration was good enough to run Final Cut Pro X well.
The 'Far Far Away' test showed that my 2009 MacBook Pro was severely limited by only having 4GB of RAM installed. Due to a bug in 10.0.9, it took
79 minutes. After I upgraded the memory to 8GB, the same test took 65 seconds.
With more recent Macs, the test is proving less useful. In another topic 'qbe' asked me to do the same test on my new Late 2013 MacBook Pro using first the internal Iris Pro graphics system, then using the discrete GeForce GT 750M 2GB GPU. The results seemed counterintuitive:
MacBook Pro late-2013 Discrete GeForce GT 750M 2GB
18.8 seconds
MacBook Pro late-2013 Intel Iris Pro 1GB
19.2 seconds
Could there be such a small advantage in using a specialised GPU? In this case there is.
On the one hand, early testing shows that Iris Pro graphics are better than many expected. Take a look at a new post at Bare Feats:
In the past we sneered at the integrated GPUs and their puny performance. Not any more. The Intel Iris and Iris Pro are every bit the match or master of discrete NVIDIA Mobile GPUs -- at least when it comes to OpenCL acceleration.
from
www.barefeats.com/rmbpc.html
Although it is great news that integrated GPUs are getting better, many are worried that the MacBook Pros should be avoided until Iris Pro has improved a little more.
In practice, the 'Far Far Away' test didn't stress the integrated GPU in the MacBook Pro 15" enough.
I've come up with a test that shows the differences between these GPUs and other Macs...
The BruceX benchmark
BruceX is a small Final Cut Pro X XML file that you import into Final Cut Pro. It creates a very short timeline at the highest possible standard resolution that Final Cut can handle: 5120 by 2700 (at 23.975 fps). It uses standard Final Cut generators, titles and transitions. As it uses many layers of complex content, it requires lots of GPU RAM.
The benchmark is based on timing how long a Mac configuration takes to export the project to disk.
To use this import the XML file at
Alex4D.com/BruceX_Test.zip
...and time the export of a 5K master file from the timeline.
1. Have both QuickTime player and Final Cut Pro X open at the same time.
3. In Final Cut Pro X, go to 'Final Cut Pro:Preferences…' – in the Playback tab make sure 'Background Render' is off.
3. Use the 'File:Import:XML...' command to import 'BruceX Test - 5K.fcpxml' to create a very short but complex 5K project.
4. Click the new 'BruceX Test - 5K ' timeline (this makes the Share command selectable)
5. Export the QuickTime movie by choosing "File:Share:Master File...'
6. In the dialogue box that appears, click the 'Settings' Tab
7. In the 'Video Codec' section choose 'H.264'
(edited to add - many people have got errors with H.264, so use ProRes422 instead you you have a problem - some times are given with different flavours of ProRes, but each person usually mentions which they've used when reporting their results)
8. In the 'Open With' section, choose 'QuickTime Player'
9. Click the 'Next' button in the bottom-right of the dialogue box
10. In the Save sheet, choose a name and location for the export - export to your fastest drive connected using your fastest connection.
11. Get your stopwatch ready and time from when you click 'Save' until you see the movie open up in QuickTime Player.
12. Quit and restart Final Cut to clear some cached renders and time the export again. Quit and restart Final Cut again, export a third time. The BruceX value is the average of the three export times.
My results
iMac mid-2011 27" 16GB RAM / Radeon HD 6970M 2 GB
73 seconds
MacBook Pro late-2013 Discrete GeForce GT 750M 2GB
88 seconds
MacBook Pro late-2013 Intel Iris Pro 1GB
163 seconds
( To choose which GPU to use for the test on my MacBook Pro with more than one, I used gfxCardStatus by Cody Krieger - a Shareware app downloaded from
gfx.io )
Here's a summary of the results so far (including those sent to me directly by Rob-ART of
Bare Feats):
Hope people find this a useful way to compare Mac configurations,
@Alex4D