Possibly, but the code that adds save in one emu may (in speculation) work in another, or similar - say relying on different defined variables from earlier. As the focus changes from addition of features to optimisation, this becomes even more the case, as each coder will have their own optimisation technique - which may not be compatible with anothers (although equally, it could be broadly similar).
Also, even if the code did work together, it would likely not be optimised after the hybrid was put together resulting in something with features but no speed...
... Which is kinda the situation we're in at the moment as it were...
For a start, the speed in Zardoz's emu is likely partly based on the fact that it is at 256 colour depth (or someone said it was - at least that transparencies didn't work), and the lack of features - not the speed hit that turning sound on in Snes9Xgp has - it halves at least on most games. Bear in mind you'll get the same kind of hit in the initial implementation of sound. Of course, Zardoz is also writing from scratch, so far as I know, which makes his emu more likely to run fast anyway, but equally, harder as a project in general.
Intellecto's optimisation, however, is especially impressive (I'm told he's just about re-coded the entire thing in ARM ASM, which is kinda a big achievement). They're both good, but the features wouldn't fit as such. I think.
Of course, I'm no coder - if someone who is and thinks I'm wrong, feel free to correct me - I'll back down instantly
