Re: Performance techniques

[Jon Paris <jon.paris@xxxxxxxxxxxxxx>]

Correct - there is not conversion when using integers. And they are the fastest.

MAIN will have an effect but not a hug impact I suspect unless the program frequently sets on LR etc.


On 2013-09-09, at 2:01 PM, Vern Hamberg <vhamberg@xxxxxxxxxxxxxxx> wrote:

> So the caution in that article about using BINARY type for looping indexes is not true for INTEGERs, right? Integers require no conversion to packed, if I read the docs right.
>
> In fact, the docs say that arithmetic processing is fastest using integers, with packed next.
>
> I don't remember hearing whether MAIN in the H spec can result in better performance - one might think so, as there has to be some overhead in the cycle.
>
> OK, back to working!
>
> Vern
>
> ----- Original Message -----
> If the binary (as opposed to integer) data type is used Vern it still gets converted to packed before any operations occur. But in those days of course there was no int.
>
> DS I/O can have a significant impact in batch runs with large numbers of reads. The higher the %age of fields actually used the bigger should be the impact. You could also add NOUNREF to the list.
>
> The other potentially big change since that was written (apart from SQL) is that you can now performance tune an app based on collected usage data.
>
> On 2013-09-09, at 11:48 AM, Vern Hamberg <vhamberg@xxxxxxxxxxxxxxx> wrote:
>
> > No danger there, Rob - no mention of Input Primary in the article.
> >
> > The article IS from 1993, so it won't have the newest stuff. I have a feeling that the techniques classified as having a large impact would not have so much of one today, with the new hardware. Nonetheless, some things still seem useful.
> >
> > There are also things like binary vs. packed - those distinctions might change completely now - at one time, everything, it seems, was converted to packed for calculations.
> >
> > Oh yeah, nothing about integer data types - only binary.
> >
> > And now we have direct IO using data structures, which could be significant.
> >
> > Hey, the writer may have had only a D02, and he/she did not have subprocedures, apparently - no mention.
> >
> > ----- Original Message -----
> > This list is for general operating system type questions and whatnot.  For 
> > RPG specific questions please take that to the RPG400 list.
> > http://lists.midrange.com/
> >
> > I am sure that the members on this list sure don't want to hear another 
> > debate on the pros/cons of Input Primary files.
> >
> >
> > Rob Berendt
> > -- 
> > IBM Certified System Administrator - IBM i 6.1
> > Group Dekko
> > Dept 1600
> > Mail to:  2505 Dekko Drive
> >         Garrett, IN 46738
> > Ship to:  Dock 108
> >         6928N 400E
> >         Kendallville, IN 46755
> > http://www.dekko.com
> >
> >
> >
> >
> >
> > From:   John Mathew <johnmathew400@xxxxxxxxx>
> > To:     Midrange Systems Technical Discussion <midrange-l@xxxxxxxxxxxx>, 
> > Date:   09/09/2013 09:10 AM
> > Subject:        Performance techniques
> > Sent by:        midrange-l-bounces@xxxxxxxxxxxx
> >
> >
> >
> >
> >
> > Hi,
> >
> > Can some please explain what performance steps generally follow while 
> > coding in RPG400 and RPG IV.
> > I had to do some performance tuning on the existing programs.
> >
> > I did Google and found some links related to performance as below.
> >
> > http://www.mcpressonline.com/programming/rpg/the-truth-about-rpg-performance-coding-techniques.html
> >
> >
> >
> > John
> > -- 
> > This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing 
> > list
> > To post a message email: MIDRANGE-L@xxxxxxxxxxxx
> > To subscribe, unsubscribe, or change list options,
> > visit: http://lists.midrange.com/mailman/listinfo/midrange-l
> > or email: MIDRANGE-L-request@xxxxxxxxxxxx
> > Before posting, please take a moment to review the archives
> > at http://archive.midrange.com/midrange-l.
> >
> >
> > -- 
> > This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
> > To post a message email: MIDRANGE-L@xxxxxxxxxxxx
> > To subscribe, unsubscribe, or change list options,
> > visit: http://lists.midrange.com/mailman/listinfo/midrange-l
> > or email: MIDRANGE-L-request@xxxxxxxxxxxx
> > Before posting, please take a moment to review the archives
> > at http://archive.midrange.com/midrange-l.
> >
> > -- 
> > This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
> > To post a message email: MIDRANGE-L@xxxxxxxxxxxx
> > To subscribe, unsubscribe, or change list options,
> > visit: http://lists.midrange.com/mailman/listinfo/midrange-l
> > or email: MIDRANGE-L-request@xxxxxxxxxxxx
> > Before posting, please take a moment to review the archives
> > at http://archive.midrange.com/midrange-l.
>
> Jon Paris
>
> www.partner400.com
> www.SystemiDeveloper.com
>
>
>
>
> -- 
> This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
> To post a message email: MIDRANGE-L@xxxxxxxxxxxx
> To subscribe, unsubscribe, or change list options,
> visit: http://lists.midrange.com/mailman/listinfo/midrange-l
> or email: MIDRANGE-L-request@xxxxxxxxxxxx
> Before posting, please take a moment to review the archives
> at http://archive.midrange.com/midrange-l.
>
> -- 
> This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
> To post a message email: MIDRANGE-L@xxxxxxxxxxxx
> To subscribe, unsubscribe, or change list options,
> visit: http://lists.midrange.com/mailman/listinfo/midrange-l
> or email: MIDRANGE-L-request@xxxxxxxxxxxx
> Before posting, please take a moment to review the archives
> at http://archive.midrange.com/midrange-l.

Jon Paris

www.partner400.com
www.SystemiDeveloper.com