Apple Silicon M1 chip in MacBook Air outperforms high-end 16-inch MacBook Pro

> The M1 chip, which belongs to a MacBook Air with 8GB RAM, features a single-core score of 1687 and a multi-core score of 7433. According to the benchmark, the M1 has a 3.2GHz base frequency.

> The Mac mini with M1 chip that was benchmarked earned a single-core score of 1682 and a multi-core score of 7067.

> Update: There's also a benchmark for the 13-inch MacBook Pro with M1 chip and 16GB RAM that has a single-core score of 1714 and a multi-core score of 6802. Like the MacBook Air , it has a 3.2GHz base frequency.

So single core we have: Air 1687, Mini 1682, Pro 1714

And multi core we have: Air 7433, Mini 7067, Pro 6802

I’m not sure what to make of these scores, but it seems wrong that the Mini and Pro significantly underperform the Air in multi core. I find it hard to imagine this benchmark is going to be representative of actual usage given the way the products are positioned, which makes it hard to know how seriously to take the comparisons to other products too.

> When compared to existing devices, the M1 chip in the MacBook Air outperforms all iOS devices. For comparison's sake, the iPhone 12 Pro earned a single-core score of 1584 and a multi-core score of 3898, while the highest ranked iOS device on Geekbench's charts, the A14 iPad Air, earned a single-core score of 1585 and a multi-core score of 4647.

This seems a bit odd too - the A14 iPad Air outperforms all iPad Pro devices?

It may be possible the variations are due to differences in the thermal environment when the tests were conducted. I would expect the pro and mini to beat the air as they should have better thermals, but that may only show up over longer term tests and environmental factors could win out in shorter tests. Just a theory.

if I recall correctly, the geekbench score does run on small bursts and it's designed to find the peak performance without taking the thermal limitations in account.

Hence why Cinebench is often used these days when evaluating real-world performance with sustained workloads.

Different tests. Different purposes.

GB deliberately avoids running up the heat because it is focused on testing the chip, not the machine's cooling ability.

Cinebench, as you say, tests "real-world" conditions, meaning the entire machine, not just the chip.

The chip's ability to run at sustained load is a part of its design also. Precisely because modern chips has to throttle in order to meet power and thermal envelopes, we should be looking at sustained performance as a more accurate measure.

In a majority of cases, burst performance only affects things like responsiveness, and those things should be measured instead for a better reflection of the benefits.