Mac OS X 10.5 vs. Ubuntu 8.10 Benchmarks

With this being our first time comparing the performance of Mac OS X to Ubuntu Linux, we were not sure going into this which one would emerge as the leader. After looking at these results from the Phoronix Test Suite, it's almost a toss-up as to which operating system is faster.

In the Nexuiz test, Ubuntu 8.10 was nearly 50% slower than Mac OS X 10.5.5 at the two resolutions tested. In the Urban Terror game, which uses the open-source Quake 3 engine, the gap between Mac OS X and Ubuntu had widened even larger. The Mac Mini with its integrated Intel 945 graphics had performed nearly four times faster under the Apple operating system. Linux losing out in this area can be attributed to the open-source Mesa stack not being too well optimized at this time for performance and the Intel Linux graphics work currently going through some radical changes.

The developers in Intel's Open-Source Technology Center have been busy working on the Graphics Execution Manager (GEM is the Intel-developed kernel memory manager after they had a falling out with TTM) and merging it into the Linux 2.6.28 kernel. They have also been busy with other invasive work such as kernel mode-setting. Once this work has been stabilized and regressions addressed, we should see better performance and a more end-user friendly experience. The G43 version of Intel's GMA X4500 isn't even working right now on Linux in some cases as we explained in this review and other users have experienced problems too. We will also see faster OpenGL performance on Linux once the open-source drivers begin using Tungsten's Gallium3D.

Moving past the graphics benchmarks, the results from the audio and video encoding benchmarks were much closer between Mac OS X and Ubuntu Linux. With MP3 encoding using LAME, Ubuntu 8.10 x86 was less than 4% slower than Mac OS X 10.5.5, but the x86_64 spin of Ubuntu 8.10 ended up being about 10% faster than Mac OS X. The x86 spin of Ubuntu was also slower at encoding an Ogg file, but the x86_64 edition ended up being noticeably faster than Mac OS X. For video encoding with FFmpeg, Ubuntu was faster in both instances.

Our compilation tests consisted of measuring the time to compile PHP 5.2.5 and ImageMagick 6.4.0. With both programs, Ubuntu was significantly faster than Mac OS X. This in part may be explained by the GCC differences with Apple's Xcode providing GCC 4.0 where as Ubuntu Intrepid is using the newest GCC 4.3 series.

Beyond the compilation tests we had ran 7-Zip and Gzip for file compression tests. The Intrepid Ibex had outscored Leopard in the Gzip compression, but Intrepid x86_64 only had a small lead in 7-Zip compression and the x86 edition was left at a loss.

Mac OS X was left on top again when it came to the Bonnie++ disk benchmarking. In both the sequential create and random read tasks, Mac OS X was able to sustain significantly more operations per second with the Serial ATA hard drive than did Ubuntu. Ubuntu was trailing Mac OS X also in the GnuPG file encryption test. The disk performance differences may be explained in part by the Journaled HFS+ file-system being used on Mac OS X compared to EXT3. Soon though Mac OS X will have EXT4 and the next-generation btrfs file-system as competitors.

While BYTE Unix Benchmark is a synthetic test, the Dhrystone 2 performance was about 25% faster in Ubuntu x86 over Mac OS X and almost twice as fast with Ubuntu x86_64.