NP> I would suggest that this difference between how virtaul memory
 NP> is handled between temp & permant swap file, could also treat
 NP> other situations in an illogical manner, and cause unnecessary
 NP> disc activity. I only use a temp swapfile, and keep my cache at
 NP> 750k, for the above reasons. 

  I've tried temporary and permanent swap files, and various sizes of
cache. It seems to me that the fastest results are a 2Mb cache under
Windows (plus 4Mb of "real" RAM, and a permanent swap file. Running
the compiler, it's *much* faster if I get it all into cache. I reduced
the sice of my RAM drive to get another meg into cache.

  The occasional swap file "dreaming" doesn't worry me. It only does
it when I'm not doing anything anyway.

 NP> On all machines I have setup, the _only_ time there is disc
 NP> activity, is when a file is actually read or written, which is
 NP> how it _should_ be. 

  The fastest overall result is how is *should* be. I'm tempted to buy
another 16Mb of RAM.

Regards,
Bob
___ Blue Wave/QWK v2.12
@EOT:

---