Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 10 Mar 2020 10:33:45 -0700
From:      Conrad Meyer <cem@freebsd.org>
To:        Theron <theron.tarigo@gmail.com>
Cc:        FreeBSD Current <freebsd-current@freebsd.org>
Subject:   Re: System clock is slow
Message-ID:  <CAG6CVpU9ihfz9sSh=2foUzaLkT4sF_iLcNcxaVapMi5NVt6EpA@mail.gmail.com>
In-Reply-To: <a2603db8-f3e9-fa9d-b2cf-a80e55dc926e@gmail.com>
References:  <989cd36a-e015-940a-dfe2-851c6fdf4734@gmail.com> <20200310053833.GD3091@server.rulingia.com> <a2603db8-f3e9-fa9d-b2cf-a80e55dc926e@gmail.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, Mar 10, 2020 at 9:24 AM Theron <theron.tarigo@gmail.com> wrote:
> I had previously used powerd to let the CPU underclock to 700MHz when
> idle.  Now, I've lost all control over CPU frequency (either by powerd
> or by sysctl) since there is some in-kernel cpufreq driver which I can't
> figure out how to disable.

It's hwpstate_intel(4).  You can add
'hint.hwpstate_intel.0.disabled="1"' in loader.conf or device.hints to
disable and use est(4), if you prefer.

The idea of the device is that hardware can do a better job managing
the frequency / power states rapidly than daemons like powerd.  But if
you like powerd/est, feel free to disable it.  Future models of Intel
CPU may not provide est(4).

If you leave it enabled, you can control the energy efficiency /
performance trade-off of hwpstate_intel on a per-core basis (or
per-package, if not disabled and hardware supports package-level
control) with 'dev.hwpstate_intel.<cpu>.epp=0-100' (sysctl or
tunable); the CPU uses this knob to control how biased it is towards
low frequency states.  100 is most efficient, 0 is most performant.
The default is 50.

Best,
Conrad



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAG6CVpU9ihfz9sSh=2foUzaLkT4sF_iLcNcxaVapMi5NVt6EpA>