Personal computing discussed
Moderators: renee, Flying Fox, morphine
bjm wrote:Anyway, for the sake of discussion, if we could replace the ISA retroactively with hindsight being 20/20, I think MIPS would make for an interesting candidate. Like ARM, it has a core that can be licensed and I hear it has a nice assembly language to go with it. But unlike ARM, its old enough to where if it had gained prominence at its inception in the early 80s, it could have supplanted x86 in the IBM PC had it been cheap enough.
just brew it! wrote:PowerPC is more of a niche product now that Apple doesn't use it any more, but it is still used in IBM servers and in high reliability military/aerospace applications.
just brew it! wrote:Agreed. What was interesting was that Wozinak came very close to using the 6800 in what became the Apple ][, before settling on the 6502 (for cost reasons, IIRC). Had he gone with the 6800, ancestor to the 68000 found in the original Macintosh, Apple would have had a much cleaner upgrade path for all the Apple ][s that were out in the world (and particularly in education), and the Apple III might have been less of a failure and more of a bridge.On the CISC side of the house, Motorola's 680x0 series was a much more sensible ISA than x86. But IBM went with the x86 for the first mass-market PC, and the rest is history.
Well, in its day Alpha was widely regarded as the clear technical winner, as it pioneered a lot of innovations that were later used by x86. And, like MIPS, there was a Windows implementation for it (and DEC even had a fancy JIT cross-compiler that enabled x86 binaries to run fairly well on it). But as you say, it was a victim of larger forces and bad management/vision -- which is pretty much part of the eulogy for any good but commercially-doomed processor design you can think of. And, speaking of DEC, they had their own ARM implementation too, which ended up going to Intel in a lawsuit. Intel called it XScale and was using it to replace their own ix60 RISC designs, before they decided to get out of that business entirely and passed it on to Marvell.With RISC I don't think there's a clear technical winner. MIPS would've probably had a better shot if they hadn't gotten sidetracked by being owned by SGI for a few years; they're currently playing second fiddle to ARM in the consumer embedded space. Alpha was a promising architecture, but Compaq effectively killed it when they sold it to Intel (some of the related tech lived on elsewhere though, e.g. the design of the original Athlon system bus). PowerPC is more of a niche product now that Apple doesn't use it any more, but it is still used in IBM servers and in high reliability military/aerospace applications. SPARC is still used by Oracle and some supercomputer vendors, but was somewhat marginalized by Sun's decline and the rise of cheap x86 servers (SPARC is interesting in that there are Open Source implementations of the ISA).
PowerPC is more of a niche product now that Apple doesn't use it any more, but it is still used in IBM servers and in high reliability military/aerospace applications.
SPARC is still used by Oracle and some supercomputer vendors, but was somewhat marginalized by Sun's decline and the rise of cheap x86 servers (SPARC is interesting in that there are Open Source implementations of the ISA).
Had Don Estridge and his team chosen, say.. Motorola's CPU at the time, I suspect that the 68k ISA would have become the x86 of today