Performance Questions and a little need for some education.

John Stires jstires at EBDS.com
Tue Aug 18 12:10:13 CDT 2009


Normally, I would be at the same conclusion you all offer.  Squeezing out he last bit of time and resources can be a huge waste.

That is precisely the problem here.  Their nightly jobs run 12 to 14 hours and are starting run into the users coming in  the morning to work.  It can be an unpleasant picture.  There are so many QTP programs involved.  One just one such job, I counted well over 65 QTP programs, most of which 1,000, to 3,500 or more lines of code.  It is not so much the number of lines of code, but the inefficiencies involved.   They, nearly, all have very similar issues:


1)      Generally, deeply cascading DEFINEs

2)      Linking as many as 20 files together in one pass and then sorting  (Boo Hiss!) Oh yes, including several outputs and subfiles all in the same request.

3)      Sorting in general in QTP

4)      Hitting the same file several times for the same data instead of gathering the data once, putting into a subfile for later use.

There are other issues as well, but these are the main ones.  When you are bumping up against time constraints like this, you do what you have to do and seek out all of the possible ways to shave time and resources.  There are a lots of techniques they have used for a long time and therefore must be  "good and sound".  I am trying desperately to move their thinking into ways that they will be receptive to in making these processes more manageable, both from the maintenance and execution time stand point.

All of which is say, marginal gains are, indeed successes.

Thanks,
John
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.sowder.com/pipermail/powerh-l/attachments/20090818/150ab0c9/attachment.htm 


More information about the powerh-l mailing list