Platform Differences with Cycles Rendering

I have been “Blenchmarking” multiple computers using the Cycles renderer. I have noticed that most computers I have tried it on (Even tried it on a Raspberry Pi! 80 minute render time), that the renderer will take tiny rectangular ‘chunks’ of a frame and render up to 4 ‘chunks’ at a time (each ‘chunk’ is a core)

I fired up my Mac Pro tower and did the same Blenchmark on the Mac Pro tower and it rendered the ENTIRE frame in cycles all at once instead of breaking it in chunks.

Is there a reason behind the different schemes in rendering?

The Raspberry Pi is ARM, the other PCs I’ve tested on were i5 and i7, and the Mac Pro is a dual socket Xeon (4 cores each CPU).