[gclist] When to collect.

Jeff Barnett jbarnett@charming.nrtc.northrop.com
Thu, 20 Nov 1997 09:23:19 PST

>Charles Fiterman wrote:
>> The problem is what I call the collect or expand
>> decision. Given a request that can't be filled
>> without expand or collect which do I do.
>> I don't like any of the current algorithms and
>> are trying to rethink from scratch.
>> First the idea is not to get by with as little
>> memory as you can but to use as much as won't
>> trigger paging. This simple program shows what
>> I mean.

I wrote a paper a long long time ago that played with
the above question:

  Barnett, J.A., Garbage collection versus swapping, in
  Operating Systems Review 13(3) (1979).

It assumed a swapping, not paging, model.  Cost models
were used to tell the programmer how big to make his
program to optimize performance.  In addition, it tried
the problem of deciding on how to split the installation
budget between buying disks and buying memory for a time
sharing machine.  (See I told you it was old.)  The
interesting result was that the optima was stable in the
sense that what was good for the installation was good
for the individual programmer.  So there was not a war
of reconfigurations.

-- Jeff Barnett