I have been working on these puzzles on my personal machine which is a high spec 16" 2020 Intel MacBook Pro. I have it dual booted between macOS and Windows so I can switch to whichever suits my current task. Usually, Windows for coding and gaming and macOS for playing with music creation.

Now that dotnet is cross platform friendly, I could run the AoC project in Visual Studio in Windows, or Rider on macOS.

Running the project in either environment produced no discernible difference in performance, so it seems the choice of OS doesn't make much difference to dotnet's use of the bare metal.

So, the Intel MacBook specs that matter are:

Processor:          Intel Core i9-9980HK
Base Speed:         2.40GHz
Cores:              8
Logical Processors: 16
RAM:                32GB

I recently got a work laptop refresh to a 13" M1 MacBook Pro with the following specs:

Processor:          Apple M1
Base Speed:         Nobody Knows
Cores:              8 (4 performance, 4 efficiency)
RAM:                16GB

I thought it would be interesting, given Apple's bold claims, to see how my code performed on this machine.

Wow.

Execution times for all 343 (7 years × 49, since Christmas day doesn't have a part 2) puzzles:

Intel MacBook Pro: 131,634ms
M1 MacBook Pro:    113,161ms

That's a 14% reduction in execution time on a MacBook that was probably around half the cost of my personal one. Seems Apple may have some substance behind their claims.