Re: [Jackit-devel] Re: Statistical methods for latency profiling

[Lee Revell]

On Sat, 2004-07-31 at 22:55, Matt Mackall wrote:
> On Sat, Jul 31, 2004 at 01:22:37AM -0400, Lee Revell wrote:
> > Hey,
> > 
> > Recently Ingo Molnar asked in one of the voluntary-preempt threads for
> > the minimum and average scheduling delay reported by jackd.  JACK does
> > not currently maintain these statistics.
> > 
> > I realized that the distribution of maximum latencies reported on each
> > process cycle is fairly normally distributed.
> 
> This is not at all what I would expect. Instead, I'd expect to see
> something like a gamma distribution, where we have everything
> clustered down close to zero, but with a very long tail in the
> positive direction falling off exponentially and obviously a hard
> limit on the other side..

Right, it is a lot closer to a gamma distribution.  It's been years
since I have used any of this, and I took stat for psych majors, vs stat
for engineers.  I was a lot more interested in playing Doom at the
time...

This looks interesting:

http://www.itl.nist.gov/div898/handbook/eda/section3/ppccplot.htm
http://www.itl.nist.gov/div898/handbook/eda/section4/eda4291.htm

I will have some numbers soon.

Lee





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