View previous topic :: View next topic |
Author |
Message |
archana Beginner
Joined: 28 May 2007 Posts: 4 Topics: 2
|
Posted: Wed Sep 17, 2008 12:30 am Post subject: Performance of a JCL PROC |
|
|
Hi All,
Can anybody tell me what is the advantage in using a Cataloged PROC instead of the utility with respect to performance of the JCL execution. |
|
Back to top |
|
|
expat Intermediate
Joined: 01 Mar 2007 Posts: 475 Topics: 9 Location: Welsh Wales
|
Posted: Wed Sep 17, 2008 2:00 am Post subject: |
|
|
PROCs are good for JCL code that needs to be used in several jobs, and thus only needs to be changed once for the changes to be reflected in all jobs that use the proc.
Some shops insist that every job that is submitted must be a proc, and some shops insist that PROCs are never used.
No discernible performance enhancements or degredations from using PROCs _________________ If it's true that we are here to help others,
then what exactly are the others here for ? |
|
Back to top |
|
|
archana Beginner
Joined: 28 May 2007 Posts: 4 Topics: 2
|
Posted: Wed Sep 17, 2008 4:18 am Post subject: |
|
|
Thanks Expat,
But as part of performance tuning we are changing the PGM=SORT to cataloged procedure and using the PROC in the JCL. So, I thought might be there will be some improvement in the performance. |
|
Back to top |
|
|
Nic Clouston Advanced
Joined: 01 Feb 2007 Posts: 1075 Topics: 7 Location: At Home
|
Posted: Wed Sep 17, 2008 4:34 am Post subject: |
|
|
There should be no difference if they are coded the same - everything is resolved as the job gets passed to the initiator - not during execution. _________________ Utility and Program control cards are NOT, repeat NOT, JCL. |
|
Back to top |
|
|
kctechpro Beginner
Joined: 27 Aug 2008 Posts: 6 Topics: 1
|
Posted: Wed Sep 17, 2008 2:02 pm Post subject: |
|
|
I'm overly detailed, so here's my two cents.
There will be a extremely minor increase in the run time. This involves the fact that JES has to expand the PROC into the JCL. If parms are added again, there will be an extremely minor increase. However, I agree there should be no decernable difference. |
|
Back to top |
|
|
Nic Clouston Advanced
Joined: 01 Feb 2007 Posts: 1075 Topics: 7 Location: At Home
|
Posted: Thu Sep 18, 2008 12:26 pm Post subject: |
|
|
There should be no increase in run time - but maybe a millisecond or two between the job coming off the reader and onto the Q where it may wait anyway. _________________ Utility and Program control cards are NOT, repeat NOT, JCL. |
|
Back to top |
|
|
jsharon1248 Intermediate
Joined: 08 Aug 2007 Posts: 291 Topics: 2 Location: Chicago
|
Posted: Thu Sep 18, 2008 3:24 pm Post subject: |
|
|
Let's forget about the performance difference for a minute because to the human observer, there is none. But I still want to add my 2 cents. Why would you want to put your SORT into a dedicated SORT PROC? I've seen a number of companies code a proc for their SORT for 'consistency', but I really don't understand it. They code symbolic parms for the dataset names and SORTOUT SPACE and so on. Some developers use the symbolics and others use DD overrides when they EXEC the PROC, and some are creative and use both. I don't see any significant advantage to coding a dedicated SORT PROC. I think coding the JCL for a SORT step is not all that complicated and I prefer to code it myself.
Ok. Back to work.
|
|
Back to top |
|
|
|
|