13 Nov
2003
13 Nov
'03
12:21 p.m.
On Thu, 2003-11-13 at 13:37, Steve Thatcher wrote:
being an engineer that worked in assembly language on
many micros
before and after the 8086, the segmented architecture was not
that hard to handle and actualy had many side benefits such as
code size and more efficient use of the bus bandwidth. The pdp11
may have been better overall, but there is no comparison on terms
of price, availability, and being able to get the job done for
the millions of PC users.
best regards, Steve Thatcher
Cool, CPU flame wars! :-)
First, you state: "...may have been better overall, but there is no
comparison on terms of price, availability, and being able to get the
job done for the millions of PC users". Intel's ability to 'get the job
done' is real, but hardly due to the 8080 legacy they dragged around.
Clearly, DEC completely ignored the chip-computer world, until it was
far too late. (I was the one who ported MSDOS to the DEC Rainbow 100A, I
could tell you some terrible DEC stories... :-(
I too wrote assembly on more minis and chips than I recall, and from a
programmers point of view, Intel sucked/s. Segmentation had hardware
advantages for Intel, for buss and peripheral component compatibility,
and that's about it. Don't you remember address normalization and
comparison woes? Memory allocation schemes? Small/Medium/Large Model
compiler "options"? Different and incompatible subroutine calls and
returns? segment:offset performance penalties? Setting up DMA
controllers to get around the 64K absolute boundaries? Hardware address
space kludges? Yuck.
I don't see where 'buss bandwidth' is affected; and if there was any
effect, it was to put the complexity into the software that had to
manage address space calcs in software. And code size efficiencies were
marginal at best; this thread I admit is outside my scope, and the realm
of many papers, related to the RISC vs. complex order code arguments of
yore.
I did little work in the DEC world (some PDP8 for fun, and the Intersil
6100) so I'm not in their "camp". In fact, 8080/8085/z80/8086 assembly
dominated for me, so if anything, I should love them more.
They were cheap and available though, which is why they were popular.
Good code could be written on it, for sure, but it wasn't "inherently"
efficient.
All CPUs suck, but some are worse than others. The nicest-looking chip
computer assembly code (in my opinion) is the Moto 6809, though I never
got to use it. The worst, besides PICs, the Signetics 8x300, or may be
the RCA 1802 (ugh)..
tomj
PS: anyone else notice the nerdily amusing puns in the Microchip PIC?
Basically it looks like a PDP8 order code, complete with paged jump
(modulo 256d, not 0200) plus the order descriptions use FORTRAN4isms:
.AND.
.OR.
!
--- Original Message ---
From: "Eric Smith" <eric(a)brouhaha.com>
To: "General Discussion: On-Topic and Off-Topic Posts"
<cctalk(a)classiccmp.org>
Date: 11/13/03 3:24:07 PM
"Hans Franke" <hans.franke(a)mch20.sbs.de> wrote:
to me
it's the way the memory is handled
that makes the 8086 the great CPU it is
What, the 64K segments that alias on paragraph boundaries?
Yecch! What a kludge! The PDP-11 had better memory management
for a 64KB address space at least seven years earlier.