Thanks for all the responses on this one, especially Jim and Peter, and all the links. I unsubscribed from this list some months ago and re-subscribed last night, now realizing how much I might have missed (week-long AMODE32 threads notwithstanding)
The info I initially received was from an IBM gent in the z processor group, and was consistent with the Processor Optimization doc pg 31. Some time ago I converted most of our MVCL*s over to MVC and XC instructions, resulting in modest cpu savings. Since that time, I began checking for page alignment of source and target addresses and routing to MVCL* rtns when page aligned, and lastly tried to make things more eligible for ADM or the 'near-memory engine' with BNDRY=PAGE which may have run headlong into fragmentation issue described or... > Whether use of the near-memory engine helps or hurts your performance > depends on how/when you were using the storage before the MVCL/MVCLE, > and how/when you use it after the MVCL/MVCLE. Thanks again! Mike -----Original Message----- From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf Of Don Poitras Sent: Friday, May 31, 2019 10:58 AM To: [email protected] Subject: Re: BNDRY=PAGE possible CPU hit? Expanding the ... leads one to (there might be more, but this is the one I found): https://www.ibm.com/developerworks/community/files/form/anonymous/api/library/ff4563be-756e-49bf-9de9-6a04a08026f1/document/07c69512-ac74-4394-87b9-a61ea201347e/media/IBMzSystemsProcessorOptimizationPrimerv2.pdf Gratuitous wrap noted. Other files in this "community" can be found at: https://www.ibm.com/developerworks/community/groups/service/html/communityview?communityUuid=9a17556c-6094-4201-acd0-d8125a3fa0db In article <of5917c456.92910c5f-on0025840b.004f9464-8525840b.004f9...@notes.na.collabserv.com> you wrote: > See page 31 in this document: > https://www.ibm.com/.../IBMzSystemsProcessorOptimizationPrimerv2.pdf > Whether use of the near-memory engine helps or hurts your performance > depends on how/when you were using the storage before the MVCL/MVCLE, > and how/when you use it after the MVCL/MVCLE. > Jim Mulder z/OS Diagnosis, Design, Development, Test IBM Corp. > Poughkeepsie NY > "IBM Mainframe Discussion List" <[email protected]> wrote on > 05/30/2019 11:12:13 PM: > > From: "Michael Hochee" <[email protected]> > > To: [email protected] > > Date: 05/31/2019 12:31 AM > > Subject: BNDRY=PAGE possible CPU hit? > > Sent by: "IBM Mainframe Discussion List" <[email protected]> > > > > Hi, > > I recently added the BNDRY=PAGE parameter to a set of STORAGE > > OBTAINS which acquire storage areas of various sizes from several > > low private subpools. My intent was a reduction of CPU used by > > subsequent MVCLE instructions, as ADM would more likely be employed > > for MVCLE executes, since the storage areas involved were page > > aligned (apparently an ADM pre-req.) Unfortunately my initial > > testing revealed a significant increase in CPU consumption, rather > > than a reduction. > > > > I did a few searches of the IBM-MAIN archive and found nothing > > involving increased CPU overhead resulting from BNDRY=PAGE usage. > > Any thoughts on what might be causing the elevated CPU? Again, the > > only change made was adding BNDRY=PAGE. (I have since backed the > > change off, tested, reapplied the change and tested with the same > results) -- Don Poitras - SAS Development - SAS Institute Inc. - SAS Campus Drive [email protected] (919) 531-5637 Cary, NC 27513 ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
