It would take an incredible amount of work to port. I don't have much emulation programming experience, but I'd venture a guess that it would probably be more practical to just rewrite it from scratch for the target console, rather than try to port all the existing code. As far as power goes, you're right for the most part, maximius, but there are some optimizations that could be made with emulation on certain consoles that would not be possible on PCs.
For example, the PS2's CPU uses the MIPS architecture, as does the N64. This has the potential to make code translation much quicker. In English, both processors speak a similar language, so not as much work is required to convert N64 code to PS2 code. Another advantage that consoles have over PCs is that their hardware is designed for high graphical performance. To that end, most consoles have much faster RAM-VRAM communication than PCs do. This could help with things that PCs really have trouble with, such as frame-buffer effects.
On the other end of the spectrum, consoles are typically not as well documented as PCs, so writing an emulator for a console requires not only emulating a poorly documented console, but also writing software for another poorly documented console. This makes development much slower and more difficult to debug. Testing out code on a console is also a lot more time-consuming than on a PC, and some special equipment is often required (modchips, special cables). I'd really love to see an N64 emulator for the PS2, but given the (unpaid) work involved, I doubt we'll see one any time soon.