That answers a question I had been meaning to ask. We upgraded (finally!) from
CICS 4.1 to CICS/TS 1.3 a few weeks ago and had seen our CPU contention
problems decrease dramatically; I had wondered if TS had improved use of
multiple processors. Apparently that was indeed one of the upsides. (
>Ted, You said that you lose 12-15% when you add the second engine and another
>15% or more for each additional engine that you add.
I said for up to a certain size and I was quoting from (poor) memory.
>Here you present a 30% reduction for the ninth CP compared to the first.
Based on LSPR,
On Sat, 10 Jun 2006 00:00:00 GMT, Ted MacNEIL
<[EMAIL PROTECTED]> wrote:
>>I won't discount that could be true with an unusual pathologically SMP
unfriendly workload but it's not normal on modern zSeries environments in
my experience.
>
>Take a look at LSPR.
>z/990 2084-308 -- 2993 MIPS
>z/990
: IBM-MAIN@BAMA.UA.EDU
Subject: Re: One or two CPUs - the pros & cons
Hi Ted,
This should not be the case. You probably have some other problem that
you
see manifesting itself as a big loss of capacity. The actual loss
should be
quite low on a z/series box, (in the area of 1% to 2% depending
>This should not be the case. You probably have some other problem that you
>see manifesting itself as a big loss of capacity. The actual loss should be
>quite low on a z/series box, (in the area of 1% to 2% depending on a number of
>factors).
The "loss" is quite consistent with LSPR.
Look at
Hi Ted,
This should not be the case. You probably have some other problem that you
see manifesting itself as a big loss of capacity. The actual loss should be
quite low on a z/series box, (in the area of 1% to 2% depending on a number
of factors).
If you are indeed getting those results then so
This was supposed to go to the list!
--Original Message--
To: Gibney, Dave
Sent: Jun 11, 2006 14:21
Subject: Re: One or two CPUs - the pros & cons
> How many installations even bother to measure the performance of the usual
> single purpose Micro$oft server box? If it does
>I'm sure all capacity planning people must have problems selling this to their
>people.
For 25 years, in my case.
I still get push-back from management.
Especially, these days with the flaws in LSPR:
1. They introduced new workloads for the z/990 (WebSphere for example).
And, they never calibrat
Gerhard Adam wrote:
> What are you losing? It isn't as if these processors are off playing
> solitaire. They're paying the cost of communication to allow more
> simultaneous operations for YOUR workload. The primary benefit of this
> approach is to reduce the queueing impacts of multiple units o
On Sat, 2006-06-10 at 15:35 -0700, Gerhard Adam wrote:
> >I disagree.
> >I lose 12-15% of the engine when I add the second one and I lose
> >another 15% (or more) for each engine thereafter, up to about 8 or 9
> >engines. Going from >a 308 to a 309 adds only about 300 MIPS when an
> >engine is (
>I disagree.
>I lose 12-15% of the engine when I add the second one and I lose
>another 15% (or more) for each engine thereafter, up to about 8 or 9
>engines. Going from >a 308 to a 309 adds only about 300 MIPS when an
>engine is (nominally) 450 MIPS.
>
>That to me is more than "almost" not wort
- Original Message -
From: "Brian Westerman" <[EMAIL PROTECTED]>
Newsgroups: bit.listserv.ibm-main
To:
Sent: Saturday, June 10, 2006 12:29 AM
Subject: Re: One or two CPUs - the pros & cons
The proper, (and probably cheaper) answer to this problem is to add
another
>I won't discount that could be true with an unusual pathologically SMP
>unfriendly workload but it's not normal on modern zSeries environments in my
>experience.
Take a look at LSPR.
z/990 2084-308 -- 2993 MIPS
z/990 2084-309 -- 3299 MIPS
Difference: 306 MIP with the engine nominally 450 MIPS.
ailto:[EMAIL PROTECTED] On
Behalf Of Ted MacNEIL
Sent: Friday, June 09, 2006 8:00 PM
To: IBM-MAIN@BAMA.UA.EDU
Subject: Re: One or two CPUs - the pros & cons
>First, for the past 10 years or so the SMP effect has been negated to
the point that it is "almost" not worth figuring, especia
>First, for the past 10 years or so the SMP effect has been negated to the
>point that it is "almost" not worth figuring, especially for anything under 10
>engines.
I disagree.
I lose 12-15% of the engine when I add the second one and I lose another 15%
(or more) for each engine thereafter, up
The proper, (and probably cheaper) answer to this problem is to add another
smaller CPU to the mix, if you have only 2 CPUs, CICS can't really take
advantage of all of the processor complex as it could. It's almost
certainly cheaper to add more engines than to upgrade, unless you have an
older box
Hi,
Yes, you are wrong. You are still focusing on too small of a picture and
missing the "greater view". My example used TCP/IP, VTAM and CICS as
individual entities to make the concepts easier for people to comprehend,
not to hide the fact that they are made up of TCBs.
