Re: Real World data on CPU cost savings with ZIIP/ZAAP engines

[Default] On Mon, 16 Oct 2017 09:45:34 -0700 (PDT), in bit.listserv.ibm-main Dan Kalmar wrote: >In our case the plan is to convert CPU intensive Cobol programs to Java in >attempt to offload work onto the ZIIPs. >We don't have ZIIP eligible workloads that we can measure

Re: Real World data on CPU cost savings with ZIIP/ZAAP engines

Thanks Bob. You can see my analysis technique from the paper available via SHARE, CMG & MXG. http://www.mxg.com/downloads/Eligible%20Time%20analysis.pdf It is from 2007 but is just as applicable today. Many customers use this same LCS analysis every month to determine when to add more zIIP

Re: Real World data on CPU cost savings with ZIIP/ZAAP engines

Please check with these kind and smart people: https://www.ibm.com/software/info/eagletco/iteconomics/ Timothy Sipples IT Architect Executive, Industry Solutions, IBM z Systems, AP/GCG/MEA

Re: Real World data on CPU cost savings with ZIIP/ZAAP engines

Subject: Re: Real World data on CPU cost savings with ZIIP/ZAAP engines I don't know of any from IBM (which does not mean there are none), I'm afraid. But you can use on PROJECTCPU in IEAOPTxx to simulate zIIPs and read the RMF reports to see what is reported as zIIP-eligible work. And, you can use

Re: Real World data on CPU cost savings with ZIIP/ZAAP engines

One of our customers had issues with Java workloads spilling over onto GCPs due to the 2-1 constraint. It cost them serious money. Moving to a z13 with faster zIIP processors solved the problem. On 16/10/2017 6:42 PM, John Eells wrote: I don't know of any from IBM (which does not mean there

Re: Real World data on CPU cost savings with ZIIP/ZAAP engines

I don't know of any from IBM (which does not mean there are none), I'm afraid. But you can use on PROJECTCPU in IEAOPTxx to simulate zIIPs and read the RMF reports to see what is reported as zIIP-eligible work. And, you can use the non-zIIP time in that same mode to estimate the probable