Instrumental to the new, ultra-thin design and Tandem OLED displays of the latest iPad Pro models, Apple today unveiled its next-generation system on a chip (SoC): M4.
It’s interesting that Apple decided to advance a full generation name with M3 to M4 with this chip as the bulk of its improvements are not with respect to pure performance but in the Neural Engine and with regard to display support.
Built using second-generation 3-nanometer technology, the M4 boasts 38 billion transistors and can be specced with up to 10 CPU cores and features a 10-core GPU.
This GPU architecture builds upon the advances introduced in M3 and brings dynamic caching, hardware-accelerated ray tracing, and hardware-accelerated mesh shading in graphics to iPad for the first time.
And while Apple notes no performance comparisons between M3 and M4 in CPU or GPU capability, the new M4 does boast Apple’s fastest Neural Engine yet. Capable of producing up to 38 trillion operations per second, the new 16-core Neural Engine brings notable performance improvements for AI-powered features like Live Captions, Visual Look Up, and more.
M4 also boasts the first dedicated Display Engine in an Apple SoC. This Display Engine was necessary to enable the Tandem OLED Ultra Retina XDR display on iPad Pro.
Apart from the capabilities the chip enables on those new iPad Pro models, details on this new chip are relatively scarce. Again, it seems like this is more of a M3 Bionic chip, but perhaps Apple feels strongly enough about the new Display and Neural Engines that it felt they merited the full generation leap from M3 to M4.
I’ll probably never need (nor be willing to pay for) this level or performance – but if it brings the price on M2 and M3 devices down.. I’m all for it!