Re: CISC vs RISC concepts -- from an assembly view
- From: Jerry Coffin <spamtrap@xxxxxxxxxx>
- Date: Wed, 26 Apr 2006 22:55:47 -0600
In article <444fcb67.17245117@xxxxxxxxxxxxxxx>,
spamtrap@xxxxxxxxxx says...
[ ... ]
In the 80s, there were only AMD and NEC which produced x86 compatible
CPUs using the original intel designs.
There was a third licensee, though they only chose to
exercise their license a few times. The third licensee
may well have been the most important one to IBM though
-- because the third licensee _was_ IBM themselves.
They produced (for one example) one of the first CPUs
developed specifically to reduce power for mobile use,
the 386SL.
NEC had made some
pin-compatible optimizations for the 8086 family called V20 and V30,
which were somewaht faster and could be used as direct 8088/8086
replacements. AMD did not design its own x86 chips until the 32bit
machines came up (afair, they started to split from intel beginning
with some 486 design).
That much coincides pretty closely with my recollection.
And IIRC, it was the other way round: Intel was not giving its design
away until they had a demand created by the IBM PC which they could
not serve on its own after it started beeing successful.
AMD and Intel had a comprehensive cross-license in place
years before IBM and the PC entered the picture at all.
The cross-license _was_ renewed about the time of the
original PC design, and it may well be that IBM demanded
(or at least asked for) it, but it was an established
fact well before.
For quite a while, the AMD/Intel cross license probably
did more for Intel than it did for AMD. CP/M was designed
for the Intel 8080, which gave them a strong market
presence, but in many people's opinions, the AMD 2901 and
2911 was really the superior design of the time (the 2901
was a bit-slice processor. Each bit slice processor
worked on 4 bits at a time (if memory serves) but they
were designed to be ganged together, so you could create
a processor of more or less arbitrary size (as long as it
was a multiple of 4 bits, of course). The 2901s had a
minuscule instruction set (3-bit opcode field -> 8
instructions?) and the 2911 was a microcode sequencer to
combine those into a higher-level instruction set.
--
Later,
Jerry.
The universe is a figment of its own imagination.
.
- References:
- CISC vs RISC concepts -- from an assembly view
- From: HellsRaison
- Re: CISC vs RISC concepts -- from an assembly view
- From: Jeffrey Schwab
- Re: CISC vs RISC concepts -- from an assembly view
- From: TS
- CISC vs RISC concepts -- from an assembly view
- Prev by Date: Re: lea with fs override
- Next by Date: Re: GNU AS and multiplatform example ...
- Previous by thread: Re: CISC vs RISC concepts -- from an assembly view
- Next by thread: Re: CISC vs RISC concepts -- from an assembly view
- Index(es):
Relevant Pages
|
