Anyway it appears that rendering HAM8 (via C2P) is not really any slower than normal 8 bit C2P, the real overheard is generating the color values to be used (which is very slow to do dynamically).
And most demos used a 160x100 HAM-mode instead of 320x200, as a 320x200 HAM8 c2p-mode would require a superhires mode and 4x the amount of bytes written to chipmem, making it pretty useless.