Archives of the TeradataForum
Message Posted: Wed, 08 Mar 2007 @ 00:08:07 GMT
| Subj: || || CPU sec variations |
| From: || || Anomy Anom |
<-- Anonymously Posted: Wednesday, March 07, 2007 18:55 -->
We are having performance problems with a particular group of users that require a sub-one minute response time. Their queries are running
from 1 minute to 8 minutes. The odd thing we have found in our testing is that when we have the same user, run the exact same query, against the
same data, on the same system, each execution is showing different CPU sec consumption. I would understand the wallclock changing based on system
being busy but we thought the CPU Sec should remain the same every time the query is run. Can anyone explain why this might happen?
Thanks for any incite.