Re: Improved Intel Turbo Boost status/control

From: Alexander Motin <mav_at_FreeBSD.org>
Date: Mon, 12 Mar 2012 22:38:16 +0200
On 03/12/12 22:22, Ivan Klymenko wrote:
> В Mon, 12 Mar 2012 22:11:28 +0200
> Alexander Motin<mav_at_FreeBSD.org>  пишет:
>
>> On 03/12/12 22:05, Ivan Klymenko wrote:
>>> В Mon, 12 Mar 2012 21:55:21 +0200
>>> Alexander Motin<mav_at_FreeBSD.org>   пишет:
>>>
>>>> On 03/12/12 21:33, Ivan Klymenko wrote:
>>>>> В Mon, 12 Mar 2012 21:15:35 +0200
>>>>> Alexander Motin<mav_at_FreeBSD.org>    пишет:
>>>>>> I'd like to note that recent r232793 change to cpufreq(4) in HEAD
>>>>>> opened simple access to the  Intel Turbo Boost status/control.
>>>>>> I've found that at least two of my desktop systems (based
>>>>>> Nehalem and SandyBridge Core i7s) with enabled Intel Turbo Boost
>>>>>> in BIOS it is not use it by default, unless powerd is enabled.
>>>>>> And before this change it was difficult to detect/fix.
>>>>>>
>>>>>> ACPI reports extra performance level with frequency 1MHz above
>>>>>> the nominal to control Intel Turbo Boost operation. It is not a
>>>>>> bug, but feature:
>>>>>> dev.cpu.0.freq_levels: 2934/106000 2933/95000 2800/82000 ...
>>>>>> In this case value 2933 means 2.93GHz, but 2934 means 3.2-3.6GHz.
>>>>>>
>>>>>> After boot with default settings I see:
>>>>>> dev.cpu.0.freq: 2933
>>>>>> , that means Turbo Boost is disabled.
>>>>>>
>>>>>> Enabling powerd or just adding to rc.conf
>>>>>> performance_cpu_freq="HIGH"
>>>>>> enables Turbo Boost and adds extra 10-20% to the system
>>>>>> performance.
>>>>>>
>>>>>> Turbo Boost operation can be monitored in run-time via the PMC
>>>>>> with command that prints number or really executed cycles per
>>>>>> CPU core: pmcstat -s unhalted-core-cycles -w 1
>>>>>>
>>>>>
>>>>> Thank you very much!
>>>>> performance_cpu_freq="HIGH"
>>>>> and as this option must be combined with state of the processor C1
>>>>> C2 C3?
>>>>> performance_cx_lowest="XX"
>>>>> economy_cx_lowest="XX"
>>>>
>>>> The more CPU cores on package are sleeping and the deeper they are
>>>> sleeping, the bigger will be boost for remaining active cores.
>>>> Without using deeper C-states boost is usually quite small (about
>>>> 100-200MHz for desktop chips). Enabling C-states increases it in
>>>> few times.
>>>>
>>>
>>> I have a Core i5 c Turbo Boost technology (enabled in BIOS)
>>> After the following:
>>> sysctl dev.cpu.0.freq_levels
>>> dev.cpu.0.freq_levels: 2301/35000 2300/35000 2000/29079 1800/25766
>>> 1600/22265 1400/18904 1225/16541 1200/15996 1050/13996 1000/12907
>>> 875/11293 800/9956 700/8711 600/7467 500/6222 400/4978 300/3733
>>> 200/2489 100/1244
>>>
>>> performance_cpu_freq="HIGH">>   /etc/rc.conf
>>>
>>> /etc/rc.d/powerd restart
>>>
>>> sysctl dev.cpu.0.freq_levels
>>> dev.cpu.0.freq_levels: 2301/35000 2300/35000 2000/29079 1800/25766
>>> 1600/22265 1400/18904 1225/16541 1200/15996 1050/13996 1000/12907
>>> 875/11293 800/9956 700/8711 600/7467 500/6222 400/4978 300/3733
>>> 200/2489 100/1244
>>>
>>> CPU frequency does not rise above 2300 Mhz
>>>
>>> What am I doing wrong?
>>
>> performance_cpu_freq variable handled not by /etc/rc.d/powerd, but
>> /etc/rc.d/power_profile.
>>
>
> ok
>
> I remove and insert power supply unit connector - nothing has changed...
>
> sysctl dev.cpu.0.freq_levels
> dev.cpu.0.freq_levels: 2301/35000 2300/35000 2000/29079 1800/25766 1600/22265 1400/18904 1225/16541 1200/15996 1050/13996 1000/12907 875/11293 800/9956 700/8711 600/7467 500/6222 400/4978 300/3733 200/2489 100/1244

What changes do you expect to see in dev.cpu.0.freq_levels? This list is 
static. It is dev.cpu.0.freq that may change and that is where 
difference between 2301 and 2300 should now have effect.

-- 
Alexander Motin
Received on Mon Mar 12 2012 - 19:38:21 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:40:24 UTC