Re: Hardware emulation of 6509 using 6502?

From: Mia Magnusson <mia_at_plea.se>
Date: Wed, 7 Mar 2018 17:38:12 +0100
Message-ID: <20180307173812.000036a4@plea.se>
Den Tue, 6 Mar 2018 21:39:09 +0000 (UTC) skrev Steve Gray
<sjgray@rogers.com>:
> I'm guessing that since the B and P series use virtually identical
> memory maps that it was easier just to put a small amount of ram in
> the same place on both machines.

Yes, but they could have mapped in a chunk of the DRAM instead of
having a separate SRAM which mostly duplicates the function of the DRAM.

> You'll also note that in the P500 the character ROM is mapped in,
> but on the B-series that space is empty (since the 6545 circuitry
> reads it directly). I actually designed a board to re-map memory from
> bank (something) into Bank 15 way back in the day. I ended up killing
> my B128 since I didn't really fully understand how the 6509/B128
> really worked back then (and not sure even now)... For it's time I
> don't think the 6509 was a bad idea. It was a CPU that's 6502
> compatible so there was a large software base, yet extended the
> memory to 1MB without changing a lot of things. Sure, they probably
> should have designed a fully 16-bit CPU but the 6509 would have been
> a fairly simple and cheap upgrade to do. Steve

The big question is if they could have done the bank switching any
better. The Commodore 128 is the first that springs to mind, and that
computer isn't really that good either.

What they could have done is firstly, as Michał already stated, a
better way to jump between code in different banks. But they could also
have made it possible to either have a variable split in the adress
space between two banks, and made it possible to select which adressing
modes that should use which of the bank registers.

If anyone want to read up on how other manufacturers solved the 64k
barrier in 8-bit computers there is the 64180 Z80 compatible CPU, and
of course the 65802/65816. Maybe there are some stuff for the 6809 CPUs
too?.

If MOS/CBM really wanted a better solution for 1MB, and if the right
people were still at MOS, they could have made something similar to how
Intel solved the barrier with the 8088. The Intel 8088 solution could
be applied to an 8-bit CPU if someone really had wanted to do that.
Actually it could be done today with just some changes to the hardware
that this thread discusses :)

-- 
(\_/) Copy the bunny to your mails to help
(O.o) him achieve world domination.
(> <) Come join the dark side.
/_|_\ We have cookies.

       Message was sent through the cbm-hackers mailing list
Received on 2018-03-07 18:05:44

Archive generated by hypermail 2.2.0.