Method R Discussion GroupMethod R helps developers, DBAs, and decision-makers optimize Oracle-based software in every phase of the software life cycle. |
||
Discuss Method R products, techniques, and events. Ask new questions or post your own responses. |
When running a similar query repeatedly, do you flush the buffer cache each time?
I don't. If you flush the cache each time, then you're always testing a "worst case" scenario...i.e. you're always saying that the data blocks needed for this query will *not* be in the buffer cache and must be physically retrieved. This should not be the case in your production environment - at least you would hope not! - and so you should test typical behavior which would be to have your tests be able to find previously accessed blocks already in the buffer cache.
I'd just note that you can easily test the worst case by flushing the cache (both the buffer cache and the shared pool too) initially, but then all subsequent tests should be executed without flushing anything. |
Powered by FogBugz