While your story about raw computing performance holds, there are two reasons why apple's M1 release is significant (i.e.: not marginal):
1. This is the first time in decades that we see a serious non-x86 competitor for general purpose consumer-level computing. This is significant in and of itself.
2. Power consumption. These things seem to consume 2/3rds as many watts to offer the same performance levels than their x86 counterparts. This is big in the laptop space.
I feel like this would be a lot more significant if you could buy the chip on its own. When you can only buy it as part of a Mac, you're really not buying the chip, you're buying a Mac with all that entails. Then who makes the chip in your Mac doesn't really change much because you buy it from Apple either way.
> Power consumption. These things seem to consume 2/3rds as many watts to offer the same performance levels than their x86 counterparts. This is big in the laptop space.
But that's much the same issue where you're comparing it to Zen 2 when Zen 3 is expected to be 20% faster at the same power, and close a lot of that gap.
It's also hard to compare this in general, for two reasons. The first is that the M1 is 5nm, so then the question is how much of the efficiency is due to that (and will accrue to AMD next year) and how much, if any, is the design?
And the second is that the M1 is big.LITTLE. The small cores are more power efficient, so when you average them in, the overall result should be more power efficient than entirely big cores. But x86-64 processors with a similar configuration are sparse, so what do you compare to? It will be interesting to see the result when there are some (both Intel and AMD have some planned).
Oh, totally. I'm not buying into apple (again) anytime soon, so the only bad thing about M1's for me is that I just won't have them, no matter how good they are.
> But that's much the same issue where you're comparing it to Zen 2 when Zen 3 is expected to be 20% faster at the same power, and close a lot of that gap.
I expect next-gen AMD laptop processors to close the gap either in performance or in power consumption, but not both. Also, power consumption of a laptop is very much a whole-package deal, and Apple really shines here.
To be clear, I dislike apple for who they are and where are they leading the personal computing industry, so I won't buy from them. Nonetheless, I have to acknowledge that they make some damn fine machines.
https://www.youtube.com/watch?v=UxSI45eeAts
Not to mention the ergonomic factor of being cooler and quieter.
This review appears to be comparing a dual core Intel i3 to the 4+4 core M1, which are only the same price because Apple doesn't sell any laptops for less than $999 even when they're a dual core Intel i3.
Is there somewhere I can find some evidence that it isn't a marginal improvement?
Apple timed the release perfectly to create the impression that it was more, because its obvious competition will be Zen 3 laptops, which are right around the corner but not actually out yet. So then you get benchmarks comparing it to 14nm Intel processors based on 5-year-old technology, and 10nm Intel processors with half as many cores as the M1 because Intel's 10nm yields are so bad they can't make high core count processors, and Zen 2 laptops when the Zen 3 ones it will actually be competing with in a couple months are expected to be 20% faster. And Zen 3 desktops, which the M1 loses to badly on multi-threaded tests, but is competitive with on single threaded tests -- as expected, since modern mobile processors actually have single-thread performance which is competitive with desktops in general, but manages to amaze anyone who doesn't know that.
Which isn't to say they've done a bad job. Making something which is even competitive with modern x86-64 processors is no small feat. But the fanboys are making it out to be some kind of coup when it does, in practice, appear to be a marginal improvement. And we still don't even have many real-world benchmarks because most of the real-world applications haven't actually been ported yet.