22

Intel infamously developed the iAPX 432, a 32-bit processor that had built-in support for OOP, GC, Multitasking and other high-level concepts. However, I can't find any articles that attest to it actually being used in real-world applications. Was any computer (home or business) ever released with this processor, or any even remotely notable embedded device?

Improve this question
10
  • 8
    Given it's poor real-world performance compared to other (and cheaper) systens, I was under impression that no one actually used it. Commented Mar 29 at 19:11
  • 2
    Interesting. I was working at Burroughs in the early 1980s (Federal and Special Systems Group), but was unaware of that. I have a bunch of iAPX 432 data books collected in the mid-1980s while working at Apollo Computer -- I guess there's no reason to hang onto them. Commented Mar 31 at 15:47
  • 4
    @dan04: No, we did not. But the sales folks were very enthusiastic about handing out the books. Coming from Burroughs, I thought the architecture looked interesting on paper, since it incorporated many of the concepts I knew from there. But it wasn't a good fit for the Apollo software architecture. Never got as far as playing with hardware. Commented Apr 1 at 1:03
  • 5
    An anecdote: A long time ago, per a roommate/friend who worked at Digital/VAX engineering ... Intel went there to tout the 432 to try to get DEC to use it. The DEC guys: You have memory segments of [only] 64 KB. How do you [linearly] go from the end of one segment to the next? Intel: Why would you want to do that? DEC: When you have arrays of more than 64 KB. Intel: Who has those? DEC: You may leave the building ... Commented Apr 2 at 21:38
  • 4
    @dan04 - IIRC memory was object-oriented - so you get yourself an (Ada) object instance and address (relatively) inside of that. But the size of those instances was limited. (But I could be very wrong.) Commented Apr 3 at 17:45

1 Answer 1

Reset to default
16
+250

Five Eden Node Machines with iAPX 432 processors were built at the University of Washington as part of the Eden Project during 1980-85. This was not a commercial product, but it was a 432-based system not built by Intel.

It also appears that High Integrity Systems in the UK built a "Multibox Computer" based on multiple 432s. Search for "Multibox" on that page for the claim. The referenced website at eight6.net resolves, but times out. A little of it is visible via the Wayback Macine.

That's all I can find, but it indicates some use.

Incidentally, the 432 development team moved on to working on another new Intel processor, the i960 family. This was fairly successful as a powerful embedded processor in the 1990s, but Intel stopped promoting it when they acquired DEC's StrongARM family, which presumably ran faster and would have been cheaper to manufacture. It remains in use in some military applications where the strong security model of the high-end versions is appreciated.

In 1990, the evolved version of the team became one of two x86 development teams and was responsible for the Pentium Pro, released in late 1995. That was the point at which x86 became extremely competitive with the RISC processors of 1995-97, outperforming all of them except Alpha and causing major rethinks at Sun and MIPS. It wasn't a great success in the Windows 95 market because it didn't run 16-bit code as fast as 32-bit, but it had major effects in the technical computing and server markets.

Improve this answer
6
  • 1
    This wayback links gets you a BIT of content: web.archive.org/web/20230326041422/https://eight6.net Commented Apr 2 at 14:21
  • @MauryMarkowitz: Thanks, added. Commented Apr 2 at 18:20
  • 2
    It might be also interesting to note that most of the 432 development team went ahead to create the i960 (re) using many of the design ideas. Quite a successful embedded CPUand the highest selling RIC around 1990. It incorporated many features ahead of time - including a tiered design, kind liketoday's RISC-V. Too bad Intel management didn't promote it as GP-CPU. The same team BTW was later responsible for the PPro, a design which saved Intel during the first years of the late 90s MHz-race. Commented Apr 2 at 22:47
  • 1
    @Raffzahn: How's this? I disagree that PPro "saved Intel" because they weren't in big trouble. It's more that it enabled them to grow into more markets and persuaded them to start their MHz war with AMD. Commented Apr 3 at 20:58
  • @JohnDallman Exactly as you said. Note I said late 1990s. Early on AMD's ability to match the original Pentiums with their 486 designs came as a shock. Intel expected a way longer time for them to catch up, but AMD effortlessly moved up to 75, 100, 133 and finally 160 MHz (at way lower prices- chip and TCO), while P5 and P54 had problems to scale (plus power issues). P55 gave a limited relief (with lots of marketing), still pressured by the K6 family. It was the 1997 P6 based Pentium II that gave intel a clear edge (consumer and professional) for a good 2+ years until the K7 finally hit. Commented Apr 3 at 22:48

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.