Thoughts on Apple M1

the Apple Silicon M1 chip all these touted statements of magical performance

Remember kids: no gameplay, no hype.

Yes, I’m confident that Apple can deliver. They’re good at their silicon for iPhone and iPad, and I’m impressed that their iPad Pro beats my own MacBook Pro in raw performance. (That said, it’s not an even comparison!)

But this is all coming from Apple. I’ll believe it, when the hardware reviewers actually run benchmarks and do real world tests comparing this and the laptops that they claim to beat. If folks like Linus or the AnandTech folks are convinced, I’m game.

I’m also worried on the software side. Processor architecture differences are a huge thing, and not all software is magically optimized to make this work. Maybe Apple’s software are enabled, but there’s a lot of tools I use that are third-party game changers that would become deal breakers if they don’t work on this.

There are also some offbeat but important use cases in my mind that I need a proper confirmation with their new chips and Big Sur. How well does it play with virtualization use cases? How well will it perform while I watch HD videos while I work on my IDE, my series of backgrounded but active Firefox / Electron sessions while it’s spinning up a series of containers? Will it overheat or actually run all of these without issues? That sort of stuff.

Finally, I’m not yet convinced about single-core performance and heavy workloads on Apple’s silicon. See the benchmarking remark up top, but this time, it’s just how Arm tends to do these kinds of stuff. It makes sense since they’re balancing power use vs performance, but still, the performance ceiling that x86–64 can reach is hard to realize on Arm.

I’ll wait for Gen 2.

comments powered by Disqus