How were 32-bit minis built in the 70s/80?

[allison]

On 05/11/2019 07:14 PM, Warren Toomey via cctalk wrote:
> I'm building my own 8-bit CPU from TTL chips, and this caused me to think:
> how were 32-bit minis built in the late 70s and early 80s? In particular,
> how was the ALU built? I know about the 74181 4-bit ALU, and I know (from
> reading A Soul of a New Machine) that PALs were also used.
> 
> Did companies get custom chips fabricated, or was it all off-the-shelf chips
> with a few PALs sprinkled in?
> 
> Thanks, Warren
> 

Lets see the VAX 11/780 hit the street in 78 and DG followed with theirs
soon after and of course the IBM 360 was 32bit so the number can be
fairly large.  Soul of a new machine was more romantic but it was of
early VAX era and the Eclipse was the result.

Reading the following woould be better as it compared and contrasts DEC
hardware and instill an idea of ISA design and then its hardware
implementation.  Its a good read and free!

http://www.bitsavers.org/pdf/dec/_Books/BellComputerEngineering.pdf

Building now is more based on what you have or can get not what was used
then as most were pushing for speed vs price and the available parts
were never fast enough and cost too much.

Not all were 74181 based, Thats an early 1972 part and but 1975 it was
already getting old though useful as it evolved to 74S and 74F series.
The 82s100 and 105 series were out there and even by 1980 the AMD 2900C
series was getting long in the tooth. Mask programable gate arrays were
in the 1000 and up gate level by 1980 and growing by doubles every 6
months to a year. Don't got get programmables like PAL/GAL logic.
There was a lot of designs and even inside DEC you might see several
approaches depending on what machine and the specific date.  For example
the 780, 750 and 730 used very different technology.  I will not go into
those that also went the ECL {10K, 100K, 1M families] route.

Your question has to be based on a specific date window and narrow at
that as keep in mind by 1978 16bit CPUs on silicon were a fact
(Ti9900, SBP9900, F11, T11, Nova, 8086).