> I’m not exactly sure how Apple was “behind the state of the art” and has caught up. Can you explain?
The M1 was when they caught up. They were behind the state of the art (of laptop/desktop chip performance) with their A series chips starting with the A4. They gradually caught up over the course of several years to the point that when they released the M1 chip (which was really not that different to the A15) they were slightly ahead.
Yeah, as the other person says, it’s weird to combine A-level chips which were mobile only (minus that short stint where they put it in a Mac Mini before the M1 was ready) with the M-level chips.
Sure, they’re all “Apple Silicon” but the chips should be compared to their equivalents, M-level with Intel/AMD and A-level with Snapdragon/Exonys. We don’t really compare Snapdragon to Core any other time, as they both optimize for vastly different things.
I would also argue that Apple is continuing to make big boosts in their A-level chips compared to their mobile counterparts.
I thought the M-level and A-level chips are pretty much identical CPU cores (just different numbers of them). As such it seems reasonable to me to compare single-threaded performance between A-level and Intel/AMD (making allowances for power envelopes). And indeed one could compare Snapdragon, except they're quite far behind.
> They were behind the state of the art (of laptop/desktop chip performance) with their A series chips starting with the A4
...but A series chips were never intended for laptops or desktops? And by all accounts, the A series nearly always out-competed comparable QC Snapdragon chips.
They weren't... but evidence is that that's only because they didn't have competitive performance for those applications. Once the performance caught up they did choose to put them in laptops and desktops.
I think A chips were quite considerably ahead of low-power Intel chips (e.g. the i7 in MBA) several years before M1 was released. But I guess they didn't believe they could scale it up to compete with higher power i7/i9s shipped in pro macbooks.
I mean that's true. But that doesn't prevent comparing performance of natively compiled ARM binaries in A chips vs natively compiled x64 binaries on x64 chips.
The M1 was when they caught up. They were behind the state of the art (of laptop/desktop chip performance) with their A series chips starting with the A4. They gradually caught up over the course of several years to the point that when they released the M1 chip (which was really not that different to the A15) they were slightly ahead.