HIGH ZIIP Usage after ZPSAVERS activation

Document ID : KB000121536
Last Modified Date : 26/11/2018
Show Technical Document Details
Hi; we have SAMS active for long time with ZIIPIT=Y in some lpars and although honestly i cannot tell you for sure that it really made a difference in CP/ZIIP utilization. If there is a way to determine that I would appreciate. But the point here is that since November 20th, I've had an excessive increase on ZIIP utilization, and this is when ZPSAVERS was installed. I don't know much about ZPSAVERS but seems to be related to SYNCSORT and SYNCSORT is also analyzing the issue, we need to determine what is the interaction between SAMS and ZPSAVERS that caused this high utilization. MVS team took a look at the generated SMF 30 data for ZIIP and it appears SAMS started to take advantage of ZPSavers features and recorded ZIIP time that otherwise would go to CP without ZP Savers. So with that, is SAMS dependent of ZPSAVERS to utilize ZIIP processors ?
I checked the SYNSORT website and found the following: "Offload up to 90% of Copy, SMS Compression, Sort CPU cycles to zIIP" This is really a nice feature from Synsort as they are offloading their sort routines off the main processor to a ZIIP processor. This will free up more cycle for the main CP. CA Vantage is a decent size user of SORT. We obtain data from other sources and SORT it to match the report you are after. Since Synsort is now using a ZIIP processor, you will see the usage go up from Vantage since we are calling SORT.
1). Vantage calls SORT.
2). Synsort offloads process to ZIIP.
 3). Data is returned to VANTAGE.
What you are seeing appears to be quite normal.