SHOOTOUT:
Five models of 2013 Mac Pro
Posted Friday, May 16th, 2014 by rob-ART morgan, mad scientist
You've seen our articles featuring various models of new 2013 Mac Pro. Today we give you a comparison of all models from 4-Core to 12-Core using Pro Apps. And a 4-Core iMac is thrown in for good measure.
GRAPH LEGEND
12c nMP 2.7 D700s = 2013 Mac Pro 2.7GHz 12-Core, 64GB 1866MHz DDR3, dual FirePro D700 GPUs
10c nMP 3.0 D700s = 2013 Mac Pro 3.0GHz 10-Core, 64GB 1866MHz DDR3, dual FirePro D700 GPUs (courtesy of EditBuilder)
8c nMP 3.0 D700s = 2013 Mac Pro 3.0GHz 8-Core, 64GB 1866MHz DDR3, dual FirePro D700 GPUs
6c nMP 3.5 D500s = 2013 Mac Pro 3.5GHz 6-Core, 64GB 1866MHz DDR3, dual FirePro D500 GPUs
4c nMP 3.7 D300s = 2013 Mac Pro 3.7GHz 4-Core, 64GB 1866MHz DDR3, dual FirePro D300 GPUs
4c iMac 3.5 G780m = 2013 iMac 3.5GHz 4-Core i7, 32GB 1600MHz DDR3, single GeForce GTX 780M GPU
DaVinci Resolve 10.1.5
We imported a 289 frame RED Digital Cinema clip (4800x2700 with 5:1 compression ratio). We playback the R3D clip with 2 blur nodes at "Half Good" playback quality. We wanted to see which models of Mac Pro fall short of the target 'real time' 23.976 FPS playback rate. (RED graph bar means FASTEST in FPS.)
Shifting gears, we used the "Deliver" function to export the R3D clip to QuickTime ProRes 4444 in HD 1920x1080 with both "Force sizing to highest quality" and "Force debayer res to highest quality" enabled. (RED graph bar means FASTEST in SECONDS.)
Final Cut Pro X 10.1.1
We imported a 1080p ProRes 422 clip. Here's how long it took to render two different video effects. (RED graph bar means FASTEST in SECONDS.)
After importing the 5K R3D clip used with Resolve, we used the "Share > Master File" function to transcode to 5K Apple ProRes 4444. (RED graph bar means FASTEST in SECONDS.)
AFTER EFFECTS CC
Enabling multiprocessing and budgeting 1.5GB of memory to each subprocess, we rendered the Total Training sample project. We will add the 12-core results to this graph soon. (RED graph bar means FASTEST in SECONDS.)
PHOTOSHOP CC
We timed how long it took to Reduce Noise in a sample RAW image. We will add the 12-core results to this graph soon. (RED graph bar means FASTEST in SECONDS.)
APERTURE
Noise Ninja plug-in was used to remove noise from 50 RAW images. Note the inverse effect of having more cores. (RED graph bar means FASTEST in SECONDS.)
INSIGHTS
The gap in performance between the different models of 2013 Mac Pro varies depending no the application and the function within that application. The short answer to "Which one do I buy?" is always, "It depends on what you run and how you run it."
A good example is to compare DaVinci Resolve's handling of playback and export of a 5K RED clip where the 12-Core Mac Pro shines to Final Cut Pro X's handling of a video effect rendering with a 2K ProRes 422 clip where the 8-Core and 10-Core Mac Pro shine.
We also ran into a phenomenon where the more cores you have, the slower the render occurs. The Aperture graph is case in point.
We included the 10-Core version of the 2013 Mac Pro courtesy of EditBuilder. It beat, matched, or was close behind the 12-core in five out of six tests featured above. My guess is that the 10-core will beat the 12-core in the Photoshop Noise Reduction test. We may not have access to the 12-core until the end of the month of May. If you have access to a 12-core sooner than that, contact us.
Yes, we are working on getting results for the 6-Core with D700s. If you have access to one, let us know. Never hurts to have more than one source.
Comments? Suggestions? Feel free to email me,
Follow me on Twitter @barefeats
WHERE TO BUY A MAC PRO or other APPLE products:
USA readers can help us earn a commission by using this Apple Store USA link or by clicking on any Apple display ad. Live outside the USA? No problem. We are affiliated with these Apple Stores:
and 
