In other words, Z380 is about as fast (in practice, due to R800 waitstates a bit faster) as an R800 while executing 'legacy' Z80 commands.
The power of the Z380 really kicks in when you use the extended instruction set and you realize the huge amount of memory you can access. (R800 is slow when using external memory, and even an expanded turboR has a 1MB limit)
In emulation it's hard (practically impossible) to make efficient use of the extended instruction set, so Z380 will be too slow to emulate R800 at the right speed.
so if u r programmer thats means u can also built an z380 instruction in UZIX OS and MSXDOS 2.20 for the speedloading? like r800 of the turbo-r with msxview/dos 2.30?
i dont agree :
99% of instructions of the r800 are z80 legacy instructions
the remaining 1% that would be trapped in case of z380 emulation are MULtiplication which are long anyway
on r800 (9 cycles + ) .
on top of that r800 is used 99.99% of the time to execute z80 only programs.
so the emulation makes sense since it would have very limited slow down and enable total
compatibility at binary level.