Re: AW: Re: DFSORT and zIIP
On 7/20/2016 10:45 PM, Peter Hunkeler wrote: I'm longing for the day when also the zIIPs disappear again and IBM has found a better way to charge software license fees. I agree it would be GREAT(!) if kneecapping of CPs was removed and replaced with PER CORE software charging like other platforms use. Until that day arrives, specialty engines are the only way for small and medium-sized companies to fully experience the current awesomeness of z Systems processor technology. BTW, some recent performance studies have shown that kneecapping can actually improve performance in larger shops by increasing the number of dispatch points. For example, take a look at the results enumerated in this "Best Session Award" presentation from SHARE in San Antonio: http://share.confex.com/data/handout/share/126/Session_18345_handout_8753_0.pdf -- Edward E Jaffe Phoenix Software International, Inc 831 Parkview Drive North El Segundo, CA 90245 http://www.phoenixsoftware.com/ -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: DFSORT and zIIP
>For DB2 Sort for z/OS, zIIP -- Peter Hunkeler >exploitation makes technical sense. For DFSORT -- except for exploiting >zIIPs on behalf of DB2 utilities and in other ancillary ways -- it doesn't >seem to make technical sense. All this speciality engine thing never made any *techincal* sence to me at all. It's a pure financial thing. >Maybe in the future, as the technologies change and evolve, it will. I'm longing for the day when also the zIIPs disappear again and IBM has found a better way to charge software license fees. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: DFsort and zIIP
>You are correct that the ZIIP dispatcher is not as sophisticated as the >regular dispatcher. I dare to contradict, not intending to question you expertise. It is my understanding that there is only one dispatcher in MVS. It handles work on the CP WUQ as well as work on the zIIP WUQ. The reason for the wait time mechanism is explained in Init Ref, IEAOPTxx. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
AW: Re: DFsort and zIIP
>DFSORT can use zIIP on behalf of DB2 utilities, but not otherwise. Here's >more information: > >At this time, IBM has no plan for enabling DFSORT to exploit the system z9 >Integrated Information Processor (zIIP). IBM realizes DFSORT remains a >prominent component of our customers' batch workloads. However, the >added controls that would need to be implemented in order to maintain our >high standards for performance, reliability and system integrity are not >justified in view of estimations that there is a low offload potential and >the value to clients may be marginal.[snip] Interesting statement. I seem to remember that SyncSort offers an Add-On package that allows certain SyncSort processing to be offloaded to zIIPs. The above statement suggest that SyncSort's perfocmance is suffering from using zIIPs (simplified and exagerated, I know). Also "IBM Sort for DB2 for z/OS" (can't remember the exact name), is offloading to zIIPs, if I remember correctly. This procuct is based on SyncSort code as far as I understand. -- Peter Hunkeler -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
