RE: Connection Pooling - Activation Groups

Scott,

Just to be more clear, the program doesn't run faster, it loads and starts
running in a shorter timeframe than it did previously because no activation
group is being created. 

Just didn't want Jason to think the program runs faster the second time. It runs
the same, although the start-to-end delta gives one the impression that the
program ran faster, it was just the removal of the steps in the path that allow
it to start up quicker.

-Bob Cozzi
www.iSeriesTV.com
Ask your Manager to watch iSeriesTV.com

-----Original Message-----
From: rpg400-l-bounces@xxxxxxxxxxxx [mailto:rpg400-l-bounces@xxxxxxxxxxxx] On
Behalf Of Scott Klement
Sent: Tuesday, September 05, 2006 3:58 PM
To: jbender@xxxxxxxxxxxxxxxxx; rpg400-l@xxxxxxxxxxxx
Subject: Re: Connection Pooling - Activation Groups


Hi Jason,

I'm sending a copy of my reply back to the RPG400-L mailing list, since 
it's where this thread started, and I feel it's important for follow-ups 
to go there as well. That way, everyone has a chance to learn, and 
everyone has a chance to make comments.

In keeping your activation group open, lets say the program is called
again, what happens with the activation group? Does it get closed or will
it try to reclaim the activation group?

ILE programs are loaded into memory (the technical term for that is called 
"activation") into an activation group.  The program then remains in 
memory until the activation group is reclaimed (i.e. deleted from memory.)

Let's say you compile your program with ACTGRP(JASON).  The first time 
this program (or any other program with the same activation group name) is 
called, the activation group gets created in memory. Your program is then 
loaded from disk into memory into this activation group. It's then run.

On subsequent calls (assuming you haven't reclaimed the activation group) 
the program is already loaded into memory, and therefore runs very fast 
because it doesn't have to be re-loaded into memory.

If you're writing an RPG program, and the program ended with *INLR off, 
the files can be left open from call to call, which speeds things up even 
further since the files don't have to be closed and re-opened on every 
call.