Using the second CPU is very hard because it has to access memory through the first CPU - if the first CPU is busy doing this, it can't do work itself so you aren't any faster.
It may be possible to give small (in memory) but complicated (in time) tasks to the second CPU and execute them from the cache, but this will be doubly difficult to utilise in an emulator - rather than in a custom application where such tasks might be easier to define. I'd be surprised if there was more than a 10% speedup by implementing the second CPU in an emu.
The situation may improve when someone rewrites the OS completely, but it's not reasonable to expect that to happen anytime soon.
EDIT: And oh, please try to read into things more before you post a question. This has been answered (better no doubt) many times before.