You're right. It turns out to be MAP (at least I think so) that is the
killer -- the pseudo assembler listing (is that turned on by MAP?). I just
took a look at a compile that I use as a standard benchmark for our virtual
compiler (www.syspointusa.com if you're interested). I don't claim it is THE
typical COBOL program, but it came from a customer, and since you can design
a benchmark to prove anything you want, I don't worry about it and just
stick with this program. But I digress. The source code is 5489 lines. (The
LIST output is a little longer because of headings.) Here are the lines, and
the percentage of the total, for each section of the output listing.
You can clearly see that MAP (or whatever it is that turns on the
pseudo-assembler) roughly triples the size of the listing. XREF turns out to
add only about a fifth to the LIST and general this-and-that output.
Lines %age
Options 58 0.3%
List 5724 26.7%
Xref-Data 813 3.8%
Xref-Procs 355 1.7%
Xref-Pgms 12 0.1%
Map-Data 904 4.2%
Map-PGT & Lits 158 0.7%
Map-Pseudo Asm 13288 62.0%
Map-TGT 105 0.5%
Warnings & RC 14 0.1%
TOTAL 21431 100.0%
Charles
-----Original Message-----
From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf
Of Tom Schmidt
Sent: Tuesday, September 26, 2006 6:10 PM
To: [email protected]
Subject: Re: Overhead caused by LE options with (XREF, MAP, LET)
On Tue, 26 Sep 2006 15:37:01 -0700, Charles Mills wrote:
I dispute that XREF double the size of the compiler SYSPRINT listing -- it
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
