The increased noise pertaining to Apple’s plans to transition their traditional compute offerings to ARM-derived solutions continues to gain steam. These rumors are normally followed by insistence that people will abandon their Apple product of choice (MacBooks, iMacs, Mac minis). The general public is very skeptical of the success of such a change. Linus Torvalds has even chimed in on why ARM-based solutions won’t succeed in the datacenter.
The prior anemic performance of Intel’s integrated graphics solution and lack of true innovation due to subtle tweaks to clock speed or minor adjustments on an extremely mature 14nm (insert the total number of +’s of your choice after said measurement) manufacturing process certainly resulted in a product that fell behind Apple’s needs. With the departure of Brian Krzanich, the acquisition of external talent including Jim Keller and Raja Koduri, and proposed changes to how Intel will implement next-generation technologies, Apple’s move to the internally developed A-series SoC’s may be a solution to a problem which won’t exist for much longer.
The current state of the next-generation of Intel graphics looks to be shaping up to provide serious competition to incumbent vendors. The leaked benchmarks for the Intel Iris Plus 940 iGPU solution show the appropriate focus finally exists for the graphics pipeline. If the stand-alone Intel GPU offerings scale this level of performance up to something that comes within a stone’s throw of what AMD and Nvidia are offering in a given price range, both the PC and Mac platforms would offer the kind of generational performance increases that haven’t been present for close to a decade.
Apple introduced the A7 chip as a “desktop class architecture” and has claimed that the current generation A12X chip in the current-generation iPad Pro to provide “console quality” graphics. A desktop variant of the A13 and beyond may indeed be competitive with what Intel offers in 2019, but the investments and strategic roadmap which Intel is executing upon may make this ARM transition an exercise in futility. Customers that have remained with the Apple ecosystem for traditional computer platforms have experienced architectural shifts before for the sake of performance. What we’re likely to see over the next 18-24 months may bring into question Apple’s architectural shift and long-term plans in this space.