[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[ih] vm vs. memory

OK, one more piece of *computing* archaeology, and then we can
get back to the Internet.

On 25/10/2017 13:38, Dan Cross wrote:
> On Tue, Oct 24, 2017 at 3:42 PM, Brian E Carpenter
> <brian.e.carpenter at gmail.com> wrote:
>> On 25/10/2017 01:12, Paul Vixie wrote:
>>> Joe Touch wrote:
>>> ...
>>>> IMO, they?re no more a stop-gap to networking than VM is to memory.
>>>> But we?re digressing from the original thread...
>>> that's hard to say, but i've forked the thread anyway.
>>> vm is an example of something that started as a workaround but
>> I disagree with that evaluation. It started in practice with the
>> famous "one-level storage system" paper from Manchester**, with the
>> specific goal of making a small high-speed memory look like a much
>> larger one. I don't think it was viewed as a work-around, but rather
>> as a brilliant engineering solution to the high cost of high-speed
>> memory, vastly easier to use than explicit overlays.
>> **One-Level Storage System, T. Kilburn, D. B. G. Edwards, M. J. Lanigan, F. H. Sumner, IRE Trans. Electronic Computers EC11(2), April 1962, 223-235.
>> Full disclosure: I am biased. Frank Sumner was my M.Sc. supervisor.
> The VM system in Atlas was, apparently, controversial. Rob Pike said
> that the decision to put VM (in Paul's sense) into Plan 9 resulted in
> raising his boss's ire, due to a bad experience with Atlas:
> https://marc.info/?l=9fans&m=111558822910429&w=2

Indeed, the thrashing problem is why Peter Denning's insights were so
important. Not that I remember thrashing being a real problem for Atlas
users by 1968 when I was in Manchester. The real problem was when the
operators accidentally tore your paper tape.

There's a rather confusing Sandy Fraser oral history transcript at
but although he mentions an Atlas file system, he doesn't mention
VM issues. There's a more coherent explanation in David Hartley's
article at http://www.cs.man.ac.uk/CCS/res/res44.htm, in which it's
clear that Fraser did this work on the Cambridge Titan, which was
a stripped-down Atlas with incomplete VM: "The Titan operating system
was necessarily different from the Atlas 1 supervisor which was built
around the hardware one-level store that we didn?t have. The Atlas 1
system programmers could just scatter programs and data all over memory
and let the one-level store do the memory management; whereas on Titan,
to keep the machine busy, we had to be very careful how we buffered data."

In other words, Fraser's bias against VM was based on experience with
something that wasn't VM. Duh!