Apple's design giving their M1 chip a wider pipeline allowed extremely good industry leading single core performance. I will give them that. 2.5 stacking memory very close to the gpu also enabled them enough bandwidth with pretty decent graphical performance as well. These are all design choices but the 5nm node really enabled some of this to happen. They can make a bigger chip while stuffing more things inside and staying cooler.
The industry is going 3d stacking in 5nm anyways which will lead to some very profound extraordinary performance gains on the x86 side so it's too early to write them off as being disrupted.
While you are correct that those are "design choices", anyone building it on the 5nm process would have done the same or mostly the same (An example would be PS5 using RDNA 1 and developing their own architecture to mimic some RDNA 2 features while XSX uses off-the-shelf AMD RDNA 2 features. Both came up with extremely similar answers and both have extremely similar performance minus the potential infinity cache-like design in the PS5 CPU and GPU cache scrubbers). The design part isn't hard (It's mostly about maximizing what you can and making sure there are no bottlenecks). What's hard is getting access to the 5nm process nodes to allow you to make those kinds of design choices. Right now, TSMC gave Apple first dibs on 5nm. If you have 2 different companies with the same goals in mind, they will produce 2 extremely similar, if not the same, designs if they have to design it on the same process node.
Anyway, Apple is receiving too much praise for things they didn't do.
Last edited: