cvs commit: src/sys/kern sched_ule.c

[Bruce Evans]

On Tue, 2 Oct 2007, John Baldwin wrote:

> On Tuesday 02 October 2007 09:49:34 am Bruce Evans wrote:
>> Apparently, there is a scaling bug for hz or extra interrupts for
>> the larger hz help, and the default preempt_thresh is not best.
>
> -current on i386 and amd64 does a very poor job of scaling stathz and profhz
> with hz, so this may explain problems at hz=100.  I have an attempt to make
> stathz and profhz more sane while also trying to not always drive the lapic
> timer at hz * 2.  I used to have a test program that would display all the
> frequencies for different 'hz' values but have misplaced it. :(

I fixed that about a year ago in my version of -current, and may have sent
you the patch, and mentioned the need to fix it before using hz = 100
in a previous reply to this thread.  Neverthless, most things including
schedulers barely noticed when stahz is completely broken -- stathz needs
to be about 100 to work as intended, but most things didn't notice when
old bugs made it 1024 and most things don't notice when current bugs make
it 13.

[Patch deleted].
[My patch sent in private mail].

Bruce