For this testing, we are comparing the performance of Mac OS X 10.5.8 to that of Mac OS X 10.6.0. For our Snow Leopard testing we were using its GM build and in fact a retail copy that we had not received until this morning. Fortunately, we have been working on this article, the Leopard testing, and OS X improvements to our testing software all week to bring you these detailed benchmarks in a timely manner. Later on we will be publishing more benchmarks, including ones that directly compare the performance of Apple's Snow Leopard to the latest Ubuntu Linux development snapshot for the Karmic Koala (9.10). As we mentioned earlier this month, in September we will also be performing a big benchmarking comparison using Mac OS X 10.6, Ubuntu Linux, potentially some other Linux distributions too, OpenSolaris, FreeBSD, and others. Stay tuned for this massive performance comparison.
Our open-source testing software, the Phoronix Test Suite, has supported benchmarking under Mac OS X since last year and we have ran other articles in the past like Ubuntu 9.04 vs. Mac OS X 10.5.6 benchmarks. However, with the new Phoronix Test Suite 2.2 Alpha 1 release, many more test profiles have been made compatible with Mac OS X and other improvements made for the Apple support. The Phoronix Test Suite ships with over 100 test profiles for Linux and other operating systems, while there are over two dozen profiles also compatible with Mac OS X 10.5/10.6. These test profiles range from running different games to monitoring the CPU usage during video playback to measuring the time needed to encrypt different files. We firmly believe this is the most comprehensive way of benchmarking Mac OS X on the desktop and servers, for at least what is publicly available. For those unfamiliar with this extensible testing software that is used by us, other publications, and many hardware/software companies, stop by Phoronix-Test-Suite.com.
Due to a lack of Apple hardware in our testing labs, for this initial Snow Leopard benchmarking we just used an older Intel-powered Mac Mini and then a newer Mac Mini that contained the NVIDIA GeForce 9400M graphics. While in the older Mac Mini setup (hereinafter in the results referred to as a "MM1"), the processor may not be as fast as those in the newest Mac computers and there are just Intel integrated graphics, these numbers should prove to be equally interesting to see how well Snow Leopard does really improve the performance on lower-end Apple hardware. MM1 is just using the 32-bit kernel due to its 32-bit EFI. With the newer Mac Mini that has the NVIDIA graphics (hereinafter referred to as "MM2"), it has a 64-bit EFI and with its GeForce 9400M graphics processor does support OpenCL. Though if you like this testing and would like to see the tests conducted on a greater selection of Apple hardware, you can help us by joining Phoronix Premium, making a donation, or using an affiliate link.