Re: PL/I RFE to vote for (or not)
On Wed, 2 Sep 2020 20:39:27 +, Robert Prins wrote: > >>> quotient = QUOTREM(dividend, divisor, remainder), >>> >>> With quotient and remainder generated in one go by the execution of a DP >>> instruction (and probably/possibly/maybe/likely also by the various other >>> divide instructions) >>> >>> https://www.ibm.com/developerworks/rfe/execute?use_case=viewRfe_ID=145001 >>> >> Might this be better handled by an optimizing compiler's collapsing >> common expressions, benefiting existing code with no source change? > >Yes, obviously, but Enterprise PL/I doesn't seem to realise this when a >division >by a value is followed by a modulo(same-value). > But likewise: quotient = QUOTREM(dividend, divisor, remainder); doesn't exist. Either would require a compiler enhancement. I favor the technique that would benefit the most users with no changes to source code that might be incompatible with older compilers. -- gil -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: PL/I RFE to vote for (or not)
There is one advantage to COBOL's rather archaic arithmetic statements: DIVIDE dividend BY divisor GIVING quotient REMAINDER remain DIVIDE divisor INTO dividend GIVING quotient REMAINDER remain From: IBM Mainframe Discussion List on behalf of Robert Prins Sent: Wednesday, September 2, 2020 1:17 PM To: IBM-MAIN@LISTSERV.UA.EDU Subject: PL/I RFE to vote for (or not) quotient = QUOTREM(dividend, divisor, remainder), With quotient and remainder generated in one go by the execution of a DP instruction (and probably/possibly/maybe/likely also by the various other divide instructions) https://www.ibm.com/developerworks/rfe/execute?use_case=viewRfe_ID=145001 -- Robert AH Prins robert.ah.prins(a)gmail.com The hitchhiking grandfather - https://prino.neocities.org/indez.html Some REXX code for use on z/OS - https://prino.neocities.org/zOS/zOS-Tools.html -- 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
Re: PL/I RFE to vote for (or not)
On 2020-09-02 17:02, Paul Gilmartin wrote: On Wed, 2 Sep 2020 19:17:21 + , Robert Prins wrote: quotient = QUOTREM(dividend, divisor, remainder), With quotient and remainder generated in one go by the execution of a DP instruction (and probably/possibly/maybe/likely also by the various other divide instructions) https://www.ibm.com/developerworks/rfe/execute?use_case=viewRfe_ID=145001 Might this be better handled by an optimizing compiler's collapsing common expressions, benefiting existing code with no source change? Yes, obviously, but Enterprise PL/I doesn't seem to realise this when a division by a value is followed by a modulo(same-value). Robert -- Robert AH Prins robert(a)prino(d)org The hitchhiking grandfather - https://prino.neocities.org/indez.html Some REXX code for use on z/OS - https://prino.neocities.org/zOS/zOS-Tools.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: PL/I RFE to vote for (or not)
On Wed, 2 Sep 2020 19:17:21 + , Robert Prins wrote: >quotient = QUOTREM(dividend, divisor, remainder), > >With quotient and remainder generated in one go by the execution of a DP >instruction (and probably/possibly/maybe/likely also by the various other >divide instructions) > >https://www.ibm.com/developerworks/rfe/execute?use_case=viewRfe_ID=145001 > Might this be better handled by an optimizing compiler's collapsing common expressions, benefiting existing code with no source change? -- gil -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
