Re: CWSPEEDNW/CPUSPEED vs. actual Cpu speed
From: Karl Arao <karlarao_at_gmail.com>
Date: Sat, 28 Jul 2012 00:09:17 -0500
Message-ID: <CACNsJneocf=47cqzQFw2gZf+4qWZVZLbaBQEAMrws8R0xaEn-w_at_mail.gmail.com>
Here's my two cents
I don't think this could be correlated to the CPU speed of a processor.. and you can modify this for specific reasons like the issue I had before where a pure OLTP system that upgraded from an old to new hardware the CPU speed
was faster on the new environment that made it to change a lot of plans - then pushing the system stats back to the old hardware value made it go back to the old plans.
Date: Sat, 28 Jul 2012 00:09:17 -0500
Message-ID: <CACNsJneocf=47cqzQFw2gZf+4qWZVZLbaBQEAMrws8R0xaEn-w_at_mail.gmail.com>
Here's my two cents
I don't think this could be correlated to the CPU speed of a processor.. and you can modify this for specific reasons like the issue I had before where a pure OLTP system that upgraded from an old to new hardware the CPU speed
was faster on the new environment that made it to change a lot of plans - then pushing the system stats back to the old hardware value made it go back to the old plans.
If you really want to have a CPU speed comparison you can make use of the cpu - SPECint_rate2006 http://goo.gl/doBI5 where as I've explained on that wiki link it makes use of the "SPECint_rate2006 per core" as a CPU speed multiplier for server sizing purposes. I've also mentioned the other ways of deriving the CPU speed on that link.
And a simple HOWTO on how to use the SPECint_rate2006 for CPU core comparison is shown here http://goo.gl/M8mVW
-- Karl Arao karlarao.wordpress.com karlarao.tiddlyspot.com -- http://www.freelists.org/webpage/oracle-lReceived on Sat Jul 28 2012 - 00:09:17 CDT