From mboxrd@z Thu Jan 1 00:00:00 1970 Date: Sun, 8 Nov 2015 12:26:27 +0100 From: Gilles Chanteperdrix Message-ID: <20151108112627.GL1798@hermes.click-hack.org> References: <55E7F013.10101@free.fr> <20151104003032.GB24848@hermes.click-hack.org> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Disposition: inline Content-Transfer-Encoding: quoted-printable In-Reply-To: <20151104003032.GB24848@hermes.click-hack.org> Subject: Re: [Xenomai] intel turbo boost List-Id: Discussions about the Xenomai project List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: =?iso-8859-1?Q?St=E9phane?= ANCELOT Cc: "xenomai@xenomai.org" On Wed, Nov 04, 2015 at 01:30:32AM +0100, Gilles Chanteperdrix wrote: > On Thu, Sep 03, 2015 at 09:00:36AM +0200, St=E9phane ANCELOT wrote: > > Hi, > >=20 > > Do you think I can use the Intel integrated turbo boost mode, without=20 > > problems. > >=20 > > I have this nice processor : > > http://ark.intel.com/fr/products/76308/Intel-Core-i5-4300U-Processor-3M= -Cache-up-to-2_90-GHz >=20 > Just experimented this, the turbo mode is a problem for applications > which rely on a precise relation between the actual count of cpu > cycles of a busy loop and the time. The sched_quota falls under that > category and simply does not work reliably when turbo mode is > enabled. Fortunately, it is possible to turn off turbo mode. For > that, in the kernel configuration > - enable CPU_FREQ > - choose performance mode by default > - enable intel pstate driver > - echo 1 > /sys/devices/system/cpu/intel_pstate/no_turbo >=20 > Other than that, turbo mode does not seem to break the latency, > which would mean that the tsc frequency does not change with turbo > mode. Turbo mode also breaks the autotune feature. So, on a machine with turbo mode, autotune should be triggered after writing to the /proc file. --=20 Gilles. https://click-hack.org