Subject: | |
From: | |
Reply To: | |
Date: | Thu, 2 Oct 1997 16:25:05 -0600 |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
I recently tried a small performance experiment in Suprtool using
the GET command with a $LOOKUP .vs. the CHAIN command. And the
winner depends upon what is more important time or CPU usage.
Please note that both of these jobs were streamed at the same
time by the same person in the same queue. Also, both of these
jobs used a TABLE (CHAIN used it as input for indexes, GET used
it in the $LOOKUP).
GET and $LOOKUP:
IN=5587824, SEL=67, OUT=21. CPU=754. Wall-Sec=974.
CHAIN:
IN=377373, SEL=67, OUT=21. CPU=171. Wall-Sec=3880.
As you can see the GET and $LOOKUP took a significant amount of
CPU resources but took very much less time. The CHAIN command
was easier on the CPU but took about 4 times longer.
Other things to point out. The index with the CHAIN command was
not the main index(unique) but a secondary search index (non
unique). I do not know how these numbers would have been
different if the index had been a main (unique) index.
Just thought someone might find this interesting....
---------------------------------------------------------------
Barry Durand "The opinions expressed in this
Systems Engineer II email are not necessarily the
Gulf South Health Plans opinions of my employer"
[log in to unmask]
---------------------------------------------------------------
|
|
|