1. March 2008 CPU Chart Published
2. Important White Paper on CPU Utilization
3. Updated zPCR
4. Tuning Letter 2007 No. 3 Update
5. EPSPT Tool for z10
6. Access to CMG Proceedings
7. Request for SMF Information
1. March 2008 CPU Chart Published
Our newest CPU Chart, which contains the latest z/OS 1.8 LSPRs and z10 processors was distributed to Tuning Letter subscribers on March 18, 2008. The CPU Chart is also available as a stand-alone publication for $250. That price also includes access to any of the older CPU Charts, and any simple updates to the chart. You can subscribe online at http://www.watsonwalker.com/chart.html.
After the CPU Chart was sent out, Larry Stout of Innovation Data Processing pointed out a typo to us. In the Excel spreadsheet, you can use column AB to sort the processors back into their original sequence (instead of column AC as suggested in the spreadsheet documentation). We’ll correct that when the next CPU Chart is distributed. Thanks, Larry!
2. Important White Paper on CPU Utilization
A frequent question I get is whether you can run z/OS effectively at 100%. Since the introduction of the Workload Manager (WLM) in 1994, I’ve said “yes, as long as you set your goals correctly.” A new White Paper (WP101208) on the WSC TechDocs site by Gary King (“Running IBM System z at High Utilization”) addresses this issue and gives you even more things to consider. It’s an excellent paper and should be read by everyone. The direct link is http://www-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/WP101208.
3. Updated zPCR
IBM has just updated their no-charge zPCR (Processor Capacity Reference) estimation tool, which now includes the new z10 processors. Look for Presentation PRS1381, “Getting Started with zPCR” on WSC TechDocs. A direct link is http://www-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/PRS1381. I’ve recommended that every System z site should always run zPCR before considering any physical hardware change, or any configuration change (such as changing the number of logical CPs, zIIPs, or zAAPs). See our Tuning Letters 2007 No. 6 and 2008 No. 1 for two thorough discussions of zPCR.
4. TUNING Letter 2007 No. 3 Update
Richard J Schwartz from State Street Corporation pointed out that we had a typo in our article on a new latent demand metric. The same incorrect calculation was seen in our Cheryl’s List #116 (http://www.watsonwalker.com/clist116.html), SHARE Hot Flashes #17, and Tuning Letter 2007 No. 3. Richard also suggested that we give a name to this new metric. Thanks, Richard!
Here is the correct calculation (we had ‘SMF70Q05 * 7’ instead of ‘8’ and simply implied the rest of the list):
In & Ready Latent Demand Metric = ((SMF70Q01 * 1) + (SMF70Q02 * 2) + (SMF70Q03 * 3) + (SMF70Q04 * 4.5) + (SMF70Q05 * 8) + (SMF70Q06 * 13) + (SMF70Q07 * 18) + (SMF70Q08 * 25.5) + (SMF70Q09 * 35.5) + (SMF70Q10 * 50.5) + (SMF70Q11 * 70.5) + (SMF70Q12 * 80)) / sum(SMF70Q00…SMF70Q12)
Where:
SMF70Q00 – # of in & ready users less than or equal to the # of CPs
SMF70Q01 – # of in & ready users is 1 more than # of CPs
SMF70Q02 – # of in & ready users is 2 more than # of CPs
SMF70Q03 – # of in & ready users is 3 more than # of CPs
SMF70Q04 – # of in & ready users is 4 or 5 more than # of CPs
SMF70Q05 – # of in & ready users is from 6 to 10 more than # of CPs
SMF70Q06 – # of in & ready users is from 11 to 15 more than # of CPs
SMF70Q07 – # of in & ready users is from 16 to 20 more than # of CPs
SMF70Q08 – # of in & ready users is from 21 to 30 more than # of CPs
SMF70Q09 – # of in & ready users is from 31 to 40 more than # of CPs
SMF70Q10 – # of in & ready users is from 41 to 60 more than # of CPs
SMF70Q11 – # of in & ready users is from 61 to 80 more than # of CPs
SMF70Q12 – # of in & ready users is greater than 80 more than # of CPs
In the calculation, I use the average of the in & ready based on these ranges. For example, the multiplier for SMF70Q08 is obtained by averaging 21 and 30 to get 25.5.
5. EPSPT Tool for z10
Marna Walle of IBM’s z/OS System Build and Install team in Poughkeepsie, recently posted this important alert on IBM-Main (http://bama.ua.edu/archives/ibm-main.html, 4Mar2008, with a subject of ‘z10 “Extract” file for the EPSPT Tool). I thought our readers would also be interested in her information:
“If you are using the EPSPT tool to verify the PTFs for the new z10, this is of interest to you.
“We have found that there was a problem during the creation of the Upgrade 2097DEVICE, Subset 2097/ZOS – Extract File. There are three PTFs identified in the 2097 PSP bucket for z/OS that were not brought forward into the extract file. The missing PTFs are for APAR OA22675, which are UA39409 (R7), UA39410 (R7), and UA39412 (R8). The R9 PTF UA39411 is in the extract file correctly. The problem was limited to only APAR OA22675.
“We intend to have the updated 2097 extract file updated tonight (March 4) or tomorrow night (March 5) at the latest. In the meantime, if you’ve used the existing 2097 extract file, note that these three PTFs were missing.”
“The incorrect 2097 extract file was introduced on February 26, and then corrected on March 5. If you have a copy of the extract file taken between these dates, then note that these three PTFs were missing, and obtain the latest 2097 extract file.”
6. Access to CMG Proceedings
The CMG (Computer Measurement Group) Proceedings contain full peer-reviewed technical papers on capacity planning, performance, systems management, and many other topics. CMG has just announced public access to conference proceedings from 1997 to 2005. You will need to register, but CMG membership is not required. Go to http://www.cmg.org/proceedings for these public documents.
While you are on the CMG site, sign up for ‘MeasureIT’, another CMG publication for members and non-members. The latest issue contains part 3 of an excellent series on SAS 9 Tips by Phil Mason. Go tohttp://www.cmg.org and select ‘Publications’, then ‘MeasureIT’.
7. Request for SMF Information
When discussing the new SMF logger function at SHARE, we realized that nobody seems to have a good idea of what volume of SMF data is being used these days. So at my ‘Hot Flashes’ session, I asked people to email or fax me a copy of any output from IFASMFDP (which shows types of records, number of records, record size, and interval). This can simply be the output from your normal run, or if you want to make a separate run, you could use either a full day or peak hours.
I’d appreciate any and all input, and will publish my findings in a later Cheryl’s List. All materials will be kept confidential.
Stay Tuned!