It reminds me of the joke that "I can do math very fast", probed with a multiplication and immediately answering some total bollocks answer. - "That's not even close" - "Yeah, but it was fast"

Sure, it's not a trivial problem, but why wouldn't we want better compilation results/developer ergonomics at the price of more compiler complexity and some minimal performance penalty?

And it's not like the performance doesn't have its own set of negatives, like header-only libraries are a hack directly manifested from this compilation model.

your topic is shifting from multi-threading benefits in language compilation, to ergonomics of C headers.

No it doesn't, try it against a language with modules support, even the oldie Turbo Pascal for MS-DOS.

Turbo pascal compiles faster because of better parallelization?