Re: Odd, inconsistent rounding error

On Jul 27, 2023, at 7:10 PM, James H. H. Lampert via RPG400-L <rpg400-l@xxxxxxxxxxxxxxxxxx> wrote:

I've got some strange rounding behavior:

Given the following fields:

DIQTSH is zoned 15.5, in a file
DIPRIC is zoned 12.5, in a file
K01108 is zoned 12.5, defined in a standalone D-spec
U01107 is zoned 15.5, in a file
U01108 is zoned 12.5, in a file
U01168 is zoned 11.2, in a file
AMOUNT is packed 15.5, defined in a C-spec


The existing program has:
eval u01108 = u01108 * u01107
eval U01168 = U01168 + U01108

while my program has:
DIQTSH MULT K01108 AMOUNT
and either
ADD AMOUNT U01168
or
EVAL U01168 = U01168 + AMOUNT

K01108 and U01108 both come from DIPRIC. U01107 comes from DIQTSH.

My code should be equivalent to the exiting code, and yet in 12 out of 30669 cases, U01168 is off by a few cents between the two.

Anybody else have any insights into this? I tried replacing the ADD with the EVAL, with no effect on the output.

--
JHHL
--
This is the RPG programming on IBM i (RPG400-L) mailing list
To post a message email: RPG400-L@xxxxxxxxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: https://lists.midrange.com/mailman/listinfo/rpg400-l
or email: RPG400-L-request@xxxxxxxxxxxxxxxxxx
Before posting, please take a moment to review the archives
at https://archive.midrange.com/rpg400-l.

Please contact support@xxxxxxxxxxxxxxxxxxxx for any subscription related questions.