By Jason Snell
July 24, 2020 11:46 AM PT
Fun With Charts: Entirely speculative charts about Apple Silicon
Note: This story has not been updated since 2020.
Last week I stared directly into the Apple marketing content and speculated about the information it was trying to impart with a WWDC slide. This week, prompted by reader David Hovis, I’d like to engage in some pure speculation on the speed potential of Apple’s new processors, using some back of the envelope calculations and existing GeekBench speed-test scores.
To get started, let’s consider the pace of Apple’s own chip development. Apple’s progress in increasing the base speed of its processors has been very consistent. That makes it fairly easy to estimate the speed of the A14 processor, which presumably will power the new iPhone models being released this fall.
Did you know that the current-model Apple A13 processor generates a faster Geekbench single-core score than the fastest single-core Mac, the 2019 i9 iMac? It’s true. Now, a Geekbench score doesn’t necessarily equate to real-world speed, because every operating system can be more or less efficient. But it’s a pretty decent proxy for us to use on the back of this envelope.
Assuming the normal pace of growth, that theoretical A14 processor will put that 2019 iMac to shame in terms of single-core performance. Now let’s imagine a processor designed for the Mac—I’m going to call it the M14, because why not—that’s based on the A14 but is capable of running at a higher clock speed due to active cooling systems that don’t exist on iOS. (Again, this is a guess—maybe many future Macs will be entirely fanless—but let’s go with it.) A reasonable boost, from a theoretical 3GHz M14 chip to a 3.5GHz M14 would skyrocket single-core performance.
That’s fun to imagine, but now let’s turn to where the rubber meets the road: multi-core performance. Again, Geekbench scores don’t tell the whole story — different operating systems behave differently when parceling out tasks to different cores. Also, Apple’s approach to multi-core processors in iOS devices is very different from the Intel processors used in Macs. Apple’s chips have two separate sets of processor cores, one set designed for high performance and the other for energy efficiency. When you really need to get work done, though, it’ll use all of them.
Still, for my exercise I decided to just focus on the performance cores, because Geekbench scores really do seem to scale directly with the number of performance cores on a processor. Looking at recent trends, I again tried to extrapolate what the two-performance-core A14 processor in the next iPhone would offer, and likewise for a theoretical four-performance-core A14X processor that would power a new iPad Pro and perhaps (in some form) new Macs. Then, looking at current ratios of single-core and multi-core Geekbench scores, I imagined that my imaginary M14 processor had eight performance cores instead of four.
Look, this envelope is getting messy. But just consider the possibilities. Simply extrapolating the growth of the iPad Pro-class processor into the A14X generates a multi-core score that’s faster than the fastest MacBook Pro Apple currently makes.
Now for giggles, toss in that eight-core “M14” processor. Now you’ve got a Mac that’s basically faster than any Intel Mac other than the very fastest Mac Pro and iMac Pro configurations.
Will it happen? Who knows what Apple’s roll-out strategy will be. But if Apple wanted to use its iPad Pro-class processor in some lower-end Macs, and a souped-up processor in some higher-end Macs, it’s not unreasonable to imagine that it could blow away the performance of almost every Intel-based Mac.
To repeat: These numbers are entirely made up. But going through the exercise makes me pretty confident that Apple will be able to deliver major upgrades to Mac performance.
[Thanks again to reader David Hovis for the suggestion.]
