Re: Maximum # of processors for Z/os
Check the IBM LSPR tables under single z/OS image. There's a 64 engine single z/OS 1.9 image in the table. I don't see an 80 engine single z/OS LPAR listed yet, but I assume that'll appear fairly soon. https://www.ibm.com/servers/resourcelink/lib03060.nsf/pages/lsprindex?OpenDocument How many engines do you want to run in your largest z/OS LPAR? - - - - - Timothy Sipples Resident Enterprise Architect STG Value Creation Complex Deals Team IBM Growth Markets (Based in Singapore) E-Mail: timothy.sipp...@us.ibm.com -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Maximum # of processors for Z/os
Has Z/os changed that number, if so what is the current number before thruput either is not helped or there is a straight line being nio help/no hurt ? I believe that the answer is: a number no lower than the maximum number of processors that z/OS (and the machine) documents that you may define to a z/OS image. That by no means implies that the incremental benefit for moving from say 31 to 32 processors is the same as the benefit for moving from 3 to 4. If you have access to machines on which you can play, you could consider checking out the MP CPU Capability adjustment factors returned by the STSI instruction (CSRSI service). Peter Relson z/OS Core Technology Design rel...@us.ibm.com -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Maximum # of processors for Z/os
On Tue, 5 Oct 2010 09:02:34 -0400 Peter Relson wrote: If you have access to machines on which you can play, you could consider checking out the MP CPU Capability adjustment factors returned by the STSI instruction (CSRSI service). I wish ... A fully populated multi-book z196 to play on mmm. I suspect a few on this list could have a fine old time with unfettered access to one of them. Shane ... -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Maximum # of processors for Z/os
And those limits, as announced, are: # Up to 80 processors per logical partition with z/OS V1.11 and later on zEnterprise servers. # Up to 60 LPARs on a single server with IBM zEnterprise System (zEnterprise), IBM System z10 Enterprise Class (z10 EC), and IBM System z9 Enterprise Class (z9 EC) servers. Peter Relson wrote: Has Z/os changed that number, if so what is the current number before thruput either is not helped or there is a straight line being nio help/no hurt ? I believe that the answer is: a number no lower than the maximum number of processors that z/OS (and the machine) documents that you may define to a z/OS image. That by no means implies that the incremental benefit for moving from say 31 to 32 processors is the same as the benefit for moving from 3 to 4. If you have access to machines on which you can play, you could consider checking out the MP CPU Capability adjustment factors returned by the STSI instruction (CSRSI service). snip -- John Eells z/OS Technical Marketing IBM Poughkeepsie ee...@us.ibm.com -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Maximum # of processors for Z/os
On 10/4/2010 11:57 AM, Ed Gould wrote: A LONG time ago(20+ years I think) IBM did some studies on the maximum number of Processors that MVS (at that time) could effectively use before there was degraded throughput. My memory says is was 15 (I could be wrong). Has Z/os changed that number, if so what is the current number before thruput either is not helped or there is a straight line being nio help/no hurt ? I am talking 1 copy of MVS on a CEC not mulriple copies. Bob Rogers gave a presentation about this (can't remember right now if it was at a TDM or at SHARE) right around the time they raised the number of logical CPs supported by z/OS to 64. It was interesting because the impact of having many logical CPs depends on the workload--with so-called traditional workloads faring somewhat worse than so-called new workloads. In any case, after studying this question IBM concluded that changes in hardware and operating system design have and will continue to reduce interlocks to the point where it makes sense for z/OS developers to invest valuable development resources supporting higher and higher numbers of logical CPs per image. The current limit is 80. -- Edward E Jaffe Phoenix Software International, Inc 831 Parkview Drive North El Segundo, CA 90245 310-338-0400 x318 edja...@phoenixsoftware.com http://www.phoenixsoftware.com/ -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Maximum # of processors for Z/os
I would suggest that it depends on the amount of processor you're willing to give up to MP effect. Certainly you should be using HiperDispatch which can give you back up to around 10% in the larger n-ways. Depends on what your world looks like and if creating another image makes sense. I would probably prefer 3 8-ways than 1 24-way. But I could be convinced of a. pair of 12-ways. nor...@desertwiz.biz Sent from my Verizon Wireless BlackBerry -Original Message- From: Edward Jaffe edja...