Havard architecture allows simultaneous access to program memory and data memory, whereas classic microprocessor (neuman or van Neuman??) must multiplex both data and program on the same address/data bus.  For DSP processors, this allows a multiply instruction to fetch constant data from program memory and variable (signal) data from RAM and multiply all in one instruction cycle.  For the PIC, it mostly just gives faster execution with a simpler internal architecture (no large multiplexers). It also allows somewhat more efficient opcodes.  Since the data address space is much smaller than the code space, the instruction can be fixed width and still encode the data address in the instruction (i.e. every instruction is 1 word long).  With the 80x86 and Motorola families, the opcode size varies depending on where in memory (and which addressing mode) you are accessing.

----------
From:   Tom Sgouros
Sent:   Tuesday, December 23, 1997 2:12 PM
To:     PICLIST@MITVMA.MIT.EDU
Subject:        pic architecture

Can someone explain in a nutshell what is the advantage to the Harvard
architecture?  Or perhaps someone who's used other microcontrollers
can give me some perspective on PIC advantages.  (Don't want to start
a flame war, though, so relax, everyone.)

I am using PICs for a couple of projects (one of which works fine now,
thanks to some list help, by the way) largely because someone whose
opinion I trust suggested it last year, when I was young and
impressionable.  I've noticed, however, that *he* doesn't use them.

-tom

----------------------------------------------------------------------
tomss at ids.net - 401-861-2831 - 42 Forge Rd, Potowomut, RI 02818 USA