It is very much dependent on the underlying architecture, but any difference will be minuscule.
If anything, I’d expect (x >= 0) to be slightly faster, as comparison with 0 comes for free on some instruction sets (such as ARM).
Of course, any sensible compiler will choose the best implementation regardless of which variant is in your source.