The controversy over the fact that Apple has changed Intel for its own processors under ARM architecture is far from over. Strategy aside, the truth is that the challenge of those of Cupertino is gigantic in terms of the software section, since X86 and its processors have little to do with ARM and theirs, why is it so difficult to change architecture and how important is the software in this?
Apple is going to play much stronger than in that 2006 when it finally left IBM to go into the arms of Intel. He jumped from an established and muscular company to go with the giant to an architecture where the entire sector works. Instead, it is now betting on ARM in an environment that is far from similar to that of that time and doubts arise among its users.
What aspect consumes more time in this change of Apple and Intel?
Without a doubt, the design and R&D that the A13 processor will include will be worthy of study, since it will be implemented in 5 nm with 12 cores and a series of improvements of which there are only rumors. But this is something where Apple can decide at all times both the development and optimization times of its architecture. On the other hand, everything related to their “bubble”, whether hardware or software, is causing them serious problems.
Therefore, the aspect where Apple is going to miss Intel the most is not in the hardware as such, but in the software section. Without wanting to delve too deeply into technical concepts, there are certain parts of the optimization process that Apple will not only have to carry in its operating system, but anyone who wants to offer support in its program will have to do it.
First, a simple installer of such programs should be able to choose between different compilations depending on the architecture or choose the correct one automatically. Also, you have to carry a huge amount of code to end up correcting a lot of it due to compilation errors and its tools.
How to force millions of developers to port their code for your ARM hardware?
Apple will manage to port its code for its operating systems and we understand that manufacturers will have to do the same for their drivers. But the challenge is not that, the challenge is in the community and professionals behind the Apple ecosystem.
Will they emulate most of the software? Will it have to be rewritten in part and wasted time and resources on it? And finally, will these companies be worth doing to stay in the apple business?
The type of language used, compiler and an endless number of variables can make carrying a program as complicated as it is inefficient, where however much power it has, it achieves very poor performance. Microsoft and Intel have been working on a kind of translator for a long time to run X86 programs in ARM without having to rewrite them, maybe Apple is working on something similar, but we don’t know if it will arrive in time and if it will work as well as expected.
Developers have to carry millions of apps to adapt to new devices, and that costs time and money, something that could shrink the Apple ecosystem in large numbers. As we see, the doubts continue and will continue until we see the first device with ARM to work and check how mature everything is.
What seems clear is that the jump between X86 and ARM at Apple will be slow and apparently irreversible, but is it really the right option? What value does the brand provide now for your hardware? Perhaps the custom design and implementation, something that would not be justified if performance problems, lack of software or poor optimization of it begin.
