Some Basic macOS 10.13 vs. Ubuntu 17.10 OpenGL Gaming Tests
Following last week's F1 2017 launch for Linux which is making use of the Vulkan graphics API on Linux and Metal API on macOS, originally I set out to compare the macOS vs. Linux performance, but that didn't go quite as planned due to MacBook Pro woes. But here are some other OpenGL game tests between macOS and Ubuntu 17.10 Linux.
Originally I planned for an interesting macOS vs. Linux gaming performance comparison, but I was only able to get my hands on a MacBook Pro 2016 (13.3) model and soon realized how even this latest-generation MBP is poor for gaming. When inquiring to Feral Interactive about it, the fears were confirmed and unfortunately the MacBook Pro is a poor candidate for gaming due to severe thermal throttling by Apple and other performance restrictions. In general they have found it very difficult to compare macOS vs. Linux gaming performance as well. With not able to get my hands on any desktop Mac Pro hardware at this time, it was left to run some less demanding gaming benchmarks with the MacBook Pro for the time being.
So I ran some Linux vs. macOS benchmarks with this MacBook Pro for some less demanding games using OpenGL on both platforms including Counter-Strike: Global Offensive, Dota 2, Team Fortress 2, and Xonotic. There's also a j2dbench run for showing the Java OpenGL performance between operating systems.
The MacBook Pro features an Intel Core i7 6700HQ, 16GB DDR4 memory, 250GB SSD, AMD Radeon Pro 450 graphics, and a 2880x1800 display. MacOS 10.13.1 was the latest High Sierra release for testing on the Apple side; all updates were applied, the Radeon Pro 450 usage was enforced rather than the HD Graphics by using gfxCardStatus, and latest Steam for macOS build used.
On the Ubuntu side was Ubuntu 17.10 with the Linux 4.13 kernel and Mesa 17.2.2. An additional run was also used when switching to Linux 4.14 Git and Mesa 17.4-dev via the Padoka PPA for a bleeding-edge Radeon graphics experience. Note that there were no hardware/settings changes made between platforms; the differences shown in the PTS system table just come down to differences in what each OS configuration exposes as their hardware values, e.g. reporting base vs. turbo clock frequencies, etc.
Take these results as you wish, but on the following pages are these surprising macOS vs. Linux gaming performance for this latest MacBook Pro.