By: RichardC (email@example.com), April 6, 2017 1:49 pm
Room: Moderated Discussions
> > So if you claim that I have a skewed picture of what things were, then right back at ya.
I specifically argue against your suggestion that the wave of early/mid-1980s
cpu's normally labelled as RISC, such as ARM/SPARC/MIPS, were not superior to
CISC designs at that time, but were adopted because of an irrational wave of
non-scientific "RISC religion" or "RISC philosophy".
That doesn't correspond to my experience of using contemporary RISC and CISC-based
machines in that era, including Vax-11/750, ARM-2, M68K Sun-3, and SPARC Sun-4.
I'm fairly sure that it doesn't correspond to benchmark data from that time, which
showed RISC'y machines considerably outperforming the CISC'y competition on integer
And most seriously, to accept that theory you'd have to believe that many major
corporations simultaneously made irrational decisions to invest huge resources in new
architectures which didn't offer a technical advantage: DEC (which had 5 RISC'y projects
in the 1980s, *and* used MIPS, *and* went on to Alpha), HP, Sun, IBM, Acorn (ok, they
weren't "major"), Fairchild, Intel, Motorola.
The evidence, and Occam's Razor, suggests that RISC'y ISAs did offer a compelling advantage
in the 1980s. Though for various (and not easily foreseeable) reasons that advantage
was very much eroded, if not completely erased, in the 1990s as Intel devoted huge
resources to both leading-edge manufacturing and leading-edge cpu design.