Nice CPUs (was: Virtual vs Physical memory (was Re: Designing around a 6502 ))

From: Dwight K. Elvey <dwightk.elvey_at_amd.com>
Date: Wed Feb 12 15:16:01 2003

>From: "Hans Franke" <Hans.Franke_at_mch20.sbs.de>
>
>> >> > A segmented architecture...
>> >> I have one word to say about segmented architecture... "Ewwwww"
>
>> >:) I like segmented systems. It's one of the best concepts for
>> >address space extension. Ok, I also think that the 8086 is one
>> >of the best 16 Bit CPUs (*1) ever, and as the 186 core at the
>
>> Then you'll love the Z8000. Its segments are non-overlapping
>> ( unless a mmu makes them so ).
>
>I never had a problem with overlaping. Programming is always
>a matter of diszipline, and well written code never accesses
>unassigned data :)
>
>The Z8000 was always a bit weired. a nice design, but weired.
>Ok, I never programmed it, only thinking around when I had to
>choose a processor for a system around 1980. The candidates
>where 16016, 68000, Z8000, 8086 and 9900.
>The 8086 droped out because of the complicated bus structure,
>the 9900 because 64K where not enough, otherwise it would have
>been my choice (and the 99000 wasn't realy available). So the
>finalists where National vs. Motorola vs. Zilog. After all,
>the NS16000 did offer the best extensibility, a complete and
>beliveable family model and a real nice and clean programming
>model. So 16016 it was.
>
>> I like the 186 myself but have to admit that the time for
>> segmented memory has passed, in general purpose computing.
>
>Well, define general purpose computing. In my eyes this
>describes real computers ... ala IBMish Mainframes, and
>Segmented memory never had a place there. When going for
>small systems, the 186 is still one of the best CPUs to
>use. Powerfull, fast, simple and high integrated. I don't
>know anything which can't be done in a megabyte of mem.

Hi Hans
 We agree here. The only hard part of working with the
186's is getting the init sequence debugged. Once that
is done they are great.
 I don't think that I've ever written
a program that took more than about 35K ( not counting data
space ). Then again, most of the larger programs I've written
were in Forth so it is hard to compare.
Dwight

>
>Gruss
>H.
>
>--
>VCF Europa 4.0 am 03./04. Mai 2003 in Muenchen
>http://www.vcfe.org/
Received on Wed Feb 12 2003 - 15:16:01 GMT

This archive was generated by hypermail 2.3.0 : Fri Oct 10 2014 - 23:35:54 BST