19 Aug
1998
19 Aug
'98
11:26 a.m.
Another problem is that nearly each machine used a
unique MIPS processor
which were more source compatible than binary compatible. Well, that isn't
exactly true---the opcodes were the same, but each MIPS processor has a
slightly different pipeline and code compiled for one MIPS chip (say, the
R3000) might not work correctly (or work at all) for another (say, the
R10000).
This really is no different from the early SPARC boxes - Sun-4 really did
not like Sun-4c, etc..
Another thing---early SGI boxes were more stable
than later ones. The one
I used for four years (SGI Personal Iris 4D/35) had the mother board
replaced once, the graphics card replaced once and the monitor was going
(everything had a slight green tinge to it).
Luckily I think I have enough spares. I even yanked the backplane out of
the donor machine. The two machines had different graphics boards - more
research I need to do (first I need to get the box out of the van!).
I enjoyed working on the box, but the hardware
flakiness got to me after
a while. I was also upset when I had to upgrade the OS from 3.3.2 (last
version to run NeWS (much better than X)) to 4.0.1 (first version based on
X). Also, avoid IRIX 5.x at all costs (it was their worst version, and
quite possibly the worst version of UNIX this side of SCO).
I think the IRIX support for the Power Series (or probably most of the 4D/
machines) died around v5.0.x.
William Donzelli
william(a)ans.net