29 Jul
2014
29 Jul
'14
11:48 a.m.
On Jul 29, 2014, at 12:13 PM, Chuck Guzis <cclist at sydex.com> wrote:
Which brings up an interesting question--how many
other machines shrunk
their instruction set during the course of evolution, rather
than the other
way around?
The MC68060 shed a number of integer mul/div, FP transendentals, and some
variants of conditional MOVE instructions that the MC68040 implemented, as
well as some privileged mode instructions like PTEST. I'm sure there are
others I've forgotten. Motorola provided exception handling software that
would emulate (many) of the unimplemented instructions.
MicroVAX dropped BCD instructions. So, I think, did x86_64.
KJ
29 Jul
29 Jul
1:21 p.m.
Ken Seefried wrote:
On Jul 29, 2014, at 12:13 PM, Chuck Guzis wrote:
And before that Motorola had the 68030 dump practically all of the
modular programming stuff that had been added to the 68020 in order to
compete with the National 16032 and the Intel iAPX286. I don't remember
if there was a software solution to handle code that used the missing
instructions.
-- Jecel
Which brings up an interesting question--how many
other machines shrunk
their instruction set during the course of evolution, rather than the other
way around?
The MC68060 shed a number of integer mul/div, FP transendentals, and some
variants of conditional MOVE instructions that the MC68040 implemented, as
well as some privileged mode instructions like PTEST. I'm sure there are
others I've forgotten. Motorola provided exception handling software that
would emulate (many) of the unimplemented instructions. 
3:02 p.m.
On Tue, Jul 29, 2014 at 2:21 PM, Jecel Assumpcao Jr.
<jecel at merlintec.com> wrote:
And before that Motorola had the 68030 dump
practically all of the
modular programming stuff that had been added to the 68020 in order to
compete with the National 16032 and the Intel iAPX286. I don't remember
if there was a software solution to handle code that used the missing
instructions.
It wasn't a problem because there *wasn't* software that used the
missing instructions (CALLM, RETM).
So making that statement will result in someone pointing out that some
obscure piece of software somewhere used them, but that basically
proves the point.
4:10 p.m.
That's interesting; I was never a "down-n-dirty" assembly language
programmer on the 68k but I always had casually assumed that the 68030 was
a strict superset of the 68k ISA as implemented in the '020... Now my
curiosity is piqued... I'll be dipping in and out of those datasheets all
night ;)
Best,
Sean
On Tue, Jul 29, 2014 at 6:02 PM, Eric Smith <spacewar at gmail.com> wrote:
On Tue, Jul 29, 2014 at 2:21 PM, Jecel Assumpcao Jr.
<jecel at merlintec.com> wrote:
And before that Motorola had the 68030 dump
practically all of the
modular programming stuff that had been added to the 68020 in order to
compete with the National 16032 and the Intel iAPX286. I don't remember
if there was a software solution to handle code that used the missing
instructions.
It wasn't a problem because there *wasn't* software that used the
missing instructions (CALLM, RETM).
So making that statement will result in someone pointing out that some
obscure piece of software somewhere used them, but that basically
proves the point.
5:04 p.m.
On Tue, Jul 29, 2014 at 5:10 PM, Sean Caron <scaron at umich.edu> wrote:
That's interesting; I was never a
"down-n-dirty" assembly language
programmer on the 68k but I always had casually assumed that the 68030 was
a strict superset of the 68k ISA as implemented in the '020... Now my
curiosity is piqued... I'll be dipping in and out of those datasheets all
night ;)
Also the 68030 doesn't implement the full instruction set or
functionality of the 68851 PMMU.
1:47 p.m.
On Jul 29, 2014, at 2:48 PM, Ken Seefried <seefriek at gmail.com> wrote:
The MC68060 shed a number of integer mul/div, FP
transendentals, and some
variants of conditional MOVE instructions that the MC68040 implemented, as
well as some privileged mode instructions like PTEST. I'm sure there are
others I've forgotten. Motorola provided exception handling software that
would emulate (many) of the unimplemented instructions.
Perhaps my memory is faulty, but do I recall that the 68040 FPU
eschewed a lot of the transcendental functions as well due to die
space constraints?
- Dave
3580
days inactive
3581
days old
5 comments
5 participants
participants (5)
-
fraveydank＠gmail.com -
jecel＠merlintec.com -
scaron＠umich.edu -
seefriek＠gmail.com -
spacewar＠gmail.com