Re: Why HZ on i386 is 100 ?

[Davide Libenzi]

On Tue, 16 Apr 2002, David Mosberger wrote:

> >>>>> On Tue, 16 Apr 2002 10:18:18 -0700 (PDT), Davide Libenzi <[e-mail]> said:
>
>   Davide> i still have pieces of paper on my desk about tests done on
>   Davide> my dual piii where by hacking HZ to 1000 the kernel build
>   Davide> time went from an average of 2min:30sec to an average
>   Davide> 2min:43sec. that is pretty close to 10%
>
> The last time I measured timer tick overhead on ia64 it was well below
> 1% of overhead.  I don't really like using kernel builds as a
> benchmark, because there are far too many variables for the results to
> have any long-term or cross-platform value.  But since it's popular, I
> did measure it quickly on a relatively slow (old) Itanium box: with
> 100Hz, the kernel compile was about 0.6% faster than with 1024Hz
> (2.4.18 UP kernel).

uhm, this is quite interesting. it's quite possible at this point that
PROC_CHANGE_PENALTY put an high cs pressure in place, with terrible cache
effects. pretty sadly i was not running the sampler that would have helped
me to detect such behaviour.

- Davide