I expect it's because VIO requires cycles. Definitely true when simulating 
a device in memory. Quite possibly also true when doing VIO to page data 
sets on disk.

What I used to call the "DIM Coffee Table Book" :-) showed VIO to Expanded 
Storage burnt CPU compared to doing it to straight temp data sets on disk. 
For MOST data points.

Martin

Martin Packer
Performance Consultant
IBM United Kingdom Ltd
+44-20-8832-5167
+44-7802-245-584
[EMAIL PROTECTED]






From:
"O'Brien, David W. (NIH/CIT) [C]" <[EMAIL PROTECTED]>
To:
[email protected]
Date:
12/05/2008 17:51
Subject:
Question concerning performance, VIO and chargeback



We are running Z/os 1.7. It recently came to my attention that VIO had 
been implemented in such a way that temp datasets no matter how small 
ALWAYS went to DASD. One of my predessesors used a storclas with 
Guaranteed Space. Anyway a user asked to be able to use VIO so I turned it 
on. Today he comes back with test results showing higher chargeback for 
CPU when using VIO. My question is - How is this possible?
 
We are processing the same numbers of records and eliminating some 
physical I/O and the CPU usage is higher.
The total CPU for the job doubled. I would have thought that eliminating 
some of the channel activity would have reduced CPU usage. Can someone 
explain this to me?
 
Regards,
Dave O'Brien

 

----------------------------------------------------------------------
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









Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 
741598. 
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU






----------------------------------------------------------------------
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

Reply via email to