RE: Performance of Zoned vs Int vs Packed

Dare wondered:

>Do you think resource (*proc, *mem, etc)
>contentions amongst other jobs on
>the system could be a factor in your TEST?

Yes, but not very much.  I ran the loops all in one run while the machine
was mostly idle.  I didn't run "s", re-compile and then run "p", etc.  I
have 6 loops in the program.  I compiled it 3 times; once for 50k, once for
500k and once more for 5m iterations.  The comparisons at the 5 million
iteration level are really the only ones that are useful, but they do seem
to make sense compared to my experience.

Here are some RPG III numbers:
                 5m
P   1   50PACK    8
    6  140ZONE   12
B  15  180BIN    35


H        1
I           IDS
I                                    P   1   50PACK
I                                        6  140ZONE
I                                    B  15  180BIN
I I            5000000               P  19  230LIMIT
C                     Z-ADDLIMIT     PACK
C                     TIME           START   60
C                     DO   PACK
C                     ENDDO
C                     TIME           STOP    60
C                     DUMP
etc.
  --buck