I still don't think this speculation is productive. The compilers exist.
Rather than guessing about a hypothetical performance issue, just run a
reasonably careful, straightforward benchmark and report actual results.
This is a fully, easily testable hypothesis. To my knowledge IBM has no
evidence in support of this hypothesis -- indeed, evidence to the contrary
-- but also to my knowledge IBM is *delighted* to receive such feedback
from customers.

>The COBOL Version 5.2 Programmers Guide states on page
>660 that NUMPROC(NOPFD) is less efficient than NUMPROC(PFD).

I assume that statement is correct. Please note carefully it's a statement
entirely within the scope of Enterprise COBOL Version 5.2, and it does not
quantify that difference. (It's quite possible the efficiency gap is
smaller in the new compiler than in the old compiler.) That statement also
does not in any way support a hypothesis that NUMPROC(NOPFD) in Enterprise
COBOL Version 5.2 is less efficient than NUMPROG(MIG) in Enterprise COBOL
Version 4.2. But even assuming a hypothetical NUMPROC(MIG) were technically
possible to implement in the new compiler (it may not be!), the Guide's
statement also does not mean that a hypothetical NUMPROC(MIG) would
necessarily yield any performance benefit in the new compiler. It's a new
compiler, with new technologies and new optimizations for systems with far
more advanced hardware capabilities. It's a different world.

I really, really don't think it's a radical notion to spot check actual
performance results before even getting mildly concerned about one compiler
parameter's retirement. Evidence matters. (So does absence of evidence,
especially over time.)

--------------------------------------------------------------------------------------------------------
Timothy Sipples
IT Architect Executive, Industry Solutions, IBM z Systems, AP/GCG/MEA
E-Mail: [email protected]

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to