joels@tekred.TEK.COM (Joel Swank) (03/09/87)
I just did a comparison of 6502 vs. Amiga timing for a real application. I ported a 6502 assembler written in C from VAX to the Amiga. I compiled it under MANX with 16 bit ints. I did a timing comparison for assembling the same 2000+ line program on the Amiga and on my Rockwell AIM with a 1Mhz 6502. The AIM assembler is written in assembler and was compiling source from external RAM disk. The Amiga assembler was compiling source from Perry's Recoverable RAM Disk. I found the Amiga to do the job 98% faster than the AIM. This is a factor of 2 increase. No doubt some of the speed advantage of the 68K was consumed by the inefficency of C and the extra overhead of the operating system. But this also shows that the 6502 is no slouch at application speed. This is especially true when the application is written in well optomized machine code. I suspect that the 6502 would have faired even better if not for a quirk in the AIM assembler that forces all the source through the display at 9600 baud. The 6502 has simple fast instructions, while the 68K has complex powerful instructions. This means that 6502 code can be better optomized to the task. As for emulating the C64, I imagine the tough part would be emulating the special chips in the C64. The actual 6502 emulation could be done fairly well with an optomized 68K machine language emulator. Even then I wouldn't expect the 68K to do any more than keep up with the 6502. Joel Swank Tektronix, Redmond, Oregon