@phoenixsoftware.com Sender: IBM Mainframe Discussion List IBM-MAIN@bama.ua.edu Date: Tue, 5 Oct 2010 06:48:40 To: IBM-MAIN@bama.ua.edu Reply-To: IBM Mainframe Discussion List IBM-MAIN@bama.ua.edu Subject: Re: Maximum # of processors for Z/os On 10/4/2010 11:57 AM, Ed Gould wrote: A LONG time ago(20+ years I think) IBM did some studies on the maximum number of Processors that MVS (at that time) could effectively use before there was degraded throughput. My memory says is was 15 (I could be wrong). Has Z/os changed that number, if so what is the current number before thruput either is not helped or there is a straight line being nio help/no hurt ? I am talking 1 copy of MVS on a CEC not mulriple copies. Bob Rogers gave a presentation about this (can't remember right now if it was at a TDM or at SHARE) right around the time they raised the number of logical CPs supported by z/OS to 64. It was interesting because the impact of having many logical CPs depends on the workload--with so-called traditional workloads faring somewhat worse than so-called new workloads. In any case, after studying this question IBM concluded that changes in hardware and operating system design have and will continue to reduce interlocks to the point where it makes sense for z/OS developers to invest valuable development resources supporting higher and higher numbers of logical CPs per image. The current limit is 80. -- Edward E Jaffe Phoenix Software International, Inc 831 Parkview Drive North El Segundo, CA 90245 310-338-0400 x318 edja...@phoenixsoftware.com http://www.phoenixsoftware.com/ -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Maximum # of processors for Z/os
--- On Tue, 10/5/10, Edward Jaffe edja...@phoenixsoftware.com wrote: From: Edward Jaffe edja...@phoenixsoftware.com Subject: Re: Maximum # of processors for Z/os To: IBM-MAIN@bama.ua.edu Date: Tuesday, October 5, 2010, 8:48 AM On 10/4/2010 11:57 AM, Ed Gould wrote: A LONG time ago(20+ years I think) IBM did some studies on the maximum number of Processors that MVS (at that time) could effectively use before there was degraded throughput. My memory says is was 15 (I could be wrong). Has Z/os changed that number, if so what is the current number before thruput either is not helped or there is a straight line being nio help/no hurt ? I am talking 1 copy of MVS on a CEC not mulriple copies. Bob Rogers gave a presentation about this (can't remember right now if it was at a TDM or at SHARE) right around the time they raised the number of logical CPs supported by z/OS to 64. It was interesting because the impact of having many logical CPs depends on the workload--with so-called traditional workloads faring somewhat worse than so-called new workloads. In any case, after studying this question IBM concluded that changes in hardware and operating system design have and will continue to reduce interlocks to the point where it makes sense for z/OS developers to invest valuable development resources supporting higher and higher numbers of logical CPs per image. The current limit is 80. Ed: That is terrific. Does anyone have the share presentation they can share with me or perhaps a manual # I can look for that talks about this some more? Thanks, Ed -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Maximum # of processors for Z/os
http://www-304.ibm.com/servers/resourcelink/lib03060.nsf/pages/lsprzOS11MIJuly2010?OpenDocumentpathID= includes all zSeries. On Tue, Oct 5, 2010 at 10:49 PM, Ed Gould ps2...@yahoo.com wrote: --- On Tue, 10/5/10, Edward Jaffe edja...@phoenixsoftware.com wrote: From: Edward Jaffe edja...@phoenixsoftware.com Subject: Re: Maximum # of processors for Z/os To: IBM-MAIN@bama.ua.edu Date: Tuesday, October 5, 2010, 8:48 AM On 10/4/2010 11:57 AM, Ed Gould wrote: A LONG time ago(20+ years I think) IBM did some studies on the maximum number of Processors that MVS (at that time) could effectively use before there was degraded throughput. My memory says is was 15 (I could be wrong). Has Z/os changed that number, if so what is the current number before thruput either is not helped or there is a straight line being nio help/no hurt ? I am talking 1 copy of MVS on a CEC not mulriple copies. Bob Rogers gave a presentation about this (can't remember right now if it was at a TDM or at SHARE) right around the time they raised the number of logical CPs supported by z/OS to 64. It was interesting because the impact of having many logical CPs depends on the workload--with so-called traditional workloads faring somewhat worse than so-called new workloads. In any case, after studying this question IBM concluded that changes in hardware and operating system design have and will continue to reduce interlocks to the point where it makes sense for z/OS developers to invest valuable development resources supporting higher and higher numbers of logical CPs per image. The current limit is 80. Ed: That is terrific. Does anyone have the share presentation they can share with me or perhaps a manual # I can look for that talks about this some more? Thanks, Ed -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- Mike A Schwab, Springfield IL USA Where do Forest Rangers go to get away from it all? -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Maximum # of processors for Z/os
On Mon, Oct 4, 2010 at 2:57 PM, Ed Gould ps2...@yahoo.com wrote: A LONG time ago(20+ years I think) IBM did some studies on the maximum number of Processors that MVS (at that time) could effectively use before there was degraded throughput. My memory says is was 15 (I could be wrong). Has Z/os changed that number, if so what is the current number before thruput either is not helped or there is a straight line being nio help/no hurt ? I am talking 1 copy of MVS on a CEC not mulriple copies. Well, technically, not helped and straight line... are the same; you mean, gets worse or is a straight line... 'Tis an interesting question. -- zMan -- I've got a mainframe and I'm not afraid to use it -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html