RE: Bifs vs Indicators

Jack,

This is not a qualified answer to your question, but the way you tested this 
will not produce conclusive results either way.  One of the problems with just 
simply running a job to record its start and end times is that you have many 
environmental influences to account for.....  Think transitional workloads, 
such as web serving, email, batch jobs, and so forth.  Also in the mix is 
memory management, where the file and program objects are paged into main 
storage for the first run, but not for subsequent runs....

It's probably better to use a performance tool like PEX (or whatever it's 
called now) to record and analyze the performance characteristics of your app.  
This should give you a much better way to understand the performance 
considerations that need to be reviewed.

I doubt that you will see any measurable difference between the indicator and 
bif versions of your app...

Good luck....

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 derhamj
Sent: Sunday, September 17, 2006 8:29 AM
To: rpg400-l@xxxxxxxxxxxx
Subject: Bifs vs Indicators


Hi List,

 

Just to check timings I ran a small test to see if the use of %EOF bif 

with a read operation would be faster or slower then the use of the classic 

indiactors.  I expected a nice clear black hat/white hat answer but alas the


V5R2 system that I ran on had to surprise me. When run as a batch job, the
Bifs

over an 8 million record file won hands down across various times an loads
of 

the work day. However when the same code was was run against the same file 

interactively, the indicators won.  Don't understand. Could someone explain
this

to me. Time Slice for both batch and interactive was 500. Run priority for 

Interactive was 20 and for Batch 50. Batch DWT was 120 and 30 for
Interactive.

 

I will continue to use the Bif syntax but it sure seems strange that the
same code 

Would provide different results depending on the type of job.

 

Jack Derham