[Info-vax] Reading Gordon Bell's VAX strategy document

[Johnny Billquist]

On 2023-09-24 16:10, John Dallman wrote:
> Gordon Bell, who was Vice-President of Engineering at DEC 1972-83 is
> still alive and documenting much of his life on the web. There's DEC
> stuff at https://gordonbell.azurewebsites.net/Digital/DECMuseum.htm
> 
> Something particularly interesting is this document on DEC strategy as of
> 1979:
> 
> https://gordonbell.azurewebsites.net/Digital/VAX%20Strategy%20c1979.pdf
> 
> At the time, DEC's other active product ranges were PDP-8, DEC-10/DEC-20
> and PDP-11. They had decided in 1975 to create an architecture that built
> upwards from the PDP-11, rather than building lower-cost DEC-10 machines.
> The reasons for doing that were the large installed base of PDP-11s and
> the convenience of 8-bit bytes for data communications, especially with
> IBM mainframes.
> 
> As of 1978/70 they had achieved this and were deciding what to do next.
> The strategy expressed in this document is to continue to sell the other
> ranges, but concentrate development efforts on the VAX family, and that's
> what basically happened. Using a single architecture is seen as a
> competitive advantage against IBM's proliferation of incompatible
> architectures, which is pretty reasonable, since IBM saw the same problem.
> 
> 
> Bell regards competition from "zero cost" microprocessors such as the
> 8086 and 68000 as likely more significant than other minicomputer
> companies, but fails to make a plan to deal with them. DEC was eventually
> defeated by 80386 and later PCs and RISC workstations, and that failure
> seems to start here. He assumes that DEC can dominate the market for
> terminals for its minis by using PDP-11 and VAX microprocessors, but
> doesn't seem to realise that compatible terminals can be built at much
> lower cost using third-party microprocessors. In any case, the
> replacement of minis by PCs and workstations meant that the terminal
> market basically vanished.
> 
> The idea of running VMS on a terminal with a total of 64KB of RAM and ROM
> in 1982 seems implausible now, but it seems to have been the reason for
> 512-byte pages. Bell praises the extremely compact VAX instruction set
> and its elaborate function calls, without appreciating the ways they will
> come to inhibit pipelining and out-of-order execution, and thus doom the
> architecture to uncompetitive performance.

It's always easy to see mistakes after the fact.
When the VAX was designed, as well as around 1980, memory was still 
rather expensive. An instruction set that lead to smaller binaries was a 
big win at that point in time.

What you could possibly argue was that DEC didn't enough see or 
anticipate the drop in price of memory, which would lead to totally 
different constraints and optimal points.

VMS was never expected to run on something with 64K. You couldn't even 
run a reasonable PDP-11 on that little memory at that point. (I said 
responable, for anyone dragging out a minimal RT-11 system.)

But VAX was most definitely designed for getting programs more memory 
efficient. More addressing modes, more things done in microcode to deal 
with things in a single instruction. Very variable length 
instructions... All was about memory cost. Which made a lot of sense 
between 1970 and 1985. After that, memory was becoming so cheap there 
was no reason for the optimization angle the VAX had taken. And you had 
the rise of RISC.

   Johnny