Native performance doesn't mean anything aside from the same program/OS running on the same system. In other words, "native performance" is the baseline for a relative speed comparison, not some absolute value that makes sense in itself.
Native means you have access to the hardware through the kernel of your choice, it's not about performance.
If there is a hypervisor in between you and your CPU - let alone emulation - you lose direct access to things like performance counters (which I need to be able to do a lot of things I enjoy programming). On top of that you also now have to trust (this isn't even security, the implementation can be buggy) a hypervisor - they effectively have to mimic the behaviour of a CPU, they aren't perfect.
>Native means you have access to the hardware through the kernel of your choice, it's not about performance.
Yeah, I know what "native" means.
I talked about the use of "native performance". That when used for comparison, it doesn't mean anything, unless it compares between the program running on the native OS and virtualized on the same machine.
So, saying
"Near native performance" i.e. not native"
(a) doesn't say anything as a general statement (of course it's less than 100% of native, as its virtualized), and
(b) is moot as an argument against the M1 (since this "non native" performance could still could be e.g. 100% or even 120% of the native performance of another machine, which would make it more than enough).
Native performance doesn't mean anything aside from the same program/OS running on the same system. In other words, "native performance" is the baseline for a relative speed comparison, not some absolute value that makes sense in itself.
Thus, "Near native" on the M1 could be faster than native on an Intel machine. Or on an ARM, for that matter: https://www.macworld.co.uk/news/apple-m1-runs-windows-faster....