 |
 |
Archives of the TeradataForum
Message Posted: Wed, 25 Jan 2012 @ 15:08:09 GMT
Subj: | | Re: Performance Metrics: What explains this kind of observation? |
|
From: | | Gorner, Tomas |
Anomy,
You don't give a lot of information, so just a blind shot - consider skew (both disk and CPU) of your tuned query. You can have an overall
reduction in the metrics, but what a hack, if all the work is done by 1 AMP?
As for your second question - it depends. The execution time of a skewed query may change dramatically depending on overall workload. So, a
query that consumes more CPU/IO/spool, but is balanced, would be probably more steady when it comes to a load/execution_time ratio.
Regards,
Tomas
| |