Date: Sat, 24 Jul 2010 11:45:30 -0700 From: Garrett Cooper <gcooper@FreeBSD.org> To: Rui Paulo <rpaulo@lavabit.com> Cc: freebsd-hackers@freebsd.org, Alexander Motin <mav@freebsd.org>, freebsd-performance@freebsd.org Subject: Re: Intel TurboBoost in practice Message-ID: <AANLkTik7XOEosrhLq9q0pyFzUKcQNedkOfW15osm=g4D@mail.gmail.com> In-Reply-To: <A30A636F-E925-456E-8866-4E46B3BA367F@lavabit.com> References: <4C4AF046.40507@FreeBSD.org> <A30A636F-E925-456E-8866-4E46B3BA367F@lavabit.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, Jul 24, 2010 at 9:18 AM, Rui Paulo <rpaulo@lavabit.com> wrote: > > On 24 Jul 2010, at 14:53, Alexander Motin wrote: > >> Hi. >> >> I've make small observations of Intel TurboBoost technology under >> FreeBSD. This technology allows Intel Core i5/i7 CPUs to rise frequency >> of some cores if other cores are idle and power/thermal conditions >> permit. CPU core counted as idle, if it has been put into C3 or deeper >> power state (may reflect ACPI C2/C3 states). So to reach maximal >> effectiveness, some tuning may be needed. >> >> Here is my test case: FreeBSD 9-CURRENT on Core i5 650 CPU, 3.2GHz + 1/2 >> TurboBoost steps (+133/+266MHz) with boxed cooler at the open air. I was >> measuring building time of the net/mpd5 from sources, using only one CPU >> core (cpuset -l 0 time make). >> >> Untuned system (hz=3D1000): =A0 =A0 14.15 sec >> Enabled ACPI C2 (hz=3D1000+C2): 13.85 sec >> Enabled ACPI C3 (hz=3D1000+C3): 13.91 sec >> Reduced HZ (hz=3D100): =A0 =A0 =A0 =A0 =A014.16 sec >> Enabled ACPI C2 (hz=3D100+C2): =A013.85 sec >> Enabled ACPI C3 (hz=3D100+C3): =A013.86 sec >> Timers tuned* (hz=3D100): =A0 =A0 =A0 14.10 sec >> Enabled ACPI C2 (hz=3D100+C2): =A013.71 sec >> Enabled ACPI C3 (hz=3D100+C3): =A013.73 sec >> >> All numbers tested few times and are repeatable up to +/-0.01sec. >> >> *) Timers were tuned to reduce interrupt rates and respectively increase >> idle cores sleep time. These lines were added to loader.conf: >> sysctl kern.eventtimer.timer1=3Di8254 >> sysctl kern.eventtimer.timer2=3DNONE >> kern.eventtimer.singlemul=3D1 >> kern.hz=3D"100" >> >> PS: In this case benefit is small, but it is the least that can be >> achieved, depending on CPU model. Some models allow frequency to be >> risen by up to 6 steps (+798MHz). > > The numbers that you are showing doesn't show much difference. Have you t= ried buildworld? Agreed. The numbers are small enough that there could be a large degree of variation just based on environmental factors alone; there are other things that go into that as well, such as disk I/O, etc, that probably shouldn't be factored into a CPU performance test. Thanks, -Garrett
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?AANLkTik7XOEosrhLq9q0pyFzUKcQNedkOfW15osm=g4D>