From mboxrd@z Thu Jan 1 00:00:00 1970 From: Alex Shi Subject: bltk-game regressions on snb laptop Date: Tue, 16 Apr 2013 15:14:31 +0800 Message-ID: <516CFA57.3060709@intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Return-path: Received: from mga03.intel.com ([143.182.124.21]:47439 "EHLO mga03.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754127Ab3DPHPG (ORCPT ); Tue, 16 Apr 2013 03:15:06 -0400 Sender: linux-pm-owner@vger.kernel.org List-Id: linux-pm@vger.kernel.org To: Linux PM list , "Brown, Len" , "Wysocki, Rafael J" , Arjan van de Ven , LKP ML , Viresh Kumar LKP found a performance and performance/watt regression on commit aa77a52764a92216b61, acpi-cpufreq: Don't set policy->related_cpus from .init. The commit removed the related_cpus setting, plus Our laptop has no coordinate type setting in BIOS. So the related_cpus is only include the cpu self, then the policy->cpus are impacted and only has self too. With ondemand governor, the bad commit cause bltk-game benchmark drop to 18fps from 50fps, the performance/watt value also dropped a lot. bltk-game runs 9 thread on the 4core*HT laptop. As Arjan and Len mentioned, the commit is correct in logical, policy->cpus should only include the cpu self. So I don't know where is the problem, maybe ondemand or some place others. Anyway, I just report the issue to you. -- Thanks Alex