Re: Physical CPU? or multicore?

From: Amar Kumar Padhi <>
Date: Wed, 29 Apr 2009 08:47:28 +0400 (GST)
Message-ID: <27408161.6981240980235005.JavaMail.seven_at_aomfe1p1>

Hi Niall, more info. We actually had this need to compare the single and multi-cores for moving our existing 100s of servers to new ones. Yes today options are limited to multi-cores, but we still needed to know that keeping all other hardware aspects somewhat constant, how many cores match a single core. I also do infrastructure recommendation and that is the reason for me getting into this.

It has worked well as we were able to draw lines, put numbers instead of vague guesswork and recommend optimal standards. Yes, we looked at the changing attributes of the Cpu and other hardware (no VM for production though) , that also accounts for the final recommendation.

All this may not be required for a new shop starting with multi-cores only, in such case benchmarking the application on the cores itself will serve the purpose. For the clients I work with, most still continue to run old boxes and the comparison will keep coming up.


-original message-

 Re: Physical CPU? or multicore?
From: "Niall Litchfield" <> Date: 28-04-2009 22:53

I'm curious as to how you arrived at this, like at least the web links if not the internal testing, though why you would do internal testing just of the CPU I'm not sure.

AFAIK you cannot buy any system that has single nehalem cores. That is, you can only buy that CPU in multi-core variants. I believe the same is true of other multi-core CPUs that are available today. (as an aside I believe that "standard" CPUs today are multicore anyway but I digress).

So I'm just unsure what you compared with what and how.

Having said all that the CPU is not the unit of measurement I am not concerned with. How many IOPS against given storage (or if really pushed Java/C# TPS) would be my measurement, and it would be on a per-server basis.




Niall Litchfield
Oracle DBA

-- Received on Tue Apr 28 2009 - 23:47:28 CDT

Original text of this message