RE: Usage of Indexes and its Effects on an RPG Program

Rich,

in order to perform these tests correctly, you should also account for the
differences between runs where the data is already paged into main storage.
Typically, you'd need to CLRPOOL to remove objects before subsequent runs.

Eric DeLong
Sally Beauty Company
MIS-Project Manager (BSG)
940-297-2863 or ext. 1863



-----Original Message-----
From: rpg400-l-bounces@xxxxxxxxxxxx
[mailto:rpg400-l-bounces@xxxxxxxxxxxx]On Behalf Of Rich Duzenbury
Sent: Wednesday, December 14, 2005 5:30 PM
To: RPG programming on the AS400 / iSeries
Subject: RE: Usage of Indexes and its Effects on an RPG Program


On Wed, 2005-12-14 at 16:05 -0600, Scott Klement wrote:
> > Therefore, I wrote two programs to test this:
> >
> > read_test
> > http://code.midrange.com/index.php?id=43c4fd6a02
> 
> Try changing READ_TEST to specify BLOCK(*YES) on the F-spec.
> 
> 
> > reade_test
> > http://code.midrange.com/index.php?id=e3203a5c0b
> 
> (Note that if you specify BLOCK(*YES) on the F-spec for READE_TEST it'll 
> give you a RNF7092 (severity 10) error in the compile listing. Look at the

> text of that message for details)
> 
> Does this change your benchmark scores appreciably?

Yes, on average, READ gets worse.  The average for READ in the latest
tests was 17 secs, and the READE average improved slightly to 15.8
secs.  

However, if I throw out the best and worst for both, READ has a bit of
an advantage, 14 secs vs 16 secs or about 12.5 %.

So, _if_ block(*yes) is specified (and doesn't RNF7092 because of other
parts of the code) _and_ reade is changed to read with a condition, you
may see some benefit, or maybe not, I suppose depending on how organized
the file is.

But certainly not to the level of 'READE is a dog!  Change it to READ +
condition for a big performance boost'.

Regards,
Rich