It's about the battery power and reduced heat that you get from having more cores with a slower clock. It's not about the actual speed the processors. Iphones are at the point right now that any increase doesnt mean much for the average app anyways.
Xeon's are mult-core beats that have low power requirements which also reduces its heat output; they have always been designed this way for a long time now. This is why the Mac Pro desktops use xeons. They are superior chips over regular desktop chips, they just dont do well in gaming, but that's the fault of the game designers not the chip designers since games don't use more than 8 cores right now, but hopefully direct x12 will help to fix this problem and eventually we'll start to see games being designed for 20+ cores.
The biggest advantages to using this design in cell phones is to reduce the power requirements to boost battery time, as well as reduce the amount of heat the phone uses which makes it alot more easier to carry around and use for longer periods of times.
If people still dont understand why i added that to the list then i wont be able to help them and they are more likely gamer's that chase the clock speed because game designers are to lazy to add true multicore support which is far superior in every way anyways. Clock speed has already reached its maximum roof already, its why multicores have taken over and why we dont see 10ghz processors since increasing the clock speed increases heat drastically which also increases power requirements. This is definitely something you don't want in the cell phone category's.
Honestly, power gating on new Intel processors (starting with Sandy Bridge) is very good. Underlying architecture in x86-64 desktop and server processors are the same really. The Mac Pro and other workstations use Xeons I suspect for ECC support rather than power considerations. Also, the workload on workstations are likely better suited for multithreading (e.g. CAD, 3D rendering, video editing, etc) than typical home use.
Not everything can be multithreaded. Iirc, even x264 (H.264/AVC video encoder) doesn't scale well beyond 32 cores. There's also overhead involved with managing a lot of parallel threads.
By the way, all things being equal (architecture, clock rate, #cores) Xeons do just as well as desktop i5s and i7s in gaming. The extra horsepower on 6- or 8-core Xeons is usually just unnecessary as the bottleneck is typically the graphics card (even in SLI/XFire multi-GPU configurations). Now if you compare, say, 6-core Nehalem/Westmere Xeon to 4-core Sandy Bridge/Ivy Bridge/Haswell for gaming, then of course the latter will be better. The newer architecture simply has better IPC (and is more power efficient to boot).
Typical core configuration I see on mobile devices is having some slow cores (low power usage for undemanding tasks) and some fast cores (high performance for demanding tasks). The device switches between the different sets of cores depending on the workload. It's not like all the cores are running at the same time. That said, this isn't the only method available for lowering power consumption. Granted, I do believe the new A10 Fusion uses this configuration (2 fast+2 slow cores).
All things considered, I think Apple does pretty well on the power management front considering the battery capacity on their iPhones.
iPhone 6s: 1715 mAh
Samsung Galaxy S6: 2550 mAh
Samsung Galaxy S7: 3000 mAh