Re: powerd forgets top CPU frequency ?

From: <rb_at_gid.co.uk>
Date: Sun, 25 Aug 2024 17:47:54 UTC
Hi,

> On 25 Aug 2024, at 18:01, Matthew D. Fuller <fullermd@over-yonder.net> wrote:
> 
> On Sun, Aug 25, 2024 at 10:38:14AM +0200 I heard the voice of
> Ronald Klop, and lo! it spake thus:
>> 
>> This is interesting by itself. According to the sysctl the CPU only
>> has 1 frequency to select, which is 2803, but it is running on 3103.
>> Maybe there is some other mechanism which influences the CPU freq on
>> your machine. I hope somebody else can shine some light on this.
> 
> This seems fairly common on more recent Intel stuff.  On some systems
> (without powerd running, since it apparently doesn't do anything
> useful in such cases);
> 
> 
> dev.cpu.0.freq_levels: 3500/-1
> dev.cpu.0.freq: 1801
> dev.cpu.7.freq_levels: 3500/-1
> dev.cpu.7.freq: 799
> 
> Xeon E3 v5 (Skylake)
> 
> 
> dev.cpu.0.freq_levels: 3500/-1
> dev.cpu.0.freq: 900
> dev.cpu.1.freq_levels: 3500/-1
> dev.cpu.1.freq: 799
> 
> Xeon E-21xx (Coffee Lake)

Another data point:

dev.cpu.0.freq_levels: 2401/100000 2400/100000 2300/93921 2200/88941 2100/84058 2000/79269 1900/74574 1800/69972 1700/64541 1600/60135 1500/55820 1400/51595 1300/47460 1200/43413 1100/38596 1000/34736
dev.cpu.0.freq: 2400

Xeon Silver 4210R (Cascade Lake) (2 packages)

…but this system is running 12.2R (and not powerd). Another similar system, running powerd (but only 1 package) shows dev.cpu.0.freq varying between 2400 and 1000 as expected. FWIW these are Intel boards.

> Seems to just all be done by the hardware.
> 
> 
> -- 
> Matthew Fuller     (MF4839)   |  fullermd@over-yonder.net
> Systems/Network Administrator |  http://www.over-yonder.net/~fullermd/
>           On the Internet, nobody can hear you scream.
> 

--
Bob Bishop
rb@gid.co.uk