OS X 10.8 vs. Ubuntu Linux: A Battle With No Clear Winner

Since Apple released OX X 10.8 "Mountain Lion" last month, there have been tests going on at Phoronix of this latest Apple operating system not only on the Retina MacBook Pro, but other Mac hardware as well. In this article is a comparison of OS X 10.8 versus Ubuntu Linux -- when trying out both Ubuntu 12.04 LTS and the latest Ubuntu 12.10 development version.

The first system being used for testing is a mid-2011 Apple Mac Mini that ships with an Intel Core i5 "Sandy Bridge" processor. The Core i5 2415M is a dual-core processor with Hyper Threading and offers Intel HD 3000 Sandy Bridge graphics, 2GB of RAM, and a 500GB Hitachi SATA HDD for storage. OS X 10.7.4 and OS X 10.8.0 were the two Apple operating systems tested on this hardware in their stock configuration. Xcode 4.4.1 was used on this system, which provides LLVM/Clang 3.1 as the default compiler environment. The mid-2011 Sandy Bridge Mac Mini was connected to an Apple Thunderbolt Display. Current Linux distributions work relatively well on this last year's Mac Mini model since it doesn't feature hybrid graphics or anything else to cause pain for Tux.

The second Apple system being used for this cross-OS benchmarking was a late-2010 Apple MacBook Pro. The Apple MacBook Pro has an Intel Core i5 520M "Arrandale" processor that's dual-core with Hyper Threading and running at 2.4GHz. This Apple laptop has 4GB of RAM, an after-market OCZ Agility 2 SSD, and switchable graphics between the Intel HD Arrandale/Ironlake graphics and NVIDIA GeForce GT 330M graphics. Unfortunately, switchable/hybrid graphics remain a pain in the ass under Linux. With Ubuntu 12.04 LTS, the 2010 MacBook Pro had problems with either open-source driver. With Ubuntu 12.10 on this nearly two-year-old Apple laptop, the open-source Intel Linux graphics driver is working with hardware acceleration "out of the box", albeit there still isn't a clean and power efficient way for handling the switch to the NVIDIA GeForce graphics in a seamless manner.

For the MacBook Pro testing, OS X 10.6.8, OS X 10.7.4, and OS X 10.8.0 were tested on the Apple Mac OS X side. For Linux we have Ubuntu 12.04 LTS and Ubuntu 12.10 20120821. Besides going back to OS X 10.6.8 for this older Apple hardware, another difference was that GCC 4.7.1 from the High Performance Computing for Mac OS X SourceForge site was used as the default compiler over Apple's Xcode. Xcode with LLVM/Clang is Apple's default and recommended compiler for OS X but the GCC 4.7 compiler was used on the MacBook Pro for another perspective when comparing to Linux on this different hardware.

Due to these differences, first up being shown are the Apple Mac Mini (mid-2011) results of OS X vs. Ubuntu followed by the Apple MacBook Pro (late-2010) results. The Retina MacBook Pro was left out of cross-OS testing for this article since Linux doesn't yet work well on the new Apple hardware.

For those interested in the graphics performance in particular, coming up next week is an interesting Intel Sandy Bridge graphics comparison that compares OS X 10.8 vs. Microsoft Windows 7 Pro vs. Ubuntu 12.04 vs. Ubuntu 12.10 vs. Ubuntu 12.10 + Git (the latest development code for the Linux kernel, Mesa, libdrm. and xf86-video-intel). These OpenGL results from OS X, Windows, and Linux are quite extensive and show some interesting numbers! All benchmarking was handled from the multi-platform open-source Phoronix Test Suite software with integration on OpenBenchmarking.org.