OS X 10.9 Mavericks Benchmarks On The 2012 Retina MBP

For those curious how the performance is impacted for the 2012 "Ivy Bridge" Retina MacBook Pro when upgrading to OS X 10.9 Mavericks, here's some fresh open-source benchmarks comparing the 10.8.5 and 10.9.0 releases for this Core i7 laptop with Intel HD 4000 and NVIDIA GeForce graphics.

For those interested in the performance of OS X Mavericks, I had delivered some OS X 10.9 benchmarks last month from the (Haswell-based) 2013 MacBook Air. However, this weekend on my main system (the 2012 Retina MacBook Pro connected to an Apple Thunderbolt Display running OS X and then Ubuntu Linux virtuaalized atop via VMware) I finally upgraded from 10.8.5 to 10.9.0. For those Apple/Linux users curious about the performance impact for this older but faster hardware than the MacBook Air, I did some before and after benchmarks.

The tests done from this Core i7 laptop with 8GB of RAM and 250GB SSD with Intel HD 4000 graphics and NVIDIA GeForce GT 650M graphics can be found on OpenBenchmarking.org via 1311183-SO-OSX109MAV08. OS X 10.8.5 vs. OS X 10.9.0 was compared while running XCode 5 and all the other same components throughout.

Before getting to the results, for those that didn't read before why I use this as my main system, at the time of purchasing the Retina MacBook Pro there wasn't a better system available in terms of its display, battery life, and form factor / weight, and excellent build quality. As far as why OS X is used and then Ubuntu virtualized, I've found it to offer better battery life than running Ubuntu bare metal on the device. There's also Thunderbolt display problems with Linux directly at last check. It's also convenient running my main system in a VMware virtual machine for backup purposes.

There's some OpenGL performance improvements to find out of this system when running OS X Mavericks.


But there also appears to be some other OpenGL workloads where the performance is unchanged or has regressed.

With both releases using XCode 5.0, the CPU tests aren't hugely interesting since there's no compiler differences and it doesn't appear there was much performance impact by the kernel changes with Mavericks for these selected workloads.

See the rest at OpenBenchmarking.org. Install the Phoronix Test Suite and run phoronix-test-suite benchmark 1311183-SO-OSX109MAV08 to see how your OS X / Linux system compares to these Core i7 Mac results.