[rescue] i860 Success

[Francisco Javier Mesa-Martinez]

On Sun, 2 May 2004, Charles Shannon Hendrix wrote:


> Yes it was best called VCISC, but the i960 is hardly RISC in the
> traditional sense.

Yah, Intel managed to use the term rather loosely on that chip though...


> The i960 came out of the Gemini/P7 project at Intel, which was based on
> the 432 design, but with changes to the ISA to lean toward RISC.  It
> still had a lot of the design and ideas in there, including the OO parts
> and tag bit.

Actually it is 2 generations behind the 432, it is almost like an
anti-432. The 960 had no tag support in its initial introduction. Tag
support was not what made the i432 OO, it also needed the complex stacks
and the weird prottection at the instruction level. Which the 960 does not
have, I think it was introduced in later iterations... the 960 is a pretty
large family of chips it seems.

> The goal was to have the feature set of the 432, but speed could beat
> something like a Motorola 68K.

This was not the 960 but the second iteration of the 432 which was canned.
This was the chip that Intel and Siemens were working for their mid 80s
venture.

> After that, Intel removed the OO parts and the tag bit, and that became
> the i960.  The i960 was a very good CPU, but Intel marketing buried it
> because it would have competed with the 80386 which was in competition
> with it at Intel.  Intel actually finished the i960 first, but many
> people claim they sat on it while the 80386 was finished.

Actually the 386 was out in 85, the 960 came out later. And it was not
geared towards general CPU, that was the 860. The 960 actually outsold any
other x86 chip for a long time before it was moved a bit by the strongARM
line. So I do not think that it was a failure in any sense, I doubt intel
would sit on a design for 4+ years just because they love the 386 so much.

> Intel had a golden opportunity to get away from x86 right there.

Highly unlikely, plus as long as people buy intel, they do not care which
line you buy. The problem was that in the 80s there were a lot of CPUs out
there, as soon as intel said here is a new CPU totally incompatible with
the x86 do you want it? Then people can look around and see that they now
have 10+ alternatives, and they are now not bounded to chose intel.

> > It was a fairly cheap, clean simple design and I assume it made it as
> > an embedded design because of those three factors. Plus intel did not
> > want anything to interfere with the x86,
>
> It was not designed as an embedded CPU, that was a marketing decision.

I will contact one of the members of the 960 team which may disagree with
this statement.

> > I believe the 960 was also developed as a contract with the DoD for
> > embedded processors.
>
> No, that was the 432.  It was developed after the DoD "everything in
> Ada" edict.

There were later contracts the 432 was late 70's the 960 was late 80's
(it is not like the DoD only contracted in the 70's), the 960 actually was
developed with hardened fabrication process in mind. The first sampled 960s
actually were the
hardened variety since military dudes were among the first clients. This
is from the fish's mouth btw. Again from talking to this dude it seems intel knew
what they were doing with the 960, the 860 was their general purpose path
from the get-go.

> http://www.brouhaha.com/~eric/retrocomputing/intel/iapx432/
>
> The P7 was used in a DoD project, but was not designed for it.
>
> These days, we have the transistor budget to make chipset like the 432
> fast enough for most any task.

The transistor count was not the problem, as the 432 was implemented in a
multipackage process. It was just a shitty design, plus it was a warning
of what happens when too much effort is put on complexity and the design
team wanders away from making the common case fast.

> After thinking about it a bit, I think maybe they pitched it as an
> alternative to x86 because almost everything we did in college and in my
> first embedded work was Motorola.  In fact, when I did see Intel or even
> Z80, it was usually when we were replacing it with Motorola hardware.

Oddly enough Intel was supposed to be dying in the 80's, I believe IBM had
to actually inject some cash in the company at some point. Motorola and
the new RISC people were supposed to be taking over, funny how things
turned out.

> Every shop we visited and most of the industry rags I read were all
> about Motorola.  It was especially true if you did your work on VMEbus.

Which sort of makes sense since VME is motorola's pretty much.

> The Intel marketing droid may have been saying, "If you find x86 a bad
> fit, we have the 860 which is much better..."

There was major criticism of the x86 in the 80s, people thought the 286
was going to fix some of the crap teh 8086 had introduced. But it just
made things even worse, then the 386 was still not what people wanted but
they bought it anyways because it could run 8086 code very fast. Then
intel said, OK here is the 860 with all the bells and whistles, now they
expected people to shut up. In fact windows NT was first designed with the
80860 as the target architecture, it was view as the next iteration of
intel in the desktop. It was that important at that time. Unfortunately
backwards compatibility is what people demanded, oh well.