[vnorilo]

The more I've written code close to metal (mostly SIMD for signal processing), the more I've grown to prefer either intrinsics or separate translation unit for assembly.

If you want your code to intertwine with what the C compiler does, intrinsics are great.

If you don't, .s is great.

[immibis]

And the only way to use new instructions is to make them slow or to wait for a compiler update.

[pjmlp]

Exactly, this is the right approach.