On 8 Oct 2009 15:32:09 -0700, in bit.listserv.ibm-main you wrote: >I am a COBOL person not an Assembler person. Don't the "grande" >instructions require a specific "architecture level set"? If so, that might >be why (as others in the thread have indicated), COBOL does NOT do what you >are asking about.
Since Enterprise COBOL requires z/OS 1.9, it requires the grande ALS. > >It would seem a "reasonable" SHARE requirement for something like > - when using *ALL* COMP-5 sending and receiving fields, (and possibly also >when using all binary fields with TRUNC(OPT)) then binary, not >packed-decimal arithmetic should be used. When such arithmetic is using >"large" binary fields, then "grande" instructions should be used." > >This assumes, however, that this would actually provide a demonstrable >advantage to IBM COBOL customers. The grande instruction set usage can be useful for even S(9) BINARY since it would be a simpler way to handle overflow with multiple operands. > >"Farley, Peter x23353" <[email protected]> wrote in message >news:<053f2631ec9c584883847c8b4970a228050da...@josqems1.jsq.bsg.ad.adp.com>. >.. >> I did not see anything about it in the updates page in the language >> reference nor the programmer's guide, but I'm wondering if anyone here >> knows if the newest release of the Enterprise COBOL compiler will >> generate "grande" arithmetic instructions for COBOL binary fields (e.g. >> will it generate an AG or AGR instruction for adding two PIC S9(18) >> BINARY fields?). >> >> TIA for any info you can provide. >> >> Peter >> > ---------------------------------------------------------------------- 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