First, for the past 10
> -Original Message-
> From: IBM Mainframe Discussion List On Behalf Of Ted MacNEIL
>
> PS: 90% of all statistics are made up on the spot!
No; that's still 74%. :-)
-jc-
--
For IBM-MAIN subscribe / signoff / archiv
On Jun 7, 2006, at 10:48 AM, Chase, John wrote:
-Original Message-
From: IBM Mainframe Discussion List On Behalf Of Hal Merritt
In a few years of doing this kind of stuff, I think the only
absolute I could think of is: 'There are absolutely no
absolutes'. ;-)
On second thought, per
>BTW, Wikipedia thinks there are only 6.5 billion folks, but that is neither
>here nor there.
Wikipedia is not necessarilly the best source.
I think we topped 7B, late last year.
PS: 90% of all statistics are made up on the spot!
-
-teD
300,000 Kilometres per Second
Not only is it a good idea
me Discussion List [mailto:[EMAIL PROTECTED] On
Behalf
>Subject: Re: One or two CPUs - the pros & cons
>
>One-third of all people who ever lived have not died.
>
>Charles
>=
>
>There have been over 70 billi
rame Discussion List [mailto:[EMAIL PROTECTED] On Behalf
Of Jeffrey D. Smith
Sent: Wednesday, June 07, 2006 1:05 PM
To: IBM-MAIN@BAMA.UA.EDU
Subject: Re: One or two CPUs - the pros & cons
=
-Original Message-
From: "Cha
Jeffrey D. Smith wrote:
=
-Original Message-
From: "Charles Mills" <[EMAIL PROTECTED]>
Sent: 6/7/2006 11:51 AM
To: "IBM-MAIN@BAMA.UA.EDU"
Subject: Re: One or two CPUs - the pros & cons
One-thir
=
-Original Message-
From: "Charles Mills" <[EMAIL PROTECTED]>
Sent: 6/7/2006 11:51 AM
To: "IBM-MAIN@BAMA.UA.EDU"
Subject: Re: One or two CPUs - the pros & cons
One-third of all people who ever
One-third of all people who ever lived have not died.
Charles
-Original Message-
From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf
Of Chase, John
Sent: Wednesday, June 07, 2006 8:48 AM
To: IBM-MAIN@BAMA.UA.EDU
Subject: Re: One or two CPUs - the pros &
Don't forget Elvis.
-Original Message-
From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED]
Behalf Of Mark Vitale
Sent: Wednesday, June 07, 2006 12:08 PM
To: IBM-MAIN@BAMA.UA.EDU
Subject: Re: One or two CPUs - the pros & cons
>
> Rule: Everybody dies.
What? No "good soldier" replies?
--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.h
>> Rule: Everybody dies.
>>
>> Exception: ?
>Enoch? Genesis 5:24, "Enoch walked with God; then he was no more,
>because God took him away."
>Elijah? 2 Kings 2:11, "As they were walking along and talking together,
>suddenly a chariot of fire and horses of fire appeared and separated the
>tw
>
> Rule: Everybody dies.
>
> Exception: ?
I see John McKown beat me to the draw as I was writing
this. Oh well. I'll post anyway, to give another
verse about Enoch:
"By faith Enoch was taken from this life,
so that he did not experience death; he
could not be found, because God ha
> -Original Message-
> From: IBM Mainframe Discussion List
> [mailto:[EMAIL PROTECTED] On Behalf Of Chase, John
> Sent: Wednesday, June 07, 2006 10:48 AM
> To: IBM-MAIN@BAMA.UA.EDU
> Subject: Re: One or two CPUs - the pros & cons
>
>
> > -O
> -Original Message-
> From: IBM Mainframe Discussion List On Behalf Of Hal Merritt
>
> In a few years of doing this kind of stuff, I think the only
> absolute I could think of is: 'There are absolutely no absolutes'. ;-)
>
>
> On second thought, perhaps one absolute: "The only rule tha
is message
in error, please notify the sender by return email.
Jon Brock <[EMAIL PROTECTED]>
Sent by: IBM Mainframe Discussion List
06/07/2006 07:21 AM
Please respond to IBM Mainframe Discussion List
To: IBM-MAIN@BAMA.UA.EDU
cc:
Subject: Re: One o
>The last time I checked, an individual CICS region was limited by the engine
>size.
That was a longgg time ago.
There are now multiple TCB's in a CICS address space.
And, for example, each DB2 thread is a TCB.
-
-teD
300,000 Kilometres per Second
Not only is it a good idea!
