[pavlov]

Laptops with this chip are not shipping yet, so it’s a bit early to judge performance.

[threeseed]

There are some early benchmarks which put its performance behind iPad:

- Snapdragon X Elite (2,574 / 12,562)

- iPad M4 (3,630 / 13,060)

[pavlov]

Since we’re talking about Linux support, the iPad benchmarks seem irrelevant. It’s a completely different kernel.

Asahi Linux benchmarks on an M3 MacBook Air would be the closest thing here.

[schaefer]

According to the asahi website, support is currently available for m1 and m2 macs only[1].

Time will tell about m3 macs.

[1]: https://asahilinux.org/fedora/#device-support

[saagarjha]

This is a CPU benchmark, not a test of what OS is running on it.

[pavlov]

It’s common to have 5-10% differences on benchmarks between Linux and Windows on the same hardware. iPadOS is even more different (for example, no swap, and many other similar fundamental optimizations for single-app usage).

[saagarjha]

It seems very unlikely that Geekbench would swap

[sspiff]

What benchmark is this? Geekbench?

Geekbench is notoriously inconsistent across platforms. Not saying the iPad chip isn't faster, but I wouldn't judge by Geekbench scores alone...

[TiredOfLife]

Do you have a source for Geekbench 6 being inconsistent across platforms?

[lozenge]

You can't benchmark a chip. It all depends on how much power and heat dissipation the actual device can provide. When the MacBook Pro comes out it should run circles round the iPad. And there are no Snapdragon X devices yet

[viraptor]

Sure you can. Measure performance when not thermally throttled and at the lowest speed / no boost. All real devices will fall in between those numbers.

[Xeamek]

What device was this benchmark run on?