Would it be worth replacing the disk accesses in an x86 emulator (file I/O) with asynchronous accesses(place a request for the handling thread for a virtual sector, continue emulation until it's needed, then when it's needed wait for the handling thread to finish processing said data, removing any I/O access times in favour of emulation time)? Or would modern operating systems make this a waste of code(doing that already, reducing read/write times to near-zero, not slowing down CPU emulation)?