[Info-vax] Raxco VMS Tuning Seminar Notes

[Bob Koehler]

In article <0vlgd6pi1pco95ufholndt2kurg5oujfuh at 4ax.com>, jls <notvalid at yahoo.com> writes:
> 
> And if you're using other mechanisms to reclaim memory, why use
> PFRATL?  The interaction between PFRATL and PFRATH almost insures that
> active processes will thrash pages in and out.  That does not help
> performance.

   Paging and swapping can be performance improvers if they happen to
   inactive processes that are using up RAM without need, causing active
   processes to page or swap needlessly.

   Tuning is the gentle art of determining at what level you get enough
   paging or swapping of idle processes so free up RAM for active
   processes without causing excessive paging or swapping of active
   processes.  In all cases I've dealt with, one never wants active
   processes to swap, but a very small page fault rate because of the
   affects of the tuning parameters on the active processes is much better
   than a large page fault rate because of idle processes consuming RAM.

   And of course, "buy more memory" avoids all of this.