RE: Question about reducing program start overhead

Plus, how much time could you possibly save by pre-opening the file? These days I usually do USROPN's after figuring out which table I want to open (not always a simple matter here!), and I still get sub-second response times; that is, the program may have to switch between files on every Enter key.

OPNDBF *may* have saved something on the B10, even an F35, but today?

Rob's right: program/application design is the key. Similar to his analogy, I took a process that was taking 3-4 hours to run; re-designed and re-programmed it - takes about 5 minutes now.

Jerry C. Adams
IBM System i Programmer/Analyst
B&W Wholesale
office: 615-995-7024
email: jerry@xxxxxxxxxxxxxxx


-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of rob@xxxxxxxxx
Sent: Wednesday, November 12, 2008 12:46 PM
To: Midrange Systems Technical Discussion
Subject: Re: Question about reducing program start overhead

Programming for performance is hard to do with simple gut checks like you
suggested.

I chopped an hour or so off of one update and got it down to just a few
minutes by significantly reducing the number of updates. There were
several "many" relations in the "from" file for each of the "one"
relations in the "to" file. So instead of updating the to file for each
read of the from file, I sorted the from file by the key of the to file
and only updated at control break time.
So, when one suggests something without the audience understanding the
full problem I don't think we can justly answer that.

Rob Berendt