After reading a Blender benchmarking, I decided to benchmark my new Macbook to compare to the other results.

I tested my Macbook and my flatmate’s Dell PC, just to give a comparison. The specs of these was as follows:

  • Dell – Intel P4 2.4Ghz HT, 512MB DDR 400. Windows XP Home SP2.
  • MacBook – Intel Core Duo 2.0Ghz, 512MB DDR2 667. Mac OS X 10.4.6.

For each build/machine two test were run, one with multiple threads off, and a second test with multiple threads turned on. This is a new feature which is part of the new render pipeline 2.42.

First up, the Dell running with Blender 2.42 RC 1" – One thread- 2:12. Multi threads- 2:13. (All times in M:SS format)

Next up I benchmarked the Macbook running the OS X PowerPC version of the Blender 2.42 RC1 release, requiring the use of Rosetta to translate the PowerPC instructions to Intel instructions. Times – One thread – 4:37. Multi threads- 3:22.

Lastly I benchmarked an Intel native version of Blender on the Macbook. This is a rather old build from around March. Some of the UI isn’t working as well as it could currently, with menus requiring you to hit F9 (Expose All Windows) twice to get them to render. Times – One thread- 1:37. Multi threads- 1.15.

As expected, the render times for the Intel Version were much faster than the PowerPC times. The times for the Dell were probably hindered by the fact the Windows renderer is known to be much slower than the Linux and OS X versions. I’ll try the linux version on the PC tomorrow, and Windows and Linux versions (If a live-CD exists yet) on the Macbook.

If anyone can point me in the direction of a newer Intel build, that would be greatly appreciated.