You've doubled the number of engines but approximately halved the
per-engine speed in moving from a 2097-504 to a 2827-408. IBM LSPR
indicates that the PCI is exactly the same for those two capacity models --
quite a coincidence, actually.

It's theoretically possible to do substantially better than the PCI
average, but I strongly suspect a reporting problem. (I'd like to be wrong,
though.)

Theoretically you could see that result if you only ever have a maximum of
4 tasks that can be dispatched at any time -- half the engines in your new
zEC12 would thus be idle, and capacity utilization could never exceed 50%
on the new machine. (I'm oversimplifying but only slightly.) Do you have a
lot of single threaded batch, for example, that would skew the utilization
reports?

Did you increase the memory to give DB2 more room to play? That can make a
substantial difference.

If you have a fair amount of Java code (as an example) that should do
substantially better than the PCI average.

--------------------------------------------------------------------------------------------------------
Timothy Sipples
Consulting Enterprise IT Architect (Based in Singapore)
E-Mail: [email protected]
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to