|
Al, Disagreeing with my QPFRADJ suggestions, in a message dated 98-02-22 18:07:14 EST, you write: > Sorry Dean. I strongly disagree. If your system is adequately sized (and > with the recently announced price/performance of the new models, there's no > excuse for an undersized system anymore). QPFRADJ should be set to 3. > Note however, when doing an upgrade (CPU, CISC to RISC, or adding memory), > you must change it to 2 for the first IPL. I stand by my two primary positions, even on the newer systems. "I'm agin' it" for the 9406 (now 6X), and "it depends". I _DO_ agree that, with the advent of the newer systems, performance tuning is far less of an issue than it has been in the past. Memory is cheap, DASD is cheap (although the disk arms seem to keep going down in number), and the processors are _FAST_. In the case of my current primary client, it just _WILL NOT_ work. For example, we have a subsystem that handles program generation for the AS/Set CASE tool. It has two generator programs running in it at all times that are driven by data queues. It also runs all of our Turnover CMS promotions. When running, these generators require a memory setting of about 25,000 in order to perform optimally. However, despite having ten developers on staff, the generators often sit idle and the memory drains away under "3" while they are not active. We also are not promoting programs every second of every day. An average program takes about 5 minutes to generate when the memory is pre- allocated. The same program takes about 25 minutes to generate with QPFRADJ turned on. Although the latter instance is far preferable to the 3 hours required back when I was using the product on a B10, it is also _FAR_ inferior to the former. We experience no degradation with the pools pre-assigned, so the system _is_ adequately configured. The same problems occur when a user kicks off a FOCUS (EIS tool) report with QPFRADJ set to "3". Pre-assign the FOCUS batch subsystem with adequate memory, no problem. Turn QPFRADJ on, and even interactive users notice the change. I cannot help but think that this would affect EDI users or anyone else that has periodic functions as well... > I have systems as small as a D10 and as large as the largest 650 in the > world that work well with this set to 3. BTW, the the latest > announcements, every RISC AS/400 currently being sold is a 9406, except for > the model 150 which is a 9401. > > This is a controversal point. You may not realize it, but IBM is improving > autp tuning with every release! If you haven't looked recently, do a > WRKSHRPOOL command, and then do an F11! We did look into it when we replaced the 320 4-way with a 530 4-way. We knew the QPFRADJ setting immediately! The CE suggested running it for a while after installation, and that lasted about two days. Of course, we're not on V4R2 yet, so there may yet be hope! Regards, Dean Asmussen Enterprise Systems Consulting, Inc. Fuquay-Varina, NC USA E-Mail: DAsmussen@aol.com "As for butter versus margarine, I trust cows more than chemists." -- Joan Gussow +--- | This is the Midrange System Mailing List! | To submit a new message, send your mail to "MIDRANGE-L@midrange.com". | To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com. | Questions should be directed to the list owner/operator: david@midrange.com +---
As an Amazon Associate we earn from qualifying purchases.
This mailing list archive is Copyright 1997-2024 by midrange.com and David Gibbs as a compilation work. Use of the archive is restricted to research of a business or technical nature. Any other uses are prohibited. Full details are available on our policy page. If you have questions about this, please contact [javascript protected email address].
Operating expenses for this site are earned using the Amazon Associate program and Google Adsense.