Hiperpool usage

Hello

Hiperpool usage
Hi all ,
I've just added 10000 pages of 4K to my hiperpool , observe that there is
much less read then written pages into the hiperpool , but also see that
there is less I/O and better performance.
Some wrote on the list that when there is much less read the written pages
into the hiperpool , its usage is not necessary.

What i observe that i have less I/O and it is better , even if a lot of
page that are moved to the hiperpool won't never be used , but those are
used don't need to be read from DASD. What do you think about it ?

---------------------------------------------------------------------------------
Welcome to the IDUG DB2-L list. To unsubscribe, go to the archives and home page at http://www.idugdb2-l.org/archives/db2-l.html. From that page select "Join or Leave the list". The IDUG DB2-L FAQ is at http://www.idugdb2-l.org. The IDUG List Admins can be reached at [login to unmask email] Find out the latest on IDUG conferences at http://conferences.idug.org/index.cfm

Martin Packer

Re: Hiperpool usage
(in response to Hello)
Daniel, the point about reads and writes is that they both represent CPU
cost. The benefits are all in the reads, of course.

If to get the read benefits you are prepared to take the read/write CPU
cost then who cares what the ratio is?

Note: The read/write RATES are readily measurable. Indeed we stack them up
in a bar chart by pool - when doing engagements with our clients. What
ISN'T measurable is the CPU cost. You have to guesstimate that: I tend to
use (VERY hand wavily and not for serious belief) 10us per page movement.
It just makes the maths simpler!

Now, if you are on a 64 bit machine, particularly if you have a "large"
subsystem, I'd be looking very seriously at Dataspace buffer pools and
removing those hiperpools. I think even with the same TOTAL pool size
(HP+VP = DP) you would get better buffer pool effectiveness. And you
certainly would save virtual storage.

Virtual storage is one of those things where you either think "what the
heck do I care about that for?" or "cripes, I'd better fix this NOW!" :-)
Actually I see a fair number of customers who are about to have a VS crunch
and aren't really aware of it. Which is why we always ask for IFCID 225
(Stats Trace Class 6) data.

Martin

Martin Packer, MBCS Martin Packer/UK/IBM
020-8832-5167 in the UK (+44) (MOBX 273643, Internal 7-325167, Mobile
07802-245584)



Daniel
<[login to unmask email]
M> To
Sent by: DB2 Data [login to unmask email]
Base Discussion cc
List
<[login to unmask email] Subject
ORG> Hiperpool usage


20-01-04 09:05


Please respond to
DB2 Database
Discussion list
at IDUG






Hi all ,
I've just added 10000 pages of 4K to my hiperpool , observe that there is
much less read then written pages into the hiperpool , but also see that
there is less I/O and better performance.
Some wrote on the list that when there is much less read the written pages
into the hiperpool , its usage is not necessary.

What i observe that i have less I/O and it is better , even if a lot of
page that are moved to the hiperpool won't never be used , but those are
used don't need to be read from DASD. What do you think about it ?

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

Welcome to the IDUG DB2-L list. To unsubscribe, go to the archives and home
page at http://www.idugdb2-l.org/archives/db2-l.html. From that page select
"Join or Leave the list". The IDUG DB2-L FAQ is at http://www.idugdb2-l.org
. The IDUG List Admins can be reached at [login to unmask email]
Find out the latest on IDUG conferences at
http://conferences.idug.org/index.cfm

---------------------------------------------------------------------------------
Welcome to the IDUG DB2-L list. To unsubscribe, go to the archives and home page at http://www.idugdb2-l.org/archives/db2-l.html. From that page select "Join or Leave the list". The IDUG DB2-L FAQ is at http://www.idugdb2-l.org. The IDUG List Admins can be reached at [login to unmask email] Find out the latest on IDUG conferences at http://conferences.idug.org/index.cfm