Apple’s chips division leaves lead developer to Intel to work on socs – Computer – News

Imagine Intel making a beautiful, high-tech SoC just as fast, efficient, or even better than Apple. Then Intel has only a third of the puzzle. The second third is a fully horizontally integrated operating system with a SoC. This is necessary for things to run efficiently, and Intel doesn’t have much experience with operating systems. The third is to urge developers to develop application software for this platform.
Make no mistake, the latter two are at least as important as success. Horizontal integration is the main reason why iPhones perform much better than Android phones. It’s not because Apple can do magic, it’s because Apple has HW and OS in-house and can improve it to the max. But even if, and that’s a big deal, others produce SOC as fast as Apple, they still need MS or Google to produce an OS. This would be generic, and therefore less optimized, i.e. slower. The M Series and A Series SOC were designed and built specifically for the Apple operating system, not a generic operating system.
If you look at what Apple has been doing in recent years, you’ll find that it’s something that Intel will have to work hard to catch up with:
– Build your own SOC, with your GPU architecture, your NPU, hardware video encoding and decoding, and shared RAM on the SOC (so that, for example, data doesn’t have to travel to and from the CPU/Processing graphics, but both are just the same data directly) and many more clever things.
Metal support in GPU only, no support for other APIs like OpenGL or Vulcan thus allowing for further optimization, no need to modify hardware to fit other APIs.
– Apple abandoned all 32-bit support a few years ago, everything is 64-bit, old software doesn’t work anymore. Supporting a legacy is like raising faster horses rather than building cars. Turn it off and you can get rid of the stables, feeds, etc. Or with a computer, it is less complex, less square mm.
Their different operating systems have converged over the years, and now it’s at the point where iOS runs on macOS (from an architectural point of view)
– OS and HW are optimized for each other, support very few devices and when support for Intel CPU and AMD GPU only ends in 5-6 years. Even a smaller code size, more optimization can be achieved
Rosetta II is 64-bit and translates all x86 64Bit programs to 64-bit ARM, allowing programs with legacy x86 code to run very quickly on their M-series
– Since they have a single SOC architecture for smartphones, tablets, watches, Set Top, etc., the overhead for developing cores in them and more functions is spread over an incredible number of SOCs. This increases profit margins compared to the competition and allows Apple to pre-purchase massive manufacturing capabilities from TSMC in its latest manufacturing operations, and possibly even fund TSMC to provide more competitive advantages.
– Most importantly, many software developers have been brought to the new platform, MS, Adobe, Affinity, Black Magic, to name a few.

Reply