Re: 6502, 65816 = RISC ?

[nathan@visi.com (Nathan Mates)]

In article <3211B31A.27EC@esis.vsnet.ch>,
Gilles Tschopp  <gilles.tschopp@esis.vsnet.ch> wrote:
>For me, I consider the 6502 (and 65816) as RISC and not CISC. 
>The RISC is only a philosophy and not a technical way of 
>manufacturing CPU. 

   True, but the RISC philosophy that's espoused by the computer
industry is roughly the following: lots of registers (16-32 is
common), and very few instructions that can operate very quickly
on all registers. The extreme RISC schools say "if an instruction
can be done by a small combination of others, throw it out." That
leads to chips with no way to increment a register-- they must add
1 to it. 

   The 6502/65816 were designed with a small number of opcodes, but
size of opcodes is not the only factor.

> I still have my Apple IIc and I always feel 
>that the 6502 is quicker than the 8086-4.77 (and even the 
>8086-9.54 from my old Toshiba portable), and is comparable with 
>a 286-8 or 10.

   Comparing raw processor speeds against each other is a BAD way to
compare chips. The number of cycles to do a similar instruction can
vary between different chips. A much better way to compare chips is to
do a lot of things, and see how long (real world time) it took to do
that.  That's the basis for benchmarks, such as Dhrystone.

   Further, the 8086 is a 4-bus clock design vs the 1 bus clock
of the 6502, something that is not part of RISC either. Basically, to
read one byte from memory on the 8086, it takes 4 clock cycles, vs the
1 cycle on the 6502. Thus, the 4.77Mhz speed acts a lot more like 1.19
Mhz. 

>Moreover, the 6502 has only five registers and can manage 
>directly the memory. So it has less opcodes to execute, compared 
>with the 8086, and can be considered as RISC.

   RISC wants more registers, because it's faster to grab stuff out of
registers than to hit on memory. Next, you should be able to do the
same operations (add, subtract, etc) to all registers. 6502/65816 fail
that test miserably.

>For you, can you consider the 6502 (and 65816) as RISC ? 

   I call it "close but not quite" to RISC.

Nathan Mates
--
<*> Nathan Mates http://www.visi.com/~nathan/      <*>
# What are the facts? Again and again and again-- what are the _facts_?
# Shun wishful thinking, avoid opinion, care not what the neighbors
# think-- what are the facts, and to how many decimal places?  -R.A. Heinlein