Hacker News Reader

Comment by ekr | Hacker News Reader

I'm personally both impressed and surprised by how well this M1 chip performs. But one thing most reviews omit to mention is that it has a fab process advantage compared to the competitors, especially against Intel. And going from TSMC 7nm (which is the latest line of AMD CPUs are using) to 5nm can bring a few tens of percentage points in power consumption and/or performance improvement.

This of course doesn't take anything away from the design, as clearly Apple has had a hefty lead against other smartphone SoC designers in the smartphone space for years.

View on HN

I suspect they do not want to mention it because fab process advantage isn't easy to understand as they're not all the same. For an example, TSMC's 7nm process is more comparable to Intel's 10/14nm process [^ 1]. It's not clear what TSMC's 5nm process is comparable to for Intel's sizes.

As you said, design itself is key also. Look at how well AMD was able to optimize Zen 3 vs. Zen 2 and it is on the same 7nm TSMC process for ~20% IPC gain without increasing power usage too much.

It's going to be interesting to see what Apple can do for the next generation, which I suspect will be on 5nm again.

And Ars did mention it in this specific Ars review:

> Lastly, let's remember that the M1 is built on TSMC's 5nm process—a smaller process than either AMD or Intel is currently using in production. Intel is probably out of the race for now, with its upcoming Rocket Lake desktop CPU expected to run on 14nm—but AMD's 5nm Zen 4 architecture should ship in 2021, and it's at least possible that Team Red will take back some performance crowns when it does.

[1]: https://www.techpowerup.com/272489/intel-14-nm-node-compared...

*At all. Amd didn't increase power draw from 3XXX to 5XXX.

> But one thing most reviews omit to mention is that it has a fab process advantage compared to the competitors, especially against Intel.

This is mentioned in the article.

If you listen to interviews with Apple’s chip team, they talk about optimizing performance of the M1 based on how Mac OS works. An example is how they handle allocation and deal location of objects. Objective-C and Swift using this a lot for loops and for their reference count memory garbage collection. The M1 is 5x faster than Intel for those operations.

that kind of optimization between hardware and software can have a big impact and is part of Apple secret sauce.

Exactly! This is why I am holding for the next generations of CPUs and GPUs, that should be 5nm.

You might be waiting a long time. Hope you have a decent setup already.

One year? AMD moves to 5nm in 2021 according to leaked memos.

What does moving to 5nm mean here? Starting to build stuff with 5 nm or releasing 5nm CPUs AND GPUs? Point is we don't know... and waiting for next gen is always an option -- what about next next gen?

At best It will be Late 2021 and widely available in 2022. AMD normally use High Performance Node and Mainstream Node Pricing. And that is assuming there will be that many capacity left, as far as I can tell, ASML hasn't been keeping up with their shipment numbers. Possibly one reason why rumours are pointing to Apple using 4nm instead of 3nm in 2022.

Reply on news.ycombinator.com