PC-BSD 10.1 vs. Ubuntu 14.10 vs. Fedora 21 Benchmarks

For your viewing pleasure today are some benchmarks of PC-BSD 10.1 compared to Ubuntu 14.10 and Fedora 21 when testing with both the GCC and LLVM/Clang compilers.

FreeBSD 10.1 was released back in November along with PC-BSD 10.1 and its new TrueOS. It took a bit of time though to get some benchmarks completed of FreeBSD/PC-BSD 10.1 due to running into issues loading the updated OS on a few test systems I frequently use for Linux testing. In particular, disk/SATA issues on multiple systems when booting the PC-BSD 10.1 installer. Fortunately, I came across one of the powerful workhorse systems that played nicely with PC-BSD 10.1

The system that played nicely and was used for today's benchmarking of Fedora, Ubuntu, and PC-BSD was an Intel Xeon E5-2687W v3 (3.10GHz base frequency; 3.5GHz turbo frequency; 10 cores + Hyper Threading), 16GB of DDR4 system memory, NVIDIA GeForce GTX TITAN graphics card, Intel M.2 60GB SSD, and an MSI X99S SLI PLUS motherboard. This X99 MSI motherboard worked great with PC-BSD/FreeBSD 10.1. All of the same hardware and BIOS/UEFI settings were used in testing across the different operating systems.

PC-BSD 10.1 ships with the FreeBSD 10.1-RELEASE-p8 x86_64 kernel, KDE 4.14.2, X.Org Server 1 .12.4, NVIDIA 340.24, the LLVM Clang 3.4.1 compiler by default, and the ZFS file-system. Given that the BSD operating systems are tending to default to LLVM Clang as the system compiler rather than GCC, both Ubuntu and Fedora were tested under the versions of GCC and Clang offered by their respective package archives. Ubuntu 14.10 was tested with GCC 4.9.1 that's the default compiler and then installing LLVM Clang 3.5.0. On Fedora 21, GCC 4.9.2 is the default compiler but also has Clang 3.5 available.

For this basic latest round of PC-BSD vs. Linux benchmarking, various open-source test profiles via the Phoronix Test Suite were run across the platforms. Let's see this initial 10.1 test data. The tests are mostly being done for curiosity sake and in the future may do a larger BSD vs. Linux performance comparison from multiple systems pending sufficient interest.