hyfsade limits...
Limits
Due to the nature of emulation, PearPC is quite slow (the client will run about 500 times slower than the host). Note that only the CPU is that slow, the speed of the emulated hardware doesn't suffer that much from the emulation; e.g. the speed of simulated harddrive/CDROM is quite good, especially when using the busmaster interface. This situation is better if you can use the JITC (about 40 times slower) but still not ready for productive use.
Because the author has only access to little endian machine, PearPC will most likely only run on little endian architectures due to some stupid assumptions in the code. This shouldn't be hard to fix, the author would fix this himself if he had big endian hardware (Hint: You should donate big endian hardware to the author if you want this to be fixed!)
Because of some equally stupid reasons PearPC will only run on 32-bit architectures. Yes, this also shouldn't be hard to fix.
A lot of unimplementated features are currently critical (i.e. will abort PearPC).
Timings are very unaccurate. This will be fixed in later versions.
No idle sleep (PearPC will consume cpu resources although the client is idle).
PearPC lacks a save/restore emulator-state feature.
No Altivec support but planned.