Re: Exadata CPU Upgrade

From: Kevin Closson <ora_kclosson_at_yahoo.com>
Date: Wed, 5 Sep 2012 13:25:08 -0700 (PDT)
Message-ID: <1346876708.63686.YahooMailNeo_at_web161705.mail.bf1.yahoo.com>



oh. I was asking about the specific quoted text where your numbers factored i 60 database and 96 cores. Lost me, that all. No biggie. I have seen your evernote page before but didn't see where that came into the thread. I guess I missed an email or something.

Karl wrote:

"I doubt that you will be able to fit that in a half rack Exadata :) let's say each of the database is coming from the same CPU as the Exadata and only needs 1core for its workload the sum of that will be 60cores / 96cores"



 From: Karl Arao <karlarao_at_gmail.com> To: Kevin Closson <ora_kclosson_at_yahoo.com> Cc: "qrasheed_at_gmail.com" <qrasheed_at_gmail.com>; Kellyn Pot'Vin <kellyn.potvin_at_ymail.com>; ORACLE-L <oracle-l_at_freelists.org> Sent: Wednesday, September 5, 2012 1:15 PM Subject: Re: Exadata CPU Upgrade  

If you check out the evernote link on this wiki http://goo.gl/M8mVW I'm also making use of logical cores from the source systems. So even if I make use of 12cores/node (real cores) as long as I'm using the same real cores number on the source I should be able to arrive at the same number. On that link the first example is an M900.. if we make this as 128 we will still arrive at the same number if we divide this by 12 and not by 24 (Exadata).

Exadata est cores requirement = source host cores * chip efficiency factor * utilization

  • 128 * 0.46 * 0.8
  • 47.104 cores

Then divide this by the number of cores of Exadata to get the number of compute nodes

=  47.104 /12

  • 3.925333333333333 nodes
-- 
Karl Arao
karlarao.wordpress.com
karlarao.tiddlyspot.com
twitter.com/karlarao


--
http://www.freelists.org/webpage/oracle-l

--
http://www.freelists.org/webpage/oracle-l
Received on Wed Sep 05 2012 - 15:25:08 CDT

Original text of this message