From mboxrd@z Thu Jan 1 00:00:00 1970 From: Vadim Rozenfeld Subject: Re: Performance issue Date: Thu, 29 Nov 2012 22:34:07 +0200 Message-ID: <201211292234.08380.vrozenfe@redhat.com> References: <20121129135610.GB17318@redhat.com> Mime-Version: 1.0 Content-Type: Text/Plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: "George-Cristian =?utf-8?q?B=C3=AErzan?=" , kvm@vger.kernel.org To: Gleb Natapov Return-path: Received: from mx1.redhat.com ([209.132.183.28]:46522 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751399Ab2K2UeR convert rfc822-to-8bit (ORCPT ); Thu, 29 Nov 2012 15:34:17 -0500 In-Reply-To: <20121129135610.GB17318@redhat.com> Sender: kvm-owner@vger.kernel.org List-ID: On Thursday, November 29, 2012 03:56:10 PM Gleb Natapov wrote: > On Thu, Nov 29, 2012 at 03:45:52PM +0200, George-Cristian B=C3=AErzan= wrote: > > On Thu, Nov 29, 2012 at 1:56 PM, Vadim Rozenfeld =20 wrote: > > >> I've also added +hv_relaxed since then, but this is the command = I'm > > >=20 > > > I would suggest activating relaxed timing for all W2K8R2/Win7 gue= sts. > >=20 > > Is there any place I can read up on the downsides of this for Linux= , > > or is Just Better? >=20 > You shouldn't use hyper-v flags for Linux guests. In theory Linux sho= uld > just ignore them, in practice there may be bugs that will prevent Lin= ux > from detecting that it runs as a guest and disable optimizations. >=20 As Gleb said, hyper-v flag are relevant to the Windows guests only.=20 IIRC spinlocks and vapic should work for Vista and higher. Relaxed timi= ng and partition reference time work for Win7/W2K8R2. > > >>>> Other than that, was looking into a profiling trace of the sof= tware > > >>=20 > > >> running and a lot of time (60%?) is spent calling two functions = from > > >> hal.dll, HalpGetPmTimerSleepModePerfCounter when I disable HPET,= and > > >> HalpHPETProgramRolloverTimer which do point at something related= to > > >> the timers. > > >=20 > > > It means that hyper-v time stamp source was not activated. > >=20 > > I recompiled the whole kernel, with your patch, and while I cannot > > check at 70Mbps now, a test stream of 20 seems to do better. Also, = now > > I don't see any of those functions, which used to account ~60% of t= he > > time spent by the program. I'm waiting for the customer to come bac= k > > and start the 'real' stream, but from my tests, time spent in hal.d= ll > > is now an order of magnitude smaller. > >=20 > > -- > > George-Cristian B=C3=AErzan >=20 > -- > Gleb.