It's the LA
Mainframe Discussion List [mailto:[EMAIL PROTECTED] On
Behalf Of Anne & Lynn Wheeler
Sent: Tuesday, June 06, 2006 5:08 PM
To: IBM-MAIN@BAMA.UA.EDU
Subject: Re: One or two CPUs - the pros & cons
Snipped the rest!
LEGAL DISCLAIMER
The information transmitted is intended solely
The last time I checked, an individual CICS region was limited by the
engine size. This means that, even if you had work that a 2 x 50 MIPS could
handle according to raw MIPS, if you had a CICS region which required 75 MIPS
to deliver acceptable response times, you were hosed on such a
ssage-
From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On
Behalf Of Charles Mills
Sent: Tuesday, June 06, 2006 1:51 PM
To: IBM-MAIN@BAMA.UA.EDU
Subject: Re: One or two CPUs - the pros & cons
I think you can state two absolutes:
NOTICE: This electronic mail message and any f
Good one, Gilbert. Thanks for a good laugh!
On Wed, 7 Jun 2006 08:42:26 -0400, Gilbert Saint-Flour
<[EMAIL PROTECTED]> wrote:
>On Wednesday 07 June 2006 07:52, Tom Marchant wrote:
>
>> There is no such thing as "MIP"
>
>Of course, there is !!!
>
>one MIP, two MIPS
>
>and similarly,
>
>one CIC,
On Wednesday 07 June 2006 07:52, Tom Marchant wrote:
> There is no such thing as "MIP"
Of course, there is !!!
one MIP, two MIPS
and similarly,
one CIC, two CICS
one IM, two IMS
and so on . . . :=)
--
Gilbert Saint-Flour
GSF Software
http://gsf-soft.com/
mailto:[EMAIL PROTECTED]
---
There is no such thing as "MIP"
--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.htm
One of the few RoT to stand the test of time:
"you'd have to have rocks in your head to run a UP".
Take it or leave it, but I advise customers against single engine
environments.
Shane ...
--
For IBM-MAIN subscribe / signoff / a
Hello Brian,
Some questions:
- Did your research took into account the common storage overhead? other
synchronization overhead?
- I know that CPU speed of the 50 Mips proc. is equal to the 100 Mips
proc. The difference is in total capacity. But its like the supermarket.
You can have 2 cashiers
Fortunately SMP has little relevence today.
This is a subject that I lecture on over (and over and over again) about
vertical vs horizontal capacity with each new client we get, and it never
seems that people are "completely" satisfied with the answers until they can
see it laid out in "living col
Charles Mills wrote:
> Are you sure? That's totally contrary to my impression.
>
> There are three states for the above machine:
>
> - both tasks waiting for I/O
> - one task waiting for I/O and the other task computing
> - either both tasks computing, or if a single CPU, one computing and the
>
PROTECTED] On Behalf
Of (IBM Mainframe Discussion List)
Sent: Tuesday, June 06, 2006 12:38 PM
To: IBM-MAIN@BAMA.UA.EDU
Subject: Re: One or two CPUs - the pros & cons
In a message dated 6/6/2006 1:52:09 P.M. Central Daylight Time,
[EMAIL PROTECTED] writes:
>1. A single 200 MIPS CPU
The only plus that I have ever gotten from two CPUs whose aggregate
power was approximately the same as a single CPU was in the even of a
high priority task getting in a tight loop. In this case, the second CPU
was often, but not always, handy to do diagnostics from.
Case in point was a vendor ST
www.ibm.com
Main page search for 'fewer faster'
Look at the 1st and 3rd entries.
--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the
In a message dated 6/6/2006 1:52:09 P.M. Central Daylight Time,
[EMAIL PROTECTED] writes:
>1. A single 200 MIPS CPU will always outperform two 100 MIPS CPUs.
Not always. As you said, it all depends on the particular workload. E.g.,
if the customer's mission-critical work is a long-runn
your particular
workload situation.
Charles
-Original Message-
From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf
Of Itschak Mugzach
Sent: Tuesday, June 06, 2006 2:47 PM
To: IBM-MAIN@BAMA.UA.EDU
Subject: One or two CPUs - the pros & cons
I wonder if there is a docu
lower engines, would acheive
the same results.
So now it workload related.
-Original Message-
From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf
Of Itschak Mugzach
Sent: Tuesday, June 06, 2006 2:47 PM
To: IBM-MAIN@BAMA.UA.EDU
Subject: One or two CPUs - the pros &
I wonder if there is a document that describes guidelines or rule of
thumbs when selecting machine configuration, engine wise. I mean, if you
get the mipsage on a uniprocessor, why have it on two? your advice is
welcome.
Itschak
--
50 matches
Mail list logo
