amd and intel both do x64 and x86 but performance varies depending on task. same with all the other ARM CPU's out there. Intel's quicksync will do some tasks a lot faster than a $500 nvidia card. nvidia will do physics better than generic intel
apple has over 1000 CPU engineers working to customize generic ARM CPU's
However, generic tasks such as decompressing zip files, rendering Cinema4D or tracing Photoshop actions are still good metric comparison for clock-to-clock performance of Intel vs AMD chip.
In the same context, unless Apple was somehow able to invent a custom chip that can accelerate web browsing specifically, which I sincerely doubt, comparing browser performance is still a relevant test, if only to test the level of optimization and polish one platform has over another. Hence why we keep seeing Anand doing those tests over and over again.
However, as stated, more GHz means more headroom for optimization. The only question that remains is how long it will take for those optimizations to arrive at a point where it matches iOS current level. At which point, I have a feeling iOS would have pulled ahead again.