And a note: Using bind variables is a necessary step but not a
sufficient step toward eliminating parse CALLS. Using bind variables
will only reduce the number of misses in the library cache ("parse count
(hard)").
If the application still puts the parse call inside a loop, it still
gets the "pars
The Parse statistics are what caught my eye. The statement is being
parsed once per execution, which is puzzling because it does use bind
variables. Of more concern, though, is the difference between CPU and
elapsed times for the parsing, indicating that there is a lot of
waiting going on. I'd s
[EMAIL PROTECTED] wrote:
>
> Oracle 8.1.7.0.0 on Win2k (4 CPU, 4GB RAM)
>
> A 3rd Party app. is experiencing very slow performance on one of our
> databases. I think I've nailed it down to slow, row-at-a-time inserts.
> The same app. performs very fast on another DB with LMTs. After switching
>
Oracle 8.1.7.0.0 on Win2k (4 CPU, 4GB RAM)
A 3rd Party app. is experiencing very slow performance on one of our
databases. I think I've nailed it down to slow, row-at-a-time inserts.
The same app. performs very fast on another DB with LMTs. After switching
the tables and indexes in the slow DB