Page Space Problems

[Ruth Milner]

In article <1991Mar17.185553.9181 at menudo.uh.edu> jet at karazm.math.uh.edu ("J. Eric Townsend") writes:
>
>I called defect(ive) support, they told me "AIX does things a little
>differently than UNIX", and that the swapper proc was supposed to be
>~6Mb on a 8Mb RAM / 32Mb paging machine.
>
>It's not a bug, it's a *feature* :-(
>
>Again, I repeat my plea for some IBM employee to come out of the woodwork
>and explain this:
>
>oiler> ps aux
>    USER   PID %CPU %MEM    SZ   RSS     TT STAT  TIME CMD
>    root     0 0.0%  78%  7592  6376      -    S  1:31 swapper

I have formed the impression (nobody told me this, though) that the SZ etc.
shown for swapper were not so much used by swapper as they were representative
of the size of the kernel. It's a paging kernel, so not all of it has to be
in memory all the time, and the size given seems to stay pretty static when
you put more memory in. That's just an impression, though.

IBM take note: the size of your kernel is a serious limitation for low-end
systems. A 320 has really respectable floating-point performance, but our
application (and probably others) needs a minimum of 32MB on an RS/6000 to 
run reasonably (performance almost doubled when we went from 16 to 32MB). With 
the price of RS/6000 memory, this really reduces its affordability as a system
for small groups at universities to run our data reduction programs.
-- 
Ruth Milner
Systems Manager                     NRAO/VLA                    Socorro NM
                            rmilner at zia.aoc.nrao.edu