Jonathan de Boyne Pollard
2010-03-22 22:19:07 UTC
I liked the 6809 instruction set, and *really* like the 68000
instruction set.
I'll second that (since no-one else in the thread has).instruction set.
OK. I spent a while once converting code from Z180 to 68HC11.
The 68- code consistently came out a few percent smaller and faster.
It seemed to be because of
* the dual A and B accumulators
* the more compact conditional branch coding.
? more compact coding for X register indexing. (Y/IY was about the same.)
data, but upon the other things mentioned elsewhere in this thread,
including the ease with which it was possible to convert from machine
code to assembly language, and the whole
let's-just-encode-all-registers-with-a-uniform-encoding idea. The
difference between X, Y, U, and S indexed and indirect addressing in the
6809 is a simple uniform set of two-bit patterns.