RE: Date formats (was Frank Doherty will be back on Tuesday (all being well))

I did run a test on a dedicated S/34 once comparing the multiplication method to the DS method. Each program looped, I think, about 10,000 times. The DS method was faster, though I don't recall the timings now. For an interactive program the difference was minimal (i.e., unnoticeable). The real kicker came in a batch program that read thousands (millions?) of records.

Jerry C. Adams
IBM System i Programmer/Analyst
--
B&W Wholesale
office: 615-995-7024
email: jerry@xxxxxxxxxxxxxxx


-----Original Message-----
From: rpg400-l-bounces@xxxxxxxxxxxx [mailto:rpg400-l-bounces@xxxxxxxxxxxx] On Behalf Of Raul A. Jager W.
Sent: Monday, December 14, 2009 1:52 PM
To: RPG programming on the IBM i / System i
Subject: Re: Date formats (was Frank Doherty will be back on Tuesday (all being well))

Old computers didn't even detect the overflow, so in the old time there
was no overhead.

Mark S. Waterbury wrote:

This "trick" of using MULT to transform from YYMMDD to MMDDYY or vice
versa was debated "to death" back around 1998-1999, as many of our
fellow midrangers had to deal with this during their Y2K conversions of
legacy code.

The main thing that makes it perform so poorly is that it causes
arithmetic overflow, which generates an exception that must be handled
by the RPG generated code and/or runtime routines (to silently ignore
this error).

Ugh.