Re: Is the Program Status Data Structure Externally Defined?

Barbara,

After this discussion started, just to check, I created a program that
simply turned on *INLR and returned, just like yours. In my test, I called
it 1000 times with a full PSDS and 1000 times with an 80-byte PSDS.

No significant difference that I could see - times were within 1-2 seconds
over 1000 calls.

Glad to know I've not been adding needless performance hits to my programs
all this time, without even writing code :)

Rory

On Wed, Dec 8, 2010 at 6:06 PM, Barbara Morris <bmorris@xxxxxxxxxx> wrote:

On 2010/12/6 9:06 PM, Rory Hewitt wrote:

...
I typically code new RPG programs from a template which includes the

above

copybook, and I've never noticed any performance hit. I assume the hit

would

simply be at startup, where it has to call e.g. MATPGMNM or its

equivalent

to get the program library etc. ...

Rory, you're right that the PSDS hit is only at startup. When that
section of the manual was written back in V3R1, there was a noticeable
difference in performance for repeated calls to a program with a maximum
sized PSDS vs one with an 80-byte PSDS, where both programs set on LR so
the initialization cost was incurred on every call.

The various APIs that the RPG runtime uses to get the information may
have got faster over time, possibly even during the CISC -> RISC change
in V3R2.

Our tests in V3R1 used programs that did nothing except set on LR. Even
then, if the programs actually did a reasonable amount of work, the
difference due to PSDS initialization was not noticeable. (So arguably,
we could have just left that information out of the manual ...)
--
This is the RPG programming on the IBM i / System i (RPG400-L) mailing list
To post a message email: RPG400-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/rpg400-l
or email: RPG400-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/rpg400-l.