On Wed, 3 Feb 2010 08:34:21 -0800, Patrick Falcone
<[email protected]> wrote:
>Really? Is engine switching overhead with SP's and GP's CP's more of a
concern versus running WAS with GP CP's and other traditional workloads from
a performance perspective?
>
>Hmm. I've been there, no SP's, and it can be very ugly. I'm not sure I
follow you...So are you saying zAAP's really don't help the overall
performance of a mix of WAS and traditional workloads when running this type
of mix on the same LPAR?
>
No - not compared to the same number of total GP engines. There is switching
overhead to move the zAAP eligible workload. In the early zAAP days, there
was quite a bit of tweaking to the JVM to determine when it made sense to
switch or not. There were also some estimation tools to help you determine
what part of your application could run on a zAAP (prior to the RMF
changes that gave you the numbers) and the switch rate (I don't know if
there are any switch rate numbers in SMF records, but I don't think there
is).
I was looking for some SHARE sessions that could explain all of this,
but the ones I have looked at don't have the slides attached ("dispatcher
changes to support zAAP", "everything zAAP" were a few I looked at).
I stopped looking after that. Maybe someone else can point you to
documentation or white papers or wants to provide more details.
Mark
--
Mark Zelden
Sr. Software and Systems Architect - z/OS Team Lead
Zurich North America / Farmers Insurance Group - ZFUS G-ITO
mailto:[email protected]
z/OS Systems Programming expert at http://expertanswercenter.techtarget.com/
Mark's MVS Utilities: http://home.flash.net/~mzelden/mvsutil.html
----------------------------------------------------------------------
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.html
