PE>> For some reason I can't get decent stats from my zip program under
 PE>> OS/2. Times taken to do a compress can take anywhere from 19 to 65
 PE>> seconds on repeat trials of the same data at the same time with
 PE>> nothing else running!!!  On a fresh system too. All data in memory
 PE>> thanks to my enormous cache.

 RS> Yeah, I noticed a similar variability too at times myself. Didnt look to
 RS> closely at it, just put it in the 'thats interesting, wonder why'
 RS> basket.

 PE>> BTW, this corresponds to the times for your zip packets to be
 PE>> compressed also.  I am back to compressing with "zip
-1", and using
 PE>> Rod's data from yesterday to do my tests, it can pack in 9 seconds or
 PE>> it can pack in 49 seconds, and it seems to be completely random which
 PE>> one it chooses. Looking at "Pulse", in either case it
goes from 0% to
 PE>> 100%, and the only thing that changes is how long it stays at 100%.
 PE>> Nothing to do with zip either.  I wrote a test program which just
 PE>> loops, and I get the same problem with it.  The test program is in the
 PE>> next message.

The above is a quote of what you sent me.  ie, the bottom bit is quotes of
what I said with nothing following.  Is that what you wanted?  BFN.

Paul

--- GoldED 2.40