Re: HZ=1024 for CPU >= 586?

Erik Corry (erik@arbat.com)
Mon, 18 Jan 1999 13:07:46 +0100

On Mon, Jan 18, 1999 at 03:42:33AM -0800, Christoph Lameter wrote:
> On Mon, 18 Jan 1999, Erik Corry wrote:
>
> > In article <19990118083931.D10800@knt.terra.vein.hu> you wrote:
> > > On Fri, Jan 15, 1999 at 11:01:09PM -0800, Christoph Lameter wrote:
> > >> I recompiled 2.2.0pre7ac4 with HZ=1024 and the responsiveness of the
> > >> Mousecursor under X is significantly improved. Lagging keyboard input etc
> > >> etc have all been reduced.
> >
> > Sounds like a bug. It shouldn't make much difference.
>
> It should and does make a difference since timeslices with HZ=100 might be
> relatively long.

Yes, but above the refresh rate of your monitor, so how come you can
see it?

> > >> Could we have the kernel increase the HZ for the modern faster CPUs?
> > >> F.e. leave it at 100 HZ for 386 486. Switch to 1024 for 586 and higher.
> >
> > Note that Linus has said that HZ is 100 to user processes
> > now and forever, so anywhere where the user space can see
> > what HZ is (some things in /proc, a few ioctls, etc.) it
> > should stay at 100. Internally you can use higher values.
>
> For alpha cpus Linux already uses 1024. User processes like top see the
> increase in HZ. And top shows wrong values...'

Sorry, I meant on Intel. On Alpha presumably user-visible HZ is
1024 now and forever. On Sparc it's a sysctl (needed anyway for
Solaris compatibility).

-- 
Erik Corry erik@arbat.com           Ceterum censeo, Microsoftem esse delendam!


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/