20 Nov
2009
20 Nov
'09
8:46 a.m.
On 11/20/09, Brian Lanning <brianlanning at gmail.com> wrote:
On Fri, Nov 20, 2009 at 10:16 AM, Al Kossow <aek at
bitsavers.org> wrote:
> Chuck Peddle told me a couple of months ago that the 6502 was never
> intended to be a general-purpose microprocessor, it was designed to
> be a replacement for hard-wired logic.
Well it certainly exceeded those expectations!
> They had a die size target to hit to get to the
price point they wanted
> and pulled out things they thought were unnecessary for its use in that
> market.
Interesting way to build it - start with a processor... discard things
until it fits in a few mm^2... profit!
> In particular, the length of the registers. I had
always wondered why they
> built a microprocessor with an 8 bit stack pointer, when the previous 6800
> design had 16.
The first two processors I worked with were the 1802 and the 6502. I
do remember the wee stack size was occasionally a problem. It would
have been nice to have been able to at least put it somewhere other
than $0100, but they probably didn't have enough room for spare
transistors to even do that.
As a little-brother to the 6800, it still does a pretty good job. I
remember wishing for 16-bit registers, but in effect, zero-page is a
wad of slow 16-bit registers.
lol Who needs multiply and divide operations anyway?
When I used to write commercial games, we used to go to great lengths
to calculate screen addresses, etc., with tables and hard-coded
multiple routines (times3, times9, done with shifts and adds) since
brute-force multiple wasn't an option.
Great info, I never knew that.
Interesting to hear the history on it.
-ethan