knudsen@ihnss.UUCP (11/06/84)
I am going to partially agree with Brad T. that the 6502 is the fastest thing
in tight corners ever made. I bought a KIM-I 6502 hacker board in '77,
3 years before my Coco, so I can talk about both chips. Yes, the 6502 is
faster in small, tight loops. I work a lot in software music synthesis
(a la Hal Chamberlain's 10/77 BYTE article) and have put his algorithms
on both micros.
However, his scheme takes only 100 cycles on 6809 vs 112 or so on
the 6502, so even here the 6809 comes out ahead, due to its 16-bit load/sore/
adds and 16-bit index registers (used as counters).
I've also written assembly graphics routines to set points, draw lines,
etc, and all those fancy addressing modes that the high-level language guys
crow about really speed things up down there in the gutter too! Multiply
instruction sure doesn't hurt either! Best of all, you needn't use all the
6809 features at first -- it's as easy to program as a 6502 -- in fact, you
almost always pretend it IS a 6502, then go back and recode -- watch your
assembler source shrink!
Motorola did make a few booboos -- no direct page indirect,
and all memory referencing instructions waste a cycle somewhere relative
to the 6502. Even so, at the same clock speed a 6809 will still come out
at least equal to the 6502, even in music, graphics, and ... GAMES!
> Nothing can touch a 6502 .....
Consider yourself touched. mike k
PS: My C benchmark would be 18.3 sec if the Coco ran at 1.00 MHz like a C64.