From mboxrd@z Thu Jan 1 00:00:00 1970 From: Nivedita Singhvi Subject: Re: Realtime linux kernel on multicore processor Date: Thu, 17 Jun 2010 13:16:44 -0700 Message-ID: <4C1A82AC.1000804@us.ibm.com> References: Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: linux-rt-users To: Neel Mehta Return-path: Received: from e5.ny.us.ibm.com ([32.97.182.145]:52532 "EHLO e5.ny.us.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S933058Ab0FQUQq (ORCPT ); Thu, 17 Jun 2010 16:16:46 -0400 Received: from d01relay03.pok.ibm.com (d01relay03.pok.ibm.com [9.56.227.235]) by e5.ny.us.ibm.com (8.14.4/8.13.1) with ESMTP id o5HJxjXE022079 for ; Thu, 17 Jun 2010 15:59:45 -0400 Received: from d01av04.pok.ibm.com (d01av04.pok.ibm.com [9.56.224.64]) by d01relay03.pok.ibm.com (8.13.8/8.13.8/NCO v10.0) with ESMTP id o5HKGjkX061190 for ; Thu, 17 Jun 2010 16:16:45 -0400 Received: from d01av04.pok.ibm.com (loopback [127.0.0.1]) by d01av04.pok.ibm.com (8.14.4/8.13.1/NCO v10.0 AVout) with ESMTP id o5HKGjSl004879 for ; Thu, 17 Jun 2010 16:16:45 -0400 In-Reply-To: Sender: linux-rt-users-owner@vger.kernel.org List-ID: Neel Mehta wrote: > Hi. I have 2 computers with ubuntu: > 1. Core 2 Duo - Non realtime linux kernel > 2. Quad core - Realtime linux kernel > > Both of them are using ubuntu. I am running a process that generates 2 > threads, both of which are continuously processing something. > 1. On core 2 duo (normal kernel): both the cores are running above 90% load. > 2. on quad core(realtime kernel): only one core runs above 90%, and > the other cores are not at such a high load. The core which is on high > load keeps on alternating among the 4 cores. > > So, my question is why doesnt 2 cores run on high load on quad core? > Is there a problem in multiprocessing in realtime linux kernel? Can > someone guide me where the source of problem could be? I'll start with the obvious - if the same amount of work is being done, but now on a larger number of cores, the load on each individual core could reduce (since you don't say what else is contributing to the load on the system, etc). As an example of this class of behavior, look at your irq distribution. The rotation of the load sounds a lot like it's being driven by interrupt round-robining, for instance. thanks, Nivedita