Re: This is a software design question - ILE related

[Paul Tuohy <tuohyp@xxxxxxxxxxxxx>]

Nathan,

You may also want to consider having the zip code file duplicated in a User
Space.
The User Space would be re-built whenever a record is add/updated or deleted on
the zip code file.
Your server program would map an array onto the user space, which would save it
having to load the array from the file and therefore save one ODP per job.

HTH

Paul Tuohy

"Nathan M. Andelin" wrote:

> I wrote some code to test the performance of some of the suggestions you
> folks have given me.
>
> Since James Kilgore suggested putting the entire file in memory, I decided
> to add that option to the mix.  I wrote a service program that exported 2
> procedures:
>
> getFromFile() - uses CHAIN to retrieve from file
> getFromArray() - uses LOOKUP to retrieve from an array
>
> The service program pre-loads a couple of arrays.  The first holds the file
> key.  The second holds the return value.  The file also has only 2 fields
> defined - the key and the lookup value.  Both file and array only held 256
> records/entries.
>
> I wrote a little program to call the procedures repetitively to test their
> speed.  Each procedure was called 76,800 times.  Following are the results.
>
> getFromFile() - 29 seconds
> getFromArray() - 10 seconds
>
> To test the speed of data queues I added a getFromQueue() procedure to the
> service program and also wrote a "server" to receive a request from a queue,
> then do an array lookup (since lookup was faster than chain), then put the
> return value to a response queue.  Both queues were non-keyed.
>
> The speed of getFromQueue() was 98 seconds.

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