Following qualcomm's announcement of their third gen Arm-laptop-SoCs, they also revealed Nuvia-designed cores will power their future laptop SoCs.

To my understanding, both 2022 and 2023 flagships will be using TSMC N4 node or a variation of it.

Now my question is, what's the real magic behind apple's outstanding Perf/W?

Is it the wide design? Then why didn't it work for Samsung's own Monngoose microarchitecture?

If it's simply their advantageous process node, then why current 5nm mobile chips still trail behind apple's 7nm ones in term of effeciency?

They both play a role, if it wasn't obvious for some reason, but which is the major factor? I remeber anadtech saying process node is the main factor while design plays a minor role.

But again, Intel's been doing pretty darn good with their relatively primitive Nodes.

Same goes for gpus, qualcoom and imagination tech both claim extremely wide shaders improve effeciency, so, why don't nvidia and amd do that? Why AMD regressed from its 4x16 wide to a 32 wide? Why is intel is doing 8wide/256bit shaders?