|What is MicroFocus doing? email@example.com (1994-07-21)|
|From:||firstname.lastname@example.org (Erick Rudiak)|
|Summary:||Strange MicroFocus COBOL behavior|
|Keywords:||Cobol, question, performance|
|Date:||Thu, 21 Jul 1994 15:47:09 GMT|
An interesting question for the compiler crowd:
There is a known 'anomaly' in MicroFocus COBOL (I have seen it first-hand
on a Pyramid, but it occurs on other UNIX platforms as well) where the
remainder operation performs miserably. A simple statement such as
DIVIDE I INTO N GIVING N1 REMAINDER REMAIN
performs about 80 -times- slower (no typo: 80 times) than
DIVIDE I INTO N GIVING N1
COMPUTE N2 = I * N1.
COMPUTE REMAIN = N2 - N1
Are they possibly doing repeated subtractions instead of division here?
Are there any other known 'anomalies' in MicroFocus that have surfaced, or
been discovered out there?
Enquiring minds would love to know!
Erick Rudiak, email@example.com
Return to the
Search the comp.compilers archives again.