Yokemate of Keyboards
Posts: 4967 from 2009/1/28
From: Delaware, USA
> people trying to adopt PPC found it difficult to get support chips.
...except when adopting a PPC SoC
PPCs SoCs sucked. They were full of bugs (being mostly untested), hard to get a hold of in any quantity, horribly underpowered, incompatible with desktop PPC instructions, or all of the above. There's a reason everyone in embedded moved on from PPC to MIPs (then ARM) in the early 00s.Quote:
> When it came to NB and SBs, generally it was a case of
> a. capable b. available c. bug-free: pick any two.
This was never true for southbridge chips.
Sure it was. Even the VIA chip the Pegasos used was a buggy piece of shit. No bugs quite as fatal as Articia, but a lot erratas to be worked around. And certainly it was never high end.Quote:
> to my understanding ARM desktop never really took off anyway.
Funnily, it was the desktop where ARM started
StrongARM maybe, but not ARM as we know it. And StrongARM CPUs, while interesting, provided no real advantage over Intel, AMD, and Motorola CPUs of the era. An Acorn Archimedes was really no faster than an Amiga 500.
Actually, the only decent Southbridge components that worked well with PPC cups and Socs were from ULi and ATI. When then X1000 was announced it was no big surprise that it utilized an SB600. That component was well documented, fairly compatible with the ULi Southbridges used on Freescale's evaluation boards, and relatively big free.
I've always favored ATI/AMD components and hope at least one of the systems supported by MorphOS NG is AMD based.
I'd still like Power9 support, though.
"Never attribute to malice what can more readily explained by incompetence"