[rescue] newest rescue

Wed Feb 3 10:59:48 CST 2016

Thanks for the info everyone, I have a better sense of the 680x0 Suns now!

John, regarding the Macs of that era, I wouldn't say Apple was making
'toys', but they had some issues around their software stack and legacy
support issues. When I got into 680x0 Macs, Apple had 24-bit color with
multiple monitor support and arbitrary resolutions and arrangements, 32-bit
addressing, multitasking (though only cooperative) and virtual memory. FPUs
were actually standard in Macs for a long time, starting with the Mac II,
and only started to be omitted much later when Apple made more of a
consumer push with the LC and IIsi line and started to do cost cutting. You
may be thinking of the fact that the Mac II included Apple's MMU in the MMU
socket rather than the Motorola PMMU, and Apple's MMU didn't support
virtual memory, so you'd have to replace it and upgrade the ROM to use
virtual memory in later OS releases. While Apple didn't initially include
TCP/IP with their OS, it was part of System 6 starting in 1988 and later on
Apple added TCP/IP support to their filesharing with AppleShare IP.

RAM was very expensive back then so while in hindsight it was a mistake to
release computers with 32-bit processors but only do 24-bit addressing on
them, I can understand why it happened. The IIci was when Apple cleaned up
the architecture, being the first Mac with a 32-bit clean ROM. With it's
support of up to 128MB of RAM, three NuBus expansion slots, and a PDS
expansion slot that came with an L2 cache card for the processor and could
be used for any number of upgrades (I have a 50Mhz 68030 accelerator in
mine right now), it was a very capable machine. Expansion card options
included 100Base-T ethernet, video cards, DSPs, sound cards, and of course
the MacIvory Lisp Machine. There was even a version of the IIci that Apple
made for the government/military which took ECC RAM.

The SE/30 was also a great machine because it was a 68030 (so integrated
PMMU) with a 68882 FPU and a 32-bit PDS slot that could take a variety of
expansion cards and accelerators as well (color video cards, networking
cards, CPU accelerators, etc.)  It could also take a maximum of 128MB of
RAM. And if you can track down an early era IIsi, it's 32-bit clean ROM is
on a card and can be swapped into an SE/30 to make it 32-bit clean. In my
experience both the IIci and SE/30 ran NetBSD well.

The IIfx was great because not only was it a 40Mhz '030, but lots of
co-processors for I/O and under AU/X had full DMA support to offload most
of the I/O. Unfortunately nobody has ever supported all the IIfx DMA under
NetBSD, so it's really just a good AU/X machine. My understanding is that
there was a lot of AU/X at NASA at one point.

NeXT hardware was great because it was like the IIfx but better, with every
NeXT having a Motorola DSP co-processor and a cube with the NeXTDimension
video card being an amazing device with terrific video in and out
capabilities and capable of offloading the Display PostScript work from the
rest of the system. And while I've never seen a comprehensive comparison of
AU/X against NeXTSTEP from that era, I think NeXTSTEP would come out on
top, not least of which because of it's good development environment.

It'd be fun to set up a 68030 Sun, Mac IIfx running AU/X, and NeXT Cube up
and see how they'd all do in a variety of tasks!

On Tue, Feb 2, 2016 at 8:28 PM, John Hudak <jjhudak at gmail.com> wrote:

> As an addendum to my own post, in comparison to Sun, Apple had used the
> 680xx cpus but, in comparison, they were making 'toys'...I  think the SE30
> was a really sweet machine as it was but apple castrated it with no FPU,
> limited memory space, and adherence to apple talk (which was never going to
> power the internet). ...It was their answer to DECs PDT, effectively a word
> processing station (which was amazing at that task, just sucked as a
> general purpose machine).
> Also, I saw a lot of Suns displace VAXes and MVAXes in my university
> setting.  Too bad there isn't a similar hw/sw 'revolution' going on
> now.....those were fun days.
>
>
> On Tue, Feb 2, 2016 at 8:18 PM, John Hudak <jjhudak at gmail.com> wrote:
>
> > well, to echo a point already made, the whole package. The 68030 cpu was,
> > as a CISC architecture, very good. It was analogous to the PDP11 CPU hw
> > architecture at the time. It had an orthogonal ISA, instruction
> composition
> > and decoding 'made sense' at the assembly language level, and Mot has a
> > very nice set of complimentary support chips that could make a 680xx
> > machine with a low chip count.  Sun was in transition from a lot of 74xx
> > 'glue logic' to PALs and even custom ASICs....IIRC the MMU for that
> machine
> > was Suns own design heavily influenced by Unix. It was clean (not
> anywhere
> > as convoluted as DECs MMU approach with the VAX, and was relatively fast.
> > Bill Joy did wonders in crafting UNIX for that machine.  I always thought
> > that Sun lost the real prize in dragging their feel (e.g. not committing
> > resources) to promoting a *really good* version of Unix and the
> networking
> > support- their slogan at one time was "The network is the computer
> > (machine?)" I believe if Sun dumped more resources into the OS and
> network
> > marketing, there would have been no Linux.
> > Sun had an excellent network view of the world and Unix, much better than
> > DECs view and definitely ahead of DEC in the UNIX world, but they ran
> into
> > the same enemy - the PEEE CEEE. - cheaper, a lot less powerful, full of
> > design holes and a very bad imitation of CMP and taught the world the
> value
> > of the three-finger salute and that print queues would be invented by MS
> > some 20 yrs later. (eventhough DEC had be doing multi tasking since mid
> > 1960s and it was a built in standard with Unix almost since its
> > inception-definitely since V6. Oh well a digression..
> > The SPARC architecture was Suns answer to the RISC architecture camps,
> and
> > it was a fairly good one.  In fact, the T1 and T2 hw architectures are
> > regularly cited and taught in EE/CS grad courses, and one can get the IP
> > for Xilinx FPGAs to emulate a T1, or multi-core T!s.  Anyway, a lot of
> > wonderful history.
> > Enjoy
> >
> > John
> >
> >
> > On Tue, Feb 2, 2016 at 5:19 PM, Nathan Raymond <nraymond at gmail.com>
> wrote:
> >
> >> So here I guess I show my (relative) youth, having never used a 680x0
> Sun
> >> workstation (just SPARC stuff), what's the draw, exactly? At that time I
> >> was excited by the Mac IIfx and AU/X and the NeXT computers, both of
> which
> >> struck me as advanced machines from a VLSI design perspective. From the
> >> photos I've seen of Sun 680x0 workstations, their motherboards have what
> >> look like an insane number of chips (surely not 74-series logic?) which
> at
> >> first blush looks primitive to me. What am I missing? Or is it the
> >> software
> >> they ran in that era?
> >>
> >> On Tue, Feb 2, 2016 at 5:05 PM, John Hudak <jjhudak at gmail.com> wrote:
> >>
> >> > OMG, nice snag....We had a bunch of those in our lab...brings back
> nice
> >> > memories.  Congrats and pics after bootup pls...
> >> >
> >> > Much jealous and envy...
> >> > -J
> >> >
> >> >
> >> > On Tue, Feb 2, 2016 at 1:51 PM, Walter Belgers <
> >> walter+rescue at belgers.com>
> >> > wrote:
> >> >
> >> > > Hi all,
> >> > >
> >> > > I just broke a promise to my wife (and myself). I promised to
> collect
> >> > Suns
> >> > > no
> >> > > bigger than then SparcCenter 1000. But when you come across a 3/470,
> >> how
> >> > > can
> >> > > you not rescue it from the garbage heap?
> >> > >
> >> > > It has a 501 1550 CPU board with a 501 1532 cg6 framebuffer (quite
> >> > special
> >> > > I
> >> > > guess - it also comes with keyboard, mouse and colour monitor).
> There
> >> s a
> >> > > 501
> >> > > 1217 SCSI controller, a 501 1102 RAM board (8MB) and a board marked
> >> > > DATARAM .
> >> > > Is this a memory caching board? It has a bunch of LEDs on it.
> >> > >
> >> > > The system supposedly does not boot up (haven t tried yet). The
> boards
> >> > have
> >> > > been pulled out and put back again, not necessarily in the same
> slot.
> >> Is
> >> > > there
> >> > > something I need to know about slot locations for these board? I
> >> vaguely
> >> > > remember some quirks with other Suns in this respect.
> >> > >
> >> > > Cheers,
> >> > > Walter.
> >> > > --
> >> > > Walter Belgers
> >> > > walter at belge.rs -=- http://belge.rs/
> >> > > _______________________________________________
> >> > > rescue list - http://www.sunhelp.org/mailman/listinfo/rescue
> >> > _______________________________________________
> >> > rescue list - http://www.sunhelp.org/mailman/listinfo/rescue
> >> _______________________________________________
> >> rescue list - http://www.sunhelp.org/mailman/listinfo/rescue
> _______________________________________________
> rescue list - http://www.sunhelp.org/mailman/listinfo/rescue