On 1/22/2020 9:00 AM, Mary Ann Horton wrote:
Even before Zilog's ZEUS, Onyx came out with a
microwave oven-sized box
based on the Z8000. They loaned one to Berkeley, and it was my first
home computer when I took it home to port the tools.
As someone who played a very minor role at UC Berkeley, I thought
I'd amplify this a little. It was amazing how the various vendors
would happily lend (or give) equipment to UC Berkeley. I don't
know how things were at the other top CS schools, but at Berkeley
I found that if I needed something, telling my faculty boss/mentor
would often make the item appear.
For example, I did an early port of Postgres to Windows NT using
a DEC PC with 16MB of RAM. With such a small amount of RAM this
was excruciatingly slow but I made enough progress to show Mike
Stonebraker. He then talked to his friends in industry and all of
a sudden a MIPS(!) box with 64MB of RAM showed up. I was then able
to get the port to the point where it could run the Wisconsin benchmark.
Something similar happened when the group wanted to port an early
version of Postgres95 to Solaris (it had been developed on DEC Alphas
running OSF/1). Mike said to write up a proposal to Sun that he'd sign
and then send in, which I did. Soon after a couple of Sparcstation 10s
(or 20, I don't remember) showed up and the port was done.
There other systems research groups at UCB in the 90s also were
the happy recipients of all kinds of great hardware. The limiting factor
for such donations was usually finding space to house them. The end
result was that the RAID, Risc, and Postgres groups weren't slowed down
by lack of hardware.
I have no idea what it's like today. Eric Allman, would you care to
comment?
Jon Forrest