Re: [QUESTION] kernel/Kconfig.hz: outdated help texts?

[John Stultz]

On Fri, Apr 24, 2026 at 11:30 AM Nathan Chancellor <nathan@xxxxxxxxxx> wrote:
> On Fri, Apr 24, 2026 at 10:18:44AM +0200, torvic9@xxxxxxxxxxx wrote:
> > Hello,
> >
> > I noticed that the help text in kernel/Kconfig.hz has barely changed since
> > configurable frequency of the timer was introduced in 2005.
> >
> > Are these help texts still accurate 21 years later?
> >
> > We have e.g.:
> >
> >   "250 Hz is a good compromise choice allowing server performance
> >    while also showing good interactive responsiveness even
> >    on SMP and NUMA systems. If you are going to be using NTSC video
> >    or multimedia, select 300 Hz instead."
> >
> > Is that NTSC/PAL thing still relevant today, in times of high resolution timers,
> > modern graphics cards, Freesync and high-refresh 4K monitors with 144 Hz?
> >

I believe the rationale there was just that the math for 60fps worked
better w/ HZ=300, as it gave you 5 ticks per frame rather than the
4.1666.. HZ=250 gives you.

hrtimers obviously makes this less critical, but it might be worth
some consideration as I know as framerates have gotten higher on
devices, using HZ=250 has been a pain point - as at 120fps, you only
8.333ms per frame and with HZ=250 the 4ms ticks doesn't give much time
(2.083 ticks) for cpu balancing to fix things if task wakeup placement
was not ideal.  That said HZ=300 doesn't really help much there, since
at 120fps you get 2.5 ticks, which is basically the same.

So for 60fps, I think HZ=300 might provide some benefit (which would
align with the Kconfig description), but I don't have any real world
numbers to show that at the moment.

thanks
-john