>I believe hyper dispatch is very different from zIIP dispatch. I stand by my >assumption that GP dispatch is very different from zIIP dispatch, or why would >there be the ZIIPAWMT parameter and have the comment about waking up after >that interval to see if there is work. When non-zIIP work comes ready and a >processor is free, the work is dispatched. No waiting. The comments in the >parameter description for ZIIPAWMT is describing a polling environment, at >least to me.
There is CCAWTM which does the same thing as ZAAPAWMT and ZIIPAWMT but for CPs. I think your statement about CP work being dispatched immediately when a CP is free, does not hold true. It was before "alternate wait management" was introduced (I don't remember when that was), when newly ready, higher priority work had interrupted lower priority work on a CP, or woken up a waiting CP. >I have looked through the internet, OK not the be-all and end-all but a >reasonable place to start, and there is a lot on what can get dispatched on a >zIIP, but no detail on how. The last five words are the very reason I started this thread. When it comes to zIIPs, there is much only fuzzy information out there. I did google and read every presentation I could get a hand on. So, coming back to one of my questions: The fuzzy information say that zIIP eligible work will suffer from a ZIIPAWMT delay when the zIIPs are overloaded and are asking for help from CPs. You mentioned the ZIIPAWMT delay *before* zIIPs idle zIIPs will start to work on newly arrived work. Ok, this is part of the delay. But as soon as there is a steady load of zIIP work, that start up delay will not occur again (think of my simplified case I initially described). Similarly the "ask for help ZIIPAWMT delay" will occur only once when there is work constantly seen on the the zIIP WUQ. I still think that, theoretically, zIIP eligible work will be served somewhat better than non zIIP eligible work *when* the zIIPs is being helped (depending on the works priority, of course). There are two processors looking after the single work queue. Oversimplified, you think? Maybe but its my starting point anyway. -- Peter Hunkeler ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
