<Do you mean a National Semi 32000 class cpu? Kewl.
Boy, those were
<hot, hot, hot when they hit the market in the early 80's. National
<did everything right on this one: Full object code compatibility
<between the 8,16 and 32 bit versions of the device; truly orthogonal
<instruction set, and so on. Mondo cool. It also was THE FIRST true
<32-bit cpu on the market (according to an EDN magazine article).
Look at that one and then look at VAX... the similarity is very strong.
I don't know; I didn't see it. To me, a processor architecture is hung off
the register structure and how the registers are used. When I read the
introductory chapters of the NS32 book, I got all excited, but when I
actually waded through the description of the machine I felt let down.
The thing about the VAX is that the entire machine is built around the
general purpose register set. Things that are special in other architectures
(immediate values, pushing and popping the stack) are simply side effects
of everyday addressing modes on the VAX. You can pop things from the
stack because you can pop them from any register; MOV R0,(R1)+ works
just as well as MOV R0,(SP)+. Immediate values are fundamentally popping
things from the PC: MOV R0,(PC)+. The NS32K required special address
modes for these operations because SP and PC weren't general purpose
registers.
I'm also not impressed by folks who claim that a 68000 is a whole lot
like a PDP-11 for the same reason...
Roger Ivie
ivie(a)cc.usu.edu