I was just stating what a theoretical new version of a theoretical compiler 
might provide.  In the specific case of E.C. 6.3 you are certainly correct.  Of 
course there is no requirement to change ARCH level upon implementation of a 
new version, even if you're at that necessary hardware level.

________________________________
From: IBM Mainframe Discussion List <IBM-MAIN@LISTSERV.UA.EDU> on behalf of 
Clark Morris <cfmt...@uniserve.com>
Sent: Tuesday, January 14, 2020 4:20 PM
To: IBM-MAIN@LISTSERV.UA.EDU <IBM-MAIN@LISTSERV.UA.EDU>
Subject: Re: Migrating to new compiler release

[Default] On 14 Jan 2020 13:58:34 -0800, in bit.listserv.ibm-main
frank.swarbr...@outlook.com (Frank Swarbrick) wrote:

>I disagree with that.  A new version may simply be new features, none 
>affecting any features the old version supported.

The 6.3 compiler supports a new ARCH level for the z15 and uses SIMD
instructions rather that the selected conversion to floating point
decimal for some calculations.  Both code sequences are an enhancement
over using standard packed decimal instructions.

Clark Morris
>
>________________________________
>From: IBM Mainframe Discussion List <IBM-MAIN@LISTSERV.UA.EDU> on behalf of 
>Charles Mills <charl...@mcn.org>
>Sent: Tuesday, January 14, 2020 2:09 PM
>To: IBM-MAIN@LISTSERV.UA.EDU <IBM-MAIN@LISTSERV.UA.EDU>
>Subject: Re: Migrating to new compiler release
>
>> My question has to do with the (probably slight) possibility that the code
>generated by one compiler would be different, for the same statement, for
>another.
>
>It certainly would. If the code generated for every statement was the same
>for both compilers, then there would be no difference between the two.
>
>Charles
>
>
>-----Original Message-----
>From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On
>Behalf Of Frank Swarbrick
>Sent: Tuesday, January 14, 2020 11:37 AM
>To: IBM-MAIN@LISTSERV.UA.EDU
>Subject: Re: Migrating to new compiler release
>
>I understand that the runtime is part of LE, and is generally shared between
>versions (at least V5 and V6 seem to share the same runtime for many/most
>functions).  Conceivably it's still possible that the code generated by a
>certain version of a compiler may have defects.  Probably less likely if the
>code is in a pre-existing feature.
>
>My question has to do with the (probably slight) possibility that the code
>generated by one compiler would be different, for the same statement, for
>another.
>
>----------------------------------------------------------------------
>For IBM-MAIN subscribe / signoff / archive access instructions,
>send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
>
>----------------------------------------------------------------------
>For IBM-MAIN subscribe / signoff / archive access instructions,
>send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Reply via email to