My Hacintosh Experiment

January 1st, 2011 by The Power Searcher

Previous: Software Installation...

Performance Benchmarking

On paper, each component selected for the "MicroMini Hackintosh" is faster than its Mac Mini's counterpart, so it seems obvious that it should completely outperform the Mac Mini: Time to quantify this "obvious" conclusion.

For performance testing the following software tools were used:
  • Activity Monitor - Part of Mac OS X, used to monitor processes and CPU use.
  • Temperature Monitor - Freeware version of a great tool for monitoring, displaying, and exporting data from temperature sensors on your Mac. From Marcel Bresink.
  • Chess - Included with Mac OS X, it's a convenient way to automatically test a CPU.
  • Xbench - Free, comprehensive utililty for benchmarking overall Mac performance. Version 1.3 was used for the test. 
  • Geekbench - Software utililty for benchmarking a computer's CPU & Memory performance. Version 2.1.11 was used for the test. 
  • MaxCPU - Two programs designed to find the limits of a Mac's components using software included with OS X, along with a few Terminal commands. Kuods to Chris Hynes for these free utilities.

Initial Benchmarking: Xbench

Xbench is a comprehensive, well-supported, widely used, and free performance benchmarking utility for Macs designed to test a computer's major components. Prior to running the test, the Xbench database was reviewed for recent test reports on 4th-generation Mac Mini computers running OS X 10.6 (aka: Macmini4,1 with 10H574 or 10D2235): The first five reports found and not listed as "partial results" were used for the comparisons.

Configuration:
Vognsens
 DB Mac Mini
 Pepcid Louis Allott
 Tom's Mac
System Version:
10.6.5 (10H574) 10.6.3 (10D2235) 10.6.5 (10H574) 10.6.3 (10D2235) 10.6.5 (10H574)
Physical RAM:
4096 8192 8192 2048 8192
Model:
Macmini4,1 Macmini4,1 Macmini4,1 Macmini4,1 Macmini4,1
Drive Type:
Patriot SSD INTEL SSD OCZ SSD Hitachi TOSHIBA
Xbench Result:
 140.38
 196.49 188.94 134.52 118.84

Of the configurations listed, Louis Allott's Mac is closest to the MicroMini: All of the others have more memory and/or use solid-state hard drives (SSD). Xbench identified the MicroMini as a MacPro3,1, corresponding to a 2008 Quad Core Mac Pro: I suspect a result of using the tonymacx86 install process. Regardless, Xbench reported an Overall Result 184.75 for the first run, and an average of 181.33 over six (6) subsequent consecutive runs.

Bottom Line:
Relative to the Mid-2010 Mac Mini (Macmini4,1), the MicroMini achieved an average score 46% higher than two stock computers and -4% versus two using a Solid-State Drive (SSD). (Excluding the "Vognsens" results, see Note below).

MicroMini Hackintosh: Xbench Comparison Details, First Run
(Click an image for larger version)

Overall, the MicroMini exceeds those using the stock 5400RPM 2.5" drive and is similar to Mac Minis using solid state drives (SSDs).
The MicroMini easily beats all but the  "Tom's Mac" configuration, which might be a Mac Mini Server running a 2.66MHz processor.
The MicroMini's Core 2 Quad Q9300 CPU easily surpasses the Mini's Core 2 Duo P8600. The MicroMini's DDR3 PC-10600 memory is faster than the Mini's DDR3 PC-8600 memory.
The MicroMini has Intel GMA X4500 integrated graphics onboard, but the dedicated 9400GT Video Card is faster than the Mini's nVidia GeForce 230M integrated graphics.
 Not sure why the MicroMini is only comparable to the stock Mac Minis in this test. Note the two higher bars are Macs with Solid State Drives (SSDs) installed.
The MicroMini definitely "feels" faster on User Interface tasks than the Mac Mini.
The MicroMini's 7200 RPM, 3.5" Seagate Barracuda drive outperforms the Mini's stock 5400 RPM, 2.5" Hitachi Travelstar drive, but pales next to the two Macs with Solid State Drives.
Note: The Disc Test results suggest that the "Vognsens" Mac Mini uses a solid-state drive (SSD) that is either configured incorrectly or possibly defective. Consequently, the Overall Result for this computer was not used in calculating performance differences.

Geekbench

Geekbench is a cross-platform performance benchmarking utility from Primate Labs that tests a computer's CPU and memory. The Macintosh version comes in 32-bit, 64-bit, and Rosetta versions: The 32-bit version is free, and the others available for a small fee.  As the MicroMini only runs in 32-bit mode, the freeware version was sufficient, and provided this score:
MicroMini Hackintosh: Geekbench 2 Summary
Compared to Geekbench's reported benchmark for Apple's Mid-2010 Mac Mini, the MicroMini scored 57.5% higher. The improvement is greater than that reported by Xbench, but is expected: Unlike Xbench, which tests a computer's CPU, memory, video, and hard drive, Geekbench only performs CPU and Memory tests.

Overall, I'm very pleased with the "out of the box" performance of the MicroMini Hackintosh: It's much faster than the Mac Mini and appears to hold it's own not only against the Mid-2010 Mac Mini Server but also against entry-level Macs using solid-state hard drives (SSDs), while having far more expansion options. From a price standpoint, the $378.16 total price of the MicroMini represents a 48% savings over the $699 entry-level Mac Mini: Almost enough to build another complete MicroMini Hackintosh.

It appears the well-worn claim made by PC supporters during the 1990's that "PCs are much faster than Macs at half the price!" still holds true today...at least for this exercise.

Next: Performance Tweaks