Hi Kurt, thanks for your response. What we do is we scan the entire SMP/E environment, so we want to make sure we optimize as much as possible. We already query for one specific zone and for one (in one case for two) specific entry type(s) and the only filtering we do is that we request only "not deleted" entries. So, there doesn't seem to be much room for improvement :-(
Our idea was that we do the dynamic allocation (using the DDDEFs) once and then perform several requests based on these allocations, thus reducing the overhead within GIMAPI (which now does the dynamic allocation for each request). One more question, though: the last level qualifier of a CSI has to be "CSI", right ? Or can it have any name, if pre-allocated with SMPCSI ? Thanks, Beate > -----Original Message----- > From: IBM Mainframe Discussion List > [mailto:[EMAIL PROTECTED] On Behalf Of Kurt Quackenbush > Sent: Wednesday, September 14, 2005 3:03 PM > To: [email protected] > Subject: Re: [IBM-MAIN] GIMAPI performance issues > > <snip> > > We thought about > > reducing this overhead by dynamically pre-allocating the DDDEF > > datasets and then call GIMAPI for several checks on these datasets. > > There have been some improvements made to the internal > processing of GIMAPI, but not since about V3.2. In any case, > I doubt the dynamic allocation of your SMPCSI data sets is > the culprit... unless you find GIMAPI is waiting for data > sets because they are migrated or in use by other tasks. > > > are there other hints and tips > > to improve the performance of GIMAPI processing ? > > You might be able to improve the performance by changing your > Query. Be specific with the zones and entry types you want > to query, and don't return more subentries than you require. > Also, Filter processing can add overhead, so use it sparingly. > > > Is there a way to detect > > whether such a dataset is a GLOBAL CSI ? > > No. > > Kurt Quackenbush -- IBM, SMP/E Development > > ---------------------------------------------------------------------- > 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 > ---------------------------------------------------------------------- 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
