dtrace discuss - Sep 2007 - about profile provider, maybe a stupid problem...

in profile provider, using "profile-xxxhz" as the name
the probe name maybe
    97, 199, 499, 997, 1999,
    4001, 4999 HZ

err... why are these numbers not multiples of 10/100/1000?

TIA

Regards,
TJ
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mail.opensolaris.org/pipermail/dtrace-discuss/attachments/20070907/308d2530/attachment.html>

?? TaoJie wrote:
> in profile provider, using "profile-xxxhz" as the name
> the probe name maybe
>     97, 199, 499, 997, 1999,
>     4001, 4999 HZ
> 
> err... why are these numbers not multiples of 10/100/1000?
> 
To skew the firing of the probes a bit so they don''t always coincide 
with the regular clock-based stuff (which runs on multiples of 10/100/1000).

Menno
-- 
Menno Lageman - Sun Microsystems - http://blogs.sun.com/menno

thank you, Menno.

err, a further step...
if the firing profile probes always coincide with the regular clock-based
stuff, they will cause what?
I mean what will happen? Would it be serious?


Regards,
TJ


2007/9/7, Menno Lageman <Menno.Lageman at
sun.com>:
>
> ?? TaoJie wrote:
> > in profile provider, using "profile-xxxhz" as the name
> > the probe name maybe
> >     97, 199, 499, 997, 1999,
> >     4001, 4999 HZ
> >
> > err... why are these numbers not multiples of 10/100/1000?
> >
>
> To skew the firing of the probes a bit so they don''t always
coincide
> with the regular clock-based stuff (which runs on multiples of
> 10/100/1000).
>
> Menno
> --
> Menno Lageman - Sun Microsystems - http://blogs.sun.com/menno
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mail.opensolaris.org/pipermail/dtrace-discuss/attachments/20070907/5fb7898a/attachment.html>

Euler,

You want to reduce interference from regular clock events in your sampling.

On 9/7/07, ?? TaoJie <eulertao at gmail.com>
wrote:
> thank you, Menno.
>
> err, a further step...
> if the firing profile probes always coincide with the regular clock-based
> stuff, they will cause what?
> I mean what will happen? Would it be serious?
>
>
> Regards,
> TJ
>
>
> 2007/9/7, Menno Lageman <Menno.Lageman at sun.com>:
> > ?? TaoJie wrote:
> > > in profile provider, using "profile-xxxhz" as the name
> > > the probe name maybe
> > >     97, 199, 499, 997, 1999,
> > >     4001, 4999 HZ
> > >
> > > err... why are these numbers not multiples of 10/100/1000?
> > >
> >
> > To skew the firing of the probes a bit so they don''t always
coincide
> > with the regular clock-based stuff (which runs on multiples of
> 10/100/1000).
> >
> > Menno
> > --
> > Menno Lageman - Sun Microsystems - http://blogs.sun.com/menno
> >
>
>
> _______________________________________________
> dtrace-discuss mailing list
> dtrace-discuss at opensolaris.org
>

-- 
Just me,
Wire ...
Blog: <prstat.blogspot.com>