Since the class I’m teaching ended today (grades are already done!), I dropped by the Apple Store to tinker a bit with the new machines. I had the opportunity to run a few tests on the new machines including running Geekbench (I also ran it on a few others for baseline comparisons). Geekbench is a fairly straightforward (and unfortunately entirely synthetic) test of processing power and memory speed, though it appears to favor raw clock power a bit more than multi-core processing in its calculations. I was pretty pleased to see the numbers on the new machines, listed below with others I tested for comparison purposes. (Several recent discussions of the Geekbench numbers for the new machines are comparing 64-bit test numbers to 32-bit numbers, which is an apples/oranges situation, so I wanted to use the same measurement across the board).
I also created a new 30 sec. iMovie project and exported that at 720p on the new Mac Mini and compared its encode time with the 17″ Macbook Pro (2.3 GHz Core i7) model nearby performing the same task. The 30 sec. iMovie encode:
Mac Mini (2.5 Core i5 w/ATI GPU): ~45 seconds
17″ Mac Pro (2.3 Core i7 ): ~35 seconds
The Geekbench numbers (all 32-bit):
Mac Mini (2.5 GHz Core i5 w/ATI GPU, 4GB RAM): 6538
Macbook Air (1.7 GHz Core i5, 4GB RAM): 5444
Macbook Pro 13″ (2.7 GHz Core i7, 4GB RAM): 6880
Macbook Pro 15″ (2.3 GHz Core i7, 4GB RAM): 10228
Mac Pro (2.8 GHz Quad-Core Xeon, 3GB RAM): 8802
Granted, the Geekbench numbers are very processor/memory focused, so the advantages of the machines with discrete graphics are minimized in such a test, but the new machines stack up pretty well overall nonetheless.
I also had the chance to run some Cinebench 11.5 tests (which use real-world applications, as opposed to Geekbench’s synthetic test) on the Mini and two iMac models:
Mini: OpenGL: 24.84 fps; CPU: 2.53 pts.
iMac 27″ (3.1GHz Core i5, 4GB RAM): 43.06 fps; 4.73 pts
iMac 27″ (2.7GHz Core i5, 4GB RAM): 45.59 fps; 4.27pts.
Oddly, the second iMac scored higher on the OpenGL test, despite having the weaker video card. I’m guessing this was due to the higher TurboBoost speeds of the 2.7 Core i5 processor in that machine. The Mini did quite respectably in this test as well, though still only about half the speed of the iMacs in more intensive operations. I’d love to see how the Mini Server, with its quad-core Core i7 would compare in these tests, especially in the OpenGL test.
The Mini didn’t feel quite as snappy as the 17″ or 13″ Macbook Pros when I was running them through tests, but it wasn’t exactly sluggish, either. As might be expected, the SSD-toting Air felt snappier than any of the others, although its raw power still lags behind its bigger brothers.
After seeing these numbers and getting my hands on these machines, I’m likely to switch my current setup (a 15″ Macbook Pro serving as both mobile machine and home hub) to a Mini/Air combo. That’s still not 100% decided, but I’m pretty close to making the switch. These machines finally have enough horsepower to at least make me think about it.