From mboxrd@z Thu Jan 1 00:00:00 1970 From: Konrad Rzeszutek Wilk Subject: Re: Xen 4.5 development update (July update) Date: Mon, 4 Aug 2014 11:44:39 -0400 Message-ID: <20140804154439.GD20097@laptop.dumpdata.com> References: <20140723002823.3F7C4BD884@laptop.dumpdata.com> <1406568922.4038.23.camel@Solace> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Received: from mail6.bemta4.messagelabs.com ([85.158.143.247]) by lists.xen.org with esmtp (Exim 4.72) (envelope-from ) id 1XEKTf-0007Sm-FL for xen-devel@lists.xenproject.org; Mon, 04 Aug 2014 15:47:11 +0000 Content-Disposition: inline In-Reply-To: <1406568922.4038.23.camel@Solace> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xen.org Errors-To: xen-devel-bounces@lists.xen.org To: Dario Faggioli Cc: artem.mygaiev@globallogic.com, msw@amazon.com, Ian.Jackson@eu.citrix.com, dongxiao.xu@intel.com, mengxu@cis.upenn.edu, JBeulich@suse.com, feng.wu@intel.com, zhigang.x.wang@oracle.com, parth.dixit@linaro.org, Paul.Skentzos@dornerworks.com, eddie.dong@intel.com, rcojocaru@bitdefender.com, guijianfeng@cn.fujitsu.com, Vijaya.Kumar@caviumnetworks.com, josh.whitehead@dornerworks.com, zoltan.kiss@citrix.com, avanzini.arianna@gmail.com, yang.z.zhang@intel.com, xen-devel@lists.xenproject.org, serge.broslavsky@linaro.org, yjhyun.yoo@samsung.com, olaf@aepfle.de, wency@cn.fujitsu.com, Ian.Campbell@citrix.com, vijay.kilari@gmail.com, stefano.stabellini@eu.citrix.com, mcgrof@suse.com, julien.grall@linaro.org, dave.scott@citrix.com, robert.vanvossen@dornerworks.com, shantong.kang@intel.com, roy.franz@linaro.org, anthony.perard@citrix.co.m, Paul.Durrant@citrix.com, ufimtseva@gmail.com List-Id: xen-devel@lists.xenproject.org On Mon, Jul 28, 2014 at 07:35:22PM +0200, Dario Faggioli wrote: > On mar, 2014-07-22 at 20:28 -0400, konrad.wilk@oracle.com wrote: > > > == x86 == > > > * HT enabled, virtualization overhead is high (Xen 4.4) (none) > > kernbench demonstrated it > > looking and tracing it > > - Dario Faggioli > > > There is a (sub-)thread sill ongoing here: > http://www.gossamer-threads.com/lists/xen/devel/339409 > > However, there have been found counterexamples, i.e., numbers showing > that virt overhead is similar with and without HT, so I think this > bullet can go away (not sure with what tracking status...) The 'This confirms, for me, that it's an SMT balancing issue that we're seen. ' is not an issue anymore? As in we can get bad and good results depending on the workload? Now here is another question - did you run the guests with 'xen_nopvspin' to see if this was an PV ticketlock (or the lack of it) issue? Also what did the " I'll try more runs, e.g. with number of VCPUs equal less than nr_corse/2 and see what happens." come out to be? Thank you! > > Dario > > -- > <> (Raistlin Majere) > ----------------------------------------------------------------- > Dario Faggioli, Ph.D, http://about.me/dario.faggioli > Senior Software Engineer, Citrix Systems R&D Ltd., Cambridge (UK) >