RE: SQL vs. traditional I/O?

["Joe Pluta" <joepluta@xxxxxxxxxxxxxxxxx>]

> From: Vern Hamberg
> 
> In the case of your tests that use SELECT INTO, it may be that the SQL
> engine is still doing some open/close activity. It'd be useful to run
the
> DBMON against your program, to see what is really going on with SQL.
This
> run would not be useful for performance comparisons, only to gather
> information on how it is running. Oh, and, before using DBMON, maybe
> STRDBG
> would give a hint of things.

I did do STRDBG one run, and it told me that SQL was using an already
open data path.  My guess is this is about as good as it can get.

Thanks for the hints on CLRPOOL though; I'll be sure to use that when I
do "from scratch" tests.  It will provide a different but just as valid
baseline.

And remember folks, don't expect me to be writing 150 tests scenarios.
I WILL start to design a generic benchmark framework (I actually started
one a while back), but I'm going to need help writing the actual tests.


> And it is possible with QAQQINI settings to
> get even more information in the job log. Most of those settings are
> documented and available in InfoCenter and probably somewhere at
> <www.iseries.ibm.com/db2>.

Perhaps they are.  However, since I'm not an SQL expert, and I'm not
advocating SQL for all things, I think I'll let the SQL evangelists do
their thing.  And just think, that knowledge could become a FAQ.  If one
of the things the IAAI ends up with is a FAQ for SQL settings, that
would probably be worth the effort right there.


Joe