From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Paul E. McKenney" Subject: Scheduler model, revisited Date: Tue, 13 Dec 2011 11:17:07 -0800 Message-ID: <20111213191707.GA10771@linux.vnet.ibm.com> Reply-To: paulmck@linux.vnet.ibm.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: linux-rt-users@vger.kernel.org To: pjt@google.com Return-path: Received: from e3.ny.us.ibm.com ([32.97.182.143]:53977 "EHLO e3.ny.us.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752318Ab1LMTTK (ORCPT ); Tue, 13 Dec 2011 14:19:10 -0500 Received: from /spool/local by e3.ny.us.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted for from ; Tue, 13 Dec 2011 14:19:09 -0500 Received: from d03av01.boulder.ibm.com (d03av01.boulder.ibm.com [9.17.195.167]) by d01relay03.pok.ibm.com (8.13.8/8.13.8/NCO v10.0) with ESMTP id pBDJIj4u208472 for ; Tue, 13 Dec 2011 14:18:45 -0500 Received: from d03av01.boulder.ibm.com (loopback [127.0.0.1]) by d03av01.boulder.ibm.com (8.14.4/8.13.1/NCO v10.0 AVout) with ESMTP id pBDJH8ct024424 for ; Tue, 13 Dec 2011 12:17:09 -0700 Content-Disposition: inline Sender: linux-rt-users-owner@vger.kernel.org List-ID: Hello, Paul, Back in Prague, you mentioned that a number of your scheduler-parameter measurements were poor fits for the exponential distribution. Are they a reasonable fit for a series of exponentially distributed Markov states? For example, would they fit something of the following form? a^2*t*e^(-a*t) Or would a three-step series fit better? (a^3*t^2*e^(-a*t))/2 And there is always the four-step series: (a^4*t^3*e^(-a*t))/6 Or, in general: (a^n*t^(n-1)*e^(-a*t))/(n-1)! But please keep in mind that higher orders multiply my state space by the number of steps in the series. ;-) Thanx, Paul