The 65816 -is- 16-bit! (Re: Does SNES really...)

[irsman@iastate.edu (Ian Schmidt)]

[crossposted so csa2 experts can have a crack at it]

>                     Sega Genesis    SNES       Lynx     IBM PC     386SX PC 
>                     ------------    ----       ----     ------     --------
>processor            68000           65816      6502     8088       80386SX
>external data bus    16 bits         8 bits     8 bits   8 bits     16 bits
>internal data bus    32 bits         16 bits    8 bits   16 bits    32 bits
>clock speed          7.6MHz          3.6MHz     4.0MHz   4.77MHz    16-33MHz


>The average number of cycles per instruction and the richness of the 
>instruction set and addressing modes determines which is faster, the 65816
>or the 68000.  I have programmed the 68000 and the 6502, and assuming the
>65816 is similar to the 6502, I'd say the 68000 in the Genesis can do raw
>data-crunching much faster than the 65816 in the SNES.  Of course, raw
>data-crunching is only part of what goes on in a game, so if you want to
>know which is better, the only way is to try both of them out.  An example
>of how raw CPU speed isn't everything, try to find any game on an original
>IBM PC that even comes close to the SNES in speed or quality.  Just by 
>looking at the CPU speed, you would expect the PC to be a little bit faster,
>but this obviously isn't the case, since the graphics hardware on a PC is
>garbage compared to the SNES.

That's correct, and the 65816 uses a lot fewer average cycles/instruction than
the 680x0 or 80x86.  The general formula for real world performance per Mhz
looks like this...

80x86 Mhz = 2 * 680x0 Mhz = 2 * 65816 Mhz.

Of course, that's solely based on raw CPU performance.  Machine design in 
general throws a lot of hooks and curves into it; game consoles typically
aren't the best designs in which to exploit a CPU's power.  Of course, neither
are a lot of computers...a 486-50 running on an 8 mhz ISA bus is pretty sad :)





-- 
                                          Ian Schmidt: irsman@iastate.edu
"I will choose a path that's     |  /~~~\              irsman@aol.com
 clear: I will choose free will" | |    < o  o  o  o  o  o  o  o  o  o  o
 - N. Peart, 1980                |  \___/              twbv4@isuvax.bitnet