Re: parsing a numeric field (timings)

["James W. Kilgore" <qappdsn@xxxxxxxxxxxxx>]

Simon,

I was just curious about performance of the opcodes.  I ran my test on a model
200 (CISC), and this time I increased the iterations to 100,000 as you did and
got 80 seconds for the Z-ADD loop and 3 seconds for the MOVE/MOVEL loop.

I'm not sure what the results mean, but it would appear that there was a huge
performance gain in the RISC boxes in Z-ADD vs MOVE.

There is a production program that gets run each night on this machine against
about 200,000 records that does a lot of Z-ADD's and takes about 25 minutes to
complete.  Just out of curiosity, since this is on a CISC box, I'm going to
replace some of the Z-ADD's with MOVE's and see what difference it makes.  I'll
report the results.

Simon Coulter wrote:

> <<snip>>
>
> Without wishing to enter yet another pissing contest I ran your code
> unmodified on my 170-2291 and the entire program finshed in less than a
> second thus STARTTIME, MIDDLETIME, and STOPTIME were identical.
>
> I increased the loop value (and the size of COUNT -- oops!) and ran each
> loop 100,000 times resulting in 0 seconds elapsed for the first loop and 1
> second elapsed for the second loop on the first run and 0 seconds elapsed
> for both loops on each subsequent run.
>
> I ran it again for 1,000,000 iterations and got 1 second elapsed for the
> first loop and 3 seconds elapsed for the second loop.  These numbers were
> consistent across multiple runs.
>

+---
| This is the RPG/400 Mailing List!
| To submit a new message, send your mail to RPG400-L@midrange.com.
| To subscribe to this list send email to RPG400-L-SUB@midrange.com.
| To unsubscribe from this list send email to RPG400-